From owner-jsoftware@lists.interlog.com Fri Jul 2 15:20:45 1999 Date: Fri, 02 Jul 1999 15:14:00 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Re: Jforum: J403 CDROMs References: <001c01beab87$6f23bba0$150114d1@Pcdburke> Mine just arrived yesterday. Nice physical packaging -- avoiding the usual situation of opening a mailer only to find a cracked jewel box -- but perhaps "J in a Box" is no longer a correct description! Chris Burke wrote: > > The J403 CDROMS will likely not be shipped until Jun 19th, so Pro > users may want to download the system first. -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 2 23:57:24 1999 Date: Fri, 02 Jul 1999 23:55:47 -0400 From: Brian Bambrough Subject: Jforum: Problems using Project Manager I'm having difficulty creating a runtime module for a system I'm developing. Below are two test scripts that give me the same kind of difficulties. The important (to me) features of these scripts are: - the script, test1, is loaded into it's own locale, test1 - test1 uses wdview, a typical j utility What I am trying to do is to build a runtime module to run on a system that doesn't have J loaded on it. Attempt 1: Here's what I did in Project Manager, on my desktop computer: Add to source files: test.ijs, test1.ijs Open project tab: add target file test.ijr Click on build. Build options are to include: project source project libraries standard libraries Close J At the DOS prompt enter c:\j403b\j.exe /rt c:\j403b\temp\test.ijr The test system runs fine. Copy j.exe, j.dll and test.ijr to directory c:\test on my laptop computer, which doesn't have j403b on it. Enter j /rt c:\test\test.ijr Get an error message "file not found: c:\j403b\temp\test1.ijs" Fair enough, I've hard coded a directory that doesn't exist. Question: How can I avoid hard coding the directory for test1.ijs in the first line of test.ijs? Attempt 2: I changed the first line of test.ijs to 'test1' load 'c:\test\test1', on the desktop computer, and rebuilt test.ijr. The laptop still couldn't find test1.ijs. Question: Why wasn't test1.ijs included in test.ijr by the Project Manager? I copied test1.ijs to the c:\test directory on the laptop. When I ran the runtime I got the message: file not found: c:\j403b\system\classes\view\jview.ijs Attempt 3: I went back to the desktop, entered Project Manager, and told it to include jview in the required files. I rebuilt test.ijr, copied it to the laptop and got the same error message. Questions: How can I get the test system to work using Project Manager? Is there a way to build test.ijr, with all needed files inside test.ijr (test1 and any other required files) without using Project Manager. Anyone who can give me step by step instructions to do this will earn my eternal gratitude. In the commercial system, I need verbs from stdlib.ijs. It seems that Project Manager doesn't have any way to include it in the .ijr file. I had to include it in the main script as: 'z' load 'winlib.ijs'. Is there a better way to do this? Script test.ijs -------------------------------------------- 'test1' load 'c:\j403b\temp\test1' TEST=: 0 : 0 pc test; xywh 132 67 57 11;cc btnTest1 button;cn "Display text window"; pas 6 6;pcenter; rem form end; ) test_run=: 3 : 0 wd TEST wd 'pshow;' ) test_close=: 3 : 0 wd'pclose' ) test_btnTest1_button=: 3 : 0 test1_run_test1_'' ) test_run'' Script test1.ijs -------------------------------------------- TEST1=: 0 : 0 pc test1; xywh 53 41 34 11;cc btnWiew button;cn "Show text"; pas 6 6;pcenter; rem form end; ) test1_run=: 3 : 0 wd TEST1 wd 'pshow;' ) test1_close=: 3 : 0 wd'pclose' ) test1_btnWiew_button=: 3 : 0 wdview 'Sample text',LF,LF,'for viewing' ) -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 02:21:51 1999 Date: Sat, 03 Jul 1999 08:56:56 +0300 From: Anssi Seppala Subject: Re: Jforum: Problems using Project Manager In-Reply-To: <377D89C3.D71AD5BD@worldnet.att.net> My PM works fine. In Build options I have checked in Include files: Project source, project libraries and standard libraries. Anssi At 23:55 2.7.1999 -0400, you wrote: >I'm having difficulty creating a runtime module for a system I'm >developing. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 06:20:46 1999 From: "Chris Burke" References: <377D89C3.D71AD5BD@worldnet.att.net> Subject: Re: Jforum: Problems using Project Manager Date: Sat, 3 Jul 1999 06:20:32 -0400 charset="iso-8859-1" Do not build an ijr file before you know that the corresponding ijs file works! It is essentially impossible to debug an ijr file. You cannot read it, and if it fails, you get no J session. So start off by building an ijs file and testing that. When you are sure it runs correctly, then build the ijr file. To find out what PM is doing when it builds a script, read the target ijs file, which can be opened from the Project tab of PM. When experimenting with PM, you may find it useful to view the target ijs file with an external editor. This makes it easy to experiment with different build options, build the target, then switch to the editor and view it. ----- Original Message ----- From: Brian Bambrough Sent: Friday, July 02, 1999 11:55 PM Subject: Jforum: Problems using Project Manager > I'm having difficulty creating a runtime module for a system I'm > developing. Below are two test scripts that give me the same kind of > difficulties. The important (to me) features of these scripts are: -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 13:15:01 1999 From: "Seymour Glass" Subject: RE: Jforum: MAPI/rtn code 11 Date: Sat, 3 Jul 1999 12:46:09 -0400 charset="iso-8859-1" In-Reply-To: <89FE544F2452D21190A50060B0684BF3899D24@szrh00305.tszrh.csfb.com> Importance: Normal The return codes from sendmail (which I guess you have aliased to send) are the ones returned by the Windows DLL. The MAPI return codes are documented in the Windows file 'mapi.h' in MSDN. Return code 11 is MAPI_E_ATTACHMENT_NOT_FOUND, with no further description. Henry Rich > -----Original Message----- > From: owner-jsoftware@lists.interlog.com > [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Ondrus, Milan > Sent: Wednesday, June 30, 1999 05:42 > To: 'Jforum' > Subject: Jforum: MAPI/rtn code 11 > > > > does anybody know what this returncode 11 means that comes from > > returncode =. sess send parameterlist (script mapi.ijs) > > its a dll return code but I could not find the meaning in the > documentation. > This code 11 I get sometimes if I send out mails containing an > attachment - and the mail is then of course not > sent. > > /Milan Ondrus > > > > > > ------------------------------------------------------------------ > -------------- > J Forum: for information about this list, see > http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 14:34:07 1999 Date: Sat, 03 Jul 1999 14:31:56 -0400 From: Brian Bambrough Subject: Re: Jforum: Problems using Project Manager References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> Chris and Anssi; Thanks for taking the time to respond. However, I am still having problems. Here is a very specific question: how do I preserve locales when I prepare a file with Table Manager? Here are two scripts. test1 is in the base locale, test2 is in the test2 locale. They each set a global variable, text. When I run test1, text is handled correctly. Now, I use Table Manager to create a single script test.ijs. I load test1 and test2 as source files. Table Manager creates one script, test, by gluing the two scripts together. When I run test everything runs in the base locale and there is only one text. ------------------------------------------------------- 'test2' load 'c:\test1\test2.ijs' text=:'Script test1' TEST1=: 0 : 0 pc test1; xywh 34 21 86 11;cc edtText edit ws_border es_autohscroll; xywh 35 4 59 11;cc btnShowtest2 button;cn "Show test2"; pas 6 6;pcenter; rem form end; ) test1_run=: 3 : 0 wd TEST1 wd 'set edtText "',text,'"' wd 'pshow;' ) test1_close=: 3 : 0 wd'pclose' ) test1_btnShowtest2_button=: 3 : 0 test2_run_test2_'' ) ------------------------------------------------------- test1_run'' text=:'Script test2' TEST2=: 0 : 0 pc test2; menupop "File"; menu new "&New" "" "" ""; menu open "&Open" "" "" ""; menusep ; menu exit "&Exit" "" "" ""; menupopz; xywh 67 23 76 11;cc edtText edit ws_border es_autohscroll; pas 6 6;pcenter; rem form end; ) test2_run=: 3 : 0 wd TEST2 wd 'set edtText "',text,'"' wd 'pshow;' ) test2_close=: 3 : 0 wd'pclose' wd 'psel test1' wd 'set edtText "',text_base_,'"' ) test2_cancel_button=: 3 : 0 test2_close'' ) -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 16:49:05 1999 From: "Chris Burke" References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> Subject: Re: Jforum: Problems using Project Manager Date: Sat, 3 Jul 1999 16:49:11 -0400 charset="iso-8859-1" 1. You could add a "cocurrent" (or "coclass") statement at the top of the test2 script, i.e. add: cocurrent 'test2' This changes the locale into which subsequent definitions will be loaded. Since the single script has test1 first, this gets loaded into base, while test2 will be loaded into test2. In this case, you would not include the line: > 'test2' load 'c:\test1\test2.ijs' 2. In general, I prefer to have each project loaded into a single locale (which can be set in the Project|Build Options dialog). An app that loads into 2 locales would then consist of 2 projects, and 2 target scripts. You would then have to ensure both get loaded, or have some way of running them together for the final build. Note that for complex builds you can add a pre or post processor, that could, for example, run the target scripts together into a single one, so you dont have to do this step manually. > Here is a very specific question: how do I preserve locales when I > prepare a file with Table Manager? > > Here are two scripts. test1 is in the base locale, test2 is in the > test2 locale. They each set a global variable, text. When I run > test1, text is handled correctly. > > Now, I use Table Manager to create a single script test.ijs. I load > test1 and test2 as source files. Table Manager creates one script, > test, by gluing the two scripts together. When I run test everything > runs in the base locale and there is only one text. > > ------------------------------------------------------- > > 'test2' load 'c:\test1\test2.ijs' > > text=:'Script test1' -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 3 17:02:08 1999 Date: Sun, 04 Jul 1999 00:01:25 +0300 From: Anssi Seppala Subject: Re: Jforum: Problems using Project Manager In-Reply-To: <377E571C.37E7EE2@worldnet.att.net> References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> I have a convention to include all definitions to one locale in one script. The first statement of a script is cocurrent 'mylocale' NB. definitions to mylocale or cocurrent 'base' NB. definitions to base locale Anssi At 14:31 3.7.1999 -0400, you wrote: >Here is a very specific question: how do I preserve locales when I >prepare a file with Table Manager? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 03:54:22 1999 Date: Sun, 04 Jul 1999 10:42:11 +0300 From: Anssi Seppala Subject: Re: Jforum: Problems using Project Manager In-Reply-To: <377EDB34.47F8586F@worldnet.att.net> References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> <001301bec595$a0f8a000$480114d1@Pcdburke> Select jview from the library and build then the project. Anssi At 23:55 3.7.1999 -0400, you wrote: >I have one more problem: In the test1 script I added a line to use >wdview. This works fine in the J environment and also works when I >invoke it as: j /rt c:\test1\test.ijs. However, when I try to do the >same thing on my laptop, which doesn't have j403b loaded, I have a >problem. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 03:54:33 1999 Date: Sun, 04 Jul 1999 10:51:49 +0300 From: Anssi Seppala Subject: Jforum: Complex builds with PM In-Reply-To: <001301bec595$a0f8a000$480114d1@Pcdburke> References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> Chris, can you give an example of a pre- or postprocessor? I do not fully understand the idea. I combine two projects by including the target of the first project as a source to the other project. In general I try to separate calculations from GUI. I keep a calculation library as one project and GUI as a other project. The target of the calculation libary is source to the GUI project and the target of the GUI project is the final application script. Anssi At 16:49 3.7.1999 -0400 Chris wrote: >Note that for complex builds you can add a pre or post processor, that >could, for example, run the target scripts together into a single one, so >you dont have to do this step manually. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 07:44:48 1999 From: "Chris Burke" References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> <001301bec595$a0f8a000$480114d1@Pcdburke> <377EDB34.47F8586F@worldnet.att.net> Subject: Re: Jforum: Problems using Project Manager Date: Sun, 4 Jul 1999 07:41:55 -0400 charset="iso-8859-1" ----- Original Message ----- From: Brian Bambrough Sent: Saturday, July 03, 1999 11:55 PM Subject: Re: Jforum: Problems using Project Manager > Definite progress is being made. Chris' and Anssi's suggestion to use > cocurrent worked fine. It seemed to me that I would want test1_run'' > to be the last line to execute, so I put the scripts into Project > Manager in the order test2, test1. Alternatively, take test1_run'' out of the scripts, and enter it in the "Append to file:" edit in Build Options. > I have one more problem: In the test1 script I added a line to use > wdview. This works fine in the J environment and also works when I > invoke it as: j /rt c:\test1\test.ijs. However, when I try to do the > same thing on my laptop, which doesn't have j403b loaded, I have a > problem. > ... > The culprit seems to be lines that the Project Manager adds after all > the included J scripts and before the test2 script. One of these > lines is 3586 (!) characters long and starts: > > PUBLIC_j_=:67 3$'colib';'c:\j403b\system\main\colib.ijs';(,'z'); > > This hard codes c:\403b\, which is not a directory on my laptop. How > can I fix this? You get this when you check one of the "Standard libraries" checkboxes in Build Options. The idea is that when you want the standard libraries, you get not only the scripts such as stdlib.ijs and winlib.ijs, but also the ability to load any of the standard supplied scripts (e.g. load 'plot'). To permit this, the PUBLIC_j_ table you see in the development system is written to the target file. It may not be useful to permit "load" in the target (for example, because of the problem with hard-coded directories), so in the J403b PM, there is an option "do not include load libraries". I'm inclined to think this should be the default. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 07:44:51 1999 From: "Chris Burke" References: <377D89C3.D71AD5BD@worldnet.att.net><002201bec53d$bf7bef80$1b0114d1@Pcdburke><377E571C.37E7EE2@worldnet.att.net> <4.1.19990704104241.00a687f0@pop.kolumbus.fi> Subject: Re: Jforum: Complex builds with PM Date: Sun, 4 Jul 1999 07:19:31 -0400 charset="iso-8859-1" The actions when you run Build in PM are: If a "pre-build" script is defined, load it. If a "target" script is defined, build it. If a "post-build" script is defined, load it. The scripts are defined in the Project tab of PM. You can put anything you like into pre and post-build. As an example, in your project, instead of including the target of the first project as a source to the other, you could create a post-build script that joins the two, something like: dat=. (freads 'target1.ijs'), LF, freads 'target2.ijs' dat fwrites 'finalapp.ijs' Sometimes, I just define a single post-build script which creates the target, and do not use the PM build facility. > Chris, can you give an example of a pre- or postprocessor? I do not fully > understand the idea. > > I combine two projects by including the target of the first project as a > source to the other project. In general I try to separate calculations from > GUI. I keep a calculation library as one project and GUI as a other > project. The target of the calculation libary is source to the GUI project > and the target of the GUI project is the final application script. > > Anssi -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 07:50:53 1999 From: "Chris Burke" References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> <001301bec595$a0f8a000$480114d1@Pcdburke> <377EDB34.47F8586F@worldnet.att.net> Subject: Re: Jforum: Problems using Project Manager Date: Sun, 4 Jul 1999 07:51:48 -0400 charset="iso-8859-1" If you would like to see the code for PM build, you can download the PM source from http://www.jsoftware.com/scripts.htm , file projman.zip, script buildapp.ijs. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 13:04:12 1999 From: "Seymour Glass" Subject: Jforum: Miscellaneous grid questions Date: Sun, 4 Jul 1999 13:01:46 -0400 charset="iso-8859-1" In-Reply-To: <000901beb3fd$441e8060$3c0114d1@Pcdburke> Importance: Normal Using Brian's code as a 'go-by' (thanks Brian!), I have started using the grid control as a selectbox with colors: output-only with support just for mouse clicks. Works OK, but I have some questions: 1. glgridspace seems to have no effect. No matter what I set it to, I get the same result. I am setting it just before I fill up the grid; is that wrong? 2. The font sizes don't match the sizes in other controls. I have to specify "courier new" 14 in my grid control to get characters that are the same size as "courier new" 11 in listboxes. Is this because of mapping modes? 3. I am having trouble understanding the stored state for gl2 commands. Here's how I think it works for grids: there is a workarea that holds the 'grid state', which includes grid data, attributes, etc. Separately, there is the gl2 graphics display buffer. When the 'glgrid' command is executed, all that happens is that a command is added to the display buffer. Data is added to the end of the display buffer by 'gl*' commands, but the added data is not available for display until a 'glshow*' is executed (so there is an internal pointer that indicates what data has been released for display by 'glshow*' - data after that is not visible yet). The buffer is traversed by 'glpaint*'. When the buffer is traversed, pixels are drawn. So, if I put a single 'glgrid' command into the display buffer, and show it, it will be there every time a repaint occurs, and will display the grid using the grid state at the time of the repaint. Is that right? Since the grid state is not double-buffered (or is it?), is there a chance that the repaint will catch a grid state that is in the middle of being modified (in other words, is the grid state used at an interrupt level?) 4. On the way to developing the theory given above, I was putting a glgrid into the display buffer every time I drew the grid. I'm not surprised this was slow, but I am surprised that I got incorrect display: the grid state seemed to be messed up, with incorrect colors. 5. Drawing the grid control takes longer than I expected. My graphics card is very cheap. Is there a way to turn off drawing the lines altogether? I have the background and the lines set to white, but I suspect the lines are being drawn anyway. 6. I am filling the grid column-by-column. Before I experiment with it, does anyone know if filling the whole grid at once is markedly faster than filling it a column at a time? Henry Rich -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 14:04:55 1999 From: "Oleg Kobchenko" References: <000101bec63e$e64c88c0$4489f7a5@seymourg> Subject: Re: Jforum: Miscellaneous grid questions Date: Sun, 4 Jul 1999 21:03:26 +0300 charset="iso-8859-1" ----- Original Message ----- From: Seymour Glass Sent: Sunday, July 04, 1999 8:01 PM Subject: Jforum: Miscellaneous grid questions > 2. The font sizes don't match the sizes in other controls. > I have to specify "courier new" 14 in my grid control to > get characters that are the same size as "courier new" 11 > in listboxes. Is this because of mapping modes? This is an overall J problem. The fonts are displayed in average -2 points off the indicated amount. I first thought it was due to the distortion of 'Large Fonts' Windows settings. However, I tried it on 'Large Fonts' and the problem remained. E.g. selecting 9pt in Session Manager console font will yield tiny 7pt or so font, whereas in other applications -- Wordpad, Visual Studio -- 9pt corresponds to a valid nice readable font. It is noticable even in the difference between the font that is displayed in the Font Preview pane of the Font Select dialog box and the actual appearence of that same font size in J itself and its GUI applications. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 20:35:59 1999 Date: Sun, 04 Jul 1999 20:33:57 -0400 From: Brian Bambrough Subject: Re: Jforum: Problems using Project Manager References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> <001301bec595$a0f8a000$480114d1@Pcdburke> <377EDB34.47F8586F@worldnet.att.net> <003501bec613$b38b9460$200114d1@Pcdburke> Success! I created a test.ijs script and got it to run on my laptop. But I had to do it without the aid of the Project Manager. The method I used was to write a batch file to assemble all the needed scripts into one big script, then to hand edit it. I'll post step-by- step details later. Chris and Anssi; Have either of you actually created test.ijs with Project Manager then loaded it onto a computer that doesn't have J loaded? To the best of my understanding there is no way that the script can work. This is because, e.g., jview.ijs has the line "require 'print'" in it. require goes off to a non-existent directory to find print.ijs, and the app crashes. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 4 21:37:05 1999 Subject: Jforum: "Pop J" From: Keith Smillie Date: Sun, 4 Jul 1999 19:34:25 -0600 (MDT) Content-Transfer-Encoding: 8bit I have just completed an article entitled �Thoughts on J and Other Languages� which was supposed to give a short description of some of my experiences with the APL, Nial and J and also some of the other programming languages I have used. When it was finished, it was 38 pages in length, and may be - although I�m not sure - an informal and rather idiosyncratic introduction to J. Downloadable Postscript, PDF and J Script files are available from my J Page at http://www.cs.ualberta.ca/~smillie/Jpage.htm Any comments or suggestions would be appreciated. Thanks. Keith -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 5 04:31:37 1999 Date: Mon, 05 Jul 1999 10:31:16 +0300 From: Anssi Seppala Subject: Re: Jforum: Problems using Project Manager In-Reply-To: <377FFD75.CE4D10F5@worldnet.att.net> References: <377D89C3.D71AD5BD@worldnet.att.net> <002201bec53d$bf7bef80$1b0114d1@Pcdburke> <377E571C.37E7EE2@worldnet.att.net> <001301bec595$a0f8a000$480114d1@Pcdburke> <377EDB34.47F8586F@worldnet.att.net> <003501bec613$b38b9460$200114d1@Pcdburke> Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id EAA17437 Brian, I have done several run-time applications distributed to several tens of systems without J. Project Manager does the work exellent and I am gratefull that it included in J system. My J work started at the same time PM was first introduced and all my J programming and code management is done with PM. PM is something that makes J a superior development system to other APL versions where you try to mange your code in workspaces. I do not understand why you did not succeed including view. I have included jview to my applications and build them with PM. If there is required scripts PM includes them too to the final script. Anssi At 20:33 4.7.1999 -0400, you wrote: >Chris and Anssi; > >Have either of you actually created test.ijs with Project Manager then >loaded it onto a computer that doesn't have J loaded? To the best of >my understanding there is no way that the script can work. This is >because, e.g., jview.ijs has the line "require 'print'" in it. require >goes off to a non-existent directory to find print.ijs, and the app >crashes. ------------------------------------------------------------------- Anssi Sepp�l�, TkT S�hk�markkinat ja s�hk�njakelu: Sovellukset ja ohjelmistot e-mail: anssi.seppala@enease.fi tel: +358 9 45400550 Enease Oy fax: +358 9 45400551 Mannerheimintie 66 A mob: +358 400 760077 FIN - 00260 HELSINKI, Finland http://www.enease.fi/ -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 5 09:35:05 1999 Date: Mon, 05 Jul 1999 09:30:29 -0400 From: Cliff Reiter Subject: Re: Jforum: "Pop J" Organization: Lafayette College Content-type: text/plain; charset=iso-8859-1 References: <19990705013435Z13372-25638+125@scapa.cs.ualberta.ca> I had the pleasure of reading both a draft of this paper and the final form -- it contains some provoking ideas and tries to articulate some substantial notions (e.g. how can we explain to a person on the street why array based languages are better?). It also contains several interesting examples. An amusing anedote from Keith's experiences in the early days appears (I won't spoil it for you) as well as some wise insights he has had as a learner of Japanese (make sure you read his 19th section: Teaching languages") -- very challenging remarks for those who take seriously the view that programming languages are languages (which may need to be taught). I'ld be interested in hearing other Jforum reader's reactions to the paper. Cliff Keith Smillie wrote: > > I have just completed an article entitled �Thoughts > on J and Other Languages� which was supposed to > give a short description of some of my experiences > with the APL, Nial and J and also some of the other > programming languages I have used. When it was > finished, it was 38 pages in length, and may be - > although I�m not sure - an informal and rather > idiosyncratic introduction to J. Downloadable > Postscript, PDF and J Script files are available > from my J Page at > http://www.cs.ualberta.ca/~smillie/Jpage.htm > > Any comments or suggestions would be > appreciated. > > Thanks. > > Keith -- Clifford A. Reiter Mathematics Department, Lafayette College Easton, PA 18042 USA, 610-330-5277 http://www.lafayette.edu/~reiterc -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 5 16:42:27 1999 Date: Mon, 05 Jul 1999 16:36:11 -0400 From: Cliff Reiter Subject: Jforum: coset enumeration Organization: Lafayette College Content-type: text/plain; charset=us-ascii References: <19990705013435Z13372-25638+125@scapa.cs.ualberta.ca> <3780B375.6D8B@lafvax.lafayette.edu> I was wondering if anyone has written coset enumeration in J? -- Clifford A. Reiter Mathematics Department, Lafayette College Easton, PA 18042 USA, 610-330-5277 http://www.lafayette.edu/~reiterc -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 5 16:47:39 1999 Date: Mon, 05 Jul 1999 16:42:41 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Jforum: MS Access 2000 ODBC driver for J 3.04b How do I go about updating the ODBC driver for Microsoft Access after upgrading to Access 2000 (under Windows 98)? I installed J 3.04b prior to upgrading from Office 97 to Office 2000. I looked on Microsoft's Web site (at the URL referenced in the J User Manual), but that seems only to point to a humungous (about 7 MB) download of something called "UDA" (Universal Data Access). -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 5 17:16:58 1999 From: gosi@centrum.is Subject: Re: Jforum: Pop J Date: Mon, 5 Jul 99 21:14:07 +0000 > Keith Smillie wrote: > > > > I have just completed an article entitled �Thoughts > > on J and Other Languages� Very good and interesting article > Downloadable > > Postscript, PDF and J Script files are available > > from my J Page at It is so easy to share articles in PDF form. Thanks /Gosi -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 09:44:50 1999 From: Andrew Nikitin Subject: RE: Jforum: Problems using Project Manager Date: Tue, 6 Jul 1999 08:40:07 -0500 Hi Brian. Some time ago I also had problems with 'require' while building scripts. Though I didn't use Project Manager, but rather scriptmake_j_ utility, my 'solution' can help you: > -----Original Message----- > From: Brian Bambrough [SMTP:b.bambrough@worldnet.att.net] > Sent: Sunday, July 04, 1999 19:34 > > This is because, e.g., jview.ijs has the line "require 'print'" in it. > require > goes off to a non-existent directory to find print.ijs, and the app > crashes. > I have inserted llines 3 : 0 '' if. ('1'=wd'qrt') do. require=:] 9!:1 <. +/ 6!:0 '' wd 'fontdef "MS Sans Serif" 8' end. ) in front of my very first script. It just blocks the 'require' subroutine (2 other lines set random seed and form font, that is not essential for the current subject). Since all nessesary libraries are already included in the runtime script, 'require' will never work correctly (i.e. skip already loaded libraries) at runtime for both PM and scriptmake_j_ load models. nsg -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 10:04:17 1999 Date: Tue, 06 Jul 1999 10:02:10 -0400 From: Brian Bambrough Subject: Re: Jforum: Problems using Project Manager References: Andrew; That's the exact same solution I came up with yesterday, though I implemented it differently. I, too, did not use Project Manager. I built a DOS batch file to assemble my run-time script. Just before I copied jview.ijs I inserted a file containing a single line: require=:3 : '''''"_ y.' I successfully built the runtime script for the alpha version of my system and emailed it to my client. I'm going to document my approach in detail and post it here later. Brian Andrew Nikitin wrote: > Hi Brian. > > Some time ago I also had problems with 'require' while building scripts. > Though I didn't use Project Manager, but rather scriptmake_j_ utility, > my 'solution' can help you: > > > -----Original Message----- > > From: Brian Bambrough [SMTP:b.bambrough@worldnet.att.net] > > Sent: Sunday, July 04, 1999 19:34 > > > > This is because, e.g., jview.ijs has the line "require 'print'" in it. > > require > > goes off to a non-existent directory to find print.ijs, and the app > > crashes. > > > I have inserted llines > > 3 : 0 '' > if. ('1'=wd'qrt') do. > require=:] > 9!:1 <. +/ 6!:0 '' > wd 'fontdef "MS Sans Serif" 8' > end. > ) > > in front of my very first script. It just blocks the 'require' > subroutine (2 other lines set random seed and form font, that is not > essential for the current subject). Since all nessesary libraries are > already included in the runtime script, 'require' will never work > correctly (i.e. skip already loaded libraries) at runtime for both PM > and scriptmake_j_ load models. > > nsg > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 12:54:02 1999 From: "HARRIS,BILL (HP-LakeStevens,ex1)" Subject: RE: Jforum: "Pop J" Date: Tue, 6 Jul 1999 10:51:12 -0600 charset="iso-8859-1" It looks like an interesting article. However, when I try to print it, page 15 won't print. Rather, it gives a "The document could not be printed. Read less image data than expected." Has anyone else had this problem? I haven't had success printing the Postscript version yet, either. Bill -- Bill Harris Year 2000 Program Office mailto: bill_harris@am.exch.hp.com Hewlett-Packard Company phone: (425) 335-2200 M/S 330C fax: (425) 335-2483 8600 Soper Hill Road web: http://hpweb.lsid.hp.com:8080/~billh/ Everett, WA 98205-1298 -----Original Message----- From: Keith Smillie [mailto:smillie@cs.ualberta.ca] Sent: Sunday, July 04, 1999 6:34 PM Subject: Jforum: "Pop J" -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 13:11:07 1999 From: "Chris Burke" References: <737091A8F919D31195E300A0C9968641667584@xboi06.boi.hp.com> Subject: Re: Jforum: "Pop J" Date: Tue, 6 Jul 1999 12:59:19 -0400 charset="iso-8859-1" It IS an interesting article! I also got the "read less image data than expected" message, but as a warning only. The article continued to print, but some of the images were completely black. ----- Original Message ----- From: HARRIS,BILL (HP-LakeStevens,ex1) Sent: Tuesday, July 06, 1999 12:51 PM Subject: RE: Jforum: "Pop J" > It looks like an interesting article. However, when I try to print it, page > 15 won't print. Rather, it gives a "The document could not be printed. > Read less image data than expected." > > Has anyone else had this problem? I haven't had success printing the > Postscript version yet, either. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 13:11:42 1999 From: "Chris Burke" References: <37820C62.CBBA0BF6@worldnet.att.net> Subject: Re: Jforum: Problems using Project Manager Date: Tue, 6 Jul 1999 13:10:26 -0400 charset="iso-8859-1" I saw this problem with "require" some time ago, and thought I had fixed it. Either I did not do a good job, or I broke it again! The problem is that a build including the standard libraries includes winlib.ijs. Ideally this should be completely standalone, but wdview (and now wdmenu) in turn require other libraries. wdview requires "print", which in turn requires "gl2". In the (misguided) interests of economy, these required libraries are not included unless requested, and also the "require" statement itself is probably left in. I need to go through this again, but unfortunately, cannot do so immediately. I'll try to get a fix up in a few days. > I, too, did not use Project Manager. I built a DOS batch file to assemble my > run-time script. Just before I copied jview.ijs I inserted a file containing a > single line: It would probably be better to create a "postbuild" script to do all this in PM, and have no target file. In this case, clicking "Build" will just run the postbuild script. This should be more convenient than running a DOS batch file, as well as being easier to program. I sometimes do this myself, when the PM build does not do exactly what I want. > require=:3 : '''''"_ y.' ----- Original Message ----- From: Brian Bambrough Sent: Tuesday, July 06, 1999 10:02 AM Subject: Re: Jforum: Problems using Project Manager > Andrew; > > That's the exact same solution I came up with yesterday, though I implemented it > differently. > > I, too, did not use Project Manager. I built a DOS batch file to assemble my > run-time script. Just before I copied jview.ijs I inserted a file containing a > single line: > > require=:3 : '''''"_ y.' > > I successfully built the runtime script for the alpha version of my system and > emailed it to my client. > > I'm going to document my approach in detail and post it here later. > > Brian -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From ljdickey Tue Jul 6 13:25:14 1999 Subject: Re: Jforum: "Pop J" I downloaded the postscript version and it printed with no difficulties. > I also got the "read less image data than expected" message, but as a > warning only. The article continued to print, but some of the images were > completely black. > > From: HARRIS,BILL (HP-LakeStevens,ex1) > To: > Sent: Tuesday, July 06, 1999 12:51 PM > Subject: RE: Jforum: "Pop J" > > > > It looks like an interesting article. However, when I try to print it, > page > > 15 won't print. Rather, it gives a "The document could not be printed. > > Read less image data than expected." > > > > Has anyone else had this problem? I haven't had success printing the > > Postscript version yet, either. From owner-jsoftware@lists.interlog.com Tue Jul 6 13:27:34 1999 From: Andrew Nikitin Subject: RE: Jforum: Problems using Project Manager Date: Tue, 6 Jul 1999 12:24:40 -0500 > -----Original Message----- > From: Chris Burke [SMTP:cdburke@interlog.com] > Sent: Tuesday, July 06, 1999 12:10 > > wdview requires "print", which > in turn requires "gl2". In the (misguided) interests of economy, these > required libraries are not included unless requested, and also the > "require" > statement itself is probably left in. Probably, require should return wether it loaded or not the requested source script. Using this information utility (wdview) may gray print button or make other nesessary decisions. Also there should be a way to say at runtime that some scripts are already loaded (so that if they checked manualy in PM, require (at runtime) will return positive result). In this case if user wants the wdview with extra functionality (print) he or she should check print script library manualy in PM. nsg -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 13:28:09 1999 From: Lee Dickey Date: Tue, 6 Jul 1999 13:25:14 -0400 (EDT) Subject: Re: Jforum: "Pop J" I downloaded the postscript version and it printed with no difficulties. > I also got the "read less image data than expected" message, but as a > warning only. The article continued to print, but some of the images were > completely black. > > From: HARRIS,BILL (HP-LakeStevens,ex1) > To: > Sent: Tuesday, July 06, 1999 12:51 PM > Subject: RE: Jforum: "Pop J" > > > > It looks like an interesting article. However, when I try to print it, > page > > 15 won't print. Rather, it gives a "The document could not be printed. > > Read less image data than expected." > > > > Has anyone else had this problem? I haven't had success printing the > > Postscript version yet, either. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 13:36:17 1999 From: "Ga�tan Godin" References: <737091A8F919D31195E300A0C9968641667584@xboi06.boi.hp.com> <008001bec7d2$773e8c40$d70114d1@Pcdburke> Subject: Re: Jforum: "Pop J" Date: Tue, 6 Jul 1999 13:33:06 -0400 charset="iso-8859-1" I also got the warning message but I got all images. I am using Acrobat reader version 4.0 and my computer has 256mb RAM. gg ----- Original Message ----- From: Chris Burke Sent: Tuesday, July 06, 1999 12:59 PM Subject: Re: Jforum: "Pop J" > It IS an interesting article! > > I also got the "read less image data than expected" message, but as a > warning only. The article continued to print, but some of the images were > completely black. > > ----- Original Message ----- > From: HARRIS,BILL (HP-LakeStevens,ex1) > To: > Sent: Tuesday, July 06, 1999 12:51 PM > Subject: RE: Jforum: "Pop J" > > > > It looks like an interesting article. However, when I try to print it, > page > > 15 won't print. Rather, it gives a "The document could not be printed. > > Read less image data than expected." > > > > Has anyone else had this problem? I haven't had success printing the > > Postscript version yet, either. > > > > -------------------------------------------------------------------------- ------ > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 18:52:41 1999 Date: Tue, 06 Jul 1999 18:47:33 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Re: Jforum: "Pop J" References: <199907061725.NAA16689@pythagoras.math.uwaterloo.ca> What printer/printer driver? Perhaps we can collectively troubleshoot this! I tried the Adobe .pdf version and, like others, got the "read less image data than expected" message and had several of the images completely black. This was with a HP LaserJet4M but with a standard Windows HPLJ (PCL) printer driver. Lee Dickey wrote: > > I downloaded the postscript version and it printed with > no difficulties. > > > I also got the "read less image data than expected" message, but as a > > warning only. The article continued to print, but some of the images were > > completely black. > > > > From: HARRIS,BILL (HP-LakeStevens,ex1) > > To: > > Sent: Tuesday, July 06, 1999 12:51 PM > > Subject: RE: Jforum: "Pop J" > > > > > > > It looks like an interesting article. However, when I try to print it, > > page > > > 15 won't print. Rather, it gives a "The document could not be printed. > > > Read less image data than expected." > > > > > > Has anyone else had this problem? I haven't had success printing the > > > Postscript version yet, either. > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 19:40:42 1999 Date: Tue, 06 Jul 1999 19:38:30 -0400 From: Jim Russell Subject: Re: Jforum: "Pop J" References: <737091A8F919D31195E300A0C9968641667584@xboi06.boi.hp.com> HARRIS,BILL (HP-LakeStevens,ex1) wrote: > > It looks like an interesting article. However, when I try to print it, page > 15 won't print. Rather, it gives a "The document could not be printed. > Read less image data than expected." > > Has anyone else had this problem? I haven't had success printing the > Postscript version yet, either. > I opened the .pdf version with acrobat reader, and printed from there. It also gave a similar error message, and left some figures garbled, but at least the text printed. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 20:27:34 1999 Date: Tue, 6 Jul 1999 20:23:22 -0400 From: "Les Redman" Subject: Re: Jforum: "Pop J" Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id UAA22370 Only 256mb -- that's awful! Shall we take up a collection? The illustrations on pages 16, 17, 19 can be printed by ticking "Print as Image" on the print dialog. Les Redman >>> "Ga (F�tan Godin" 07/06 1:33 PM >>> I also got the warning message but I got all images. I am using Acrobat reader version 4.0 and my computer has 256mb RAM. gg ----- Original Message ----- From: Chris Burke Sent: Tuesday, July 06, 1999 12:59 PM Subject: Re: Jforum: "Pop J" > It IS an interesting article! > > I also got the "read less image data than expected" message, but as a > warning only. The article continued to print, but some of the images were > completely black. > > ----- Original Message ----- > From: HARRIS,BILL (HP-LakeStevens,ex1) > To: > Sent: Tuesday, July 06, 1999 12:51 PM > Subject: RE: Jforum: "Pop J" > > > > It looks like an interesting article. However, when I try to print it, > page > > 15 won't print. Rather, it gives a "The document could not be printed. > > Read less image data than expected." > > > > Has anyone else had this problem? I haven't had success printing the > > Postscript version yet, either. > > > > -------------------------------------------------------------------------- ------ > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 21:50:29 1999 From: "Ga�tan Godin" References: Subject: Re: Jforum: "Pop J" Date: Tue, 6 Jul 1999 21:47:03 -0400 charset="iso-8859-1" Content-Transfer-Encoding: 8bit Interesting comment in light of today's memory costs. When using a language that allows full use of memory in manipulating its objects, spending $159 for an extra 128 mb does not seem extravagant. gg ----- Original Message ----- From: Les Redman Sent: Tuesday, July 06, 1999 8:23 PM Subject: Re: Jforum: "Pop J" Only 256mb -- that's awful! Shall we take up a collection? The illustrations on pages 16, 17, 19 can be printed by ticking "Print as Image" on the print dialog. Les Redman >>> "Ga (F�tan Godin" 07/06 1:33 PM >>> I also got the warning message but I got all images. I am using Acrobat reader version 4.0 and my computer has 256mb RAM. gg ----- Original Message ----- From: Chris Burke Sent: Tuesday, July 06, 1999 12:59 PM Subject: Re: Jforum: "Pop J" > It IS an interesting article! > > I also got the "read less image data than expected" message, but as a > warning only. The article continued to print, but some of the images were > completely black. > > ----- Original Message ----- > From: HARRIS,BILL (HP-LakeStevens,ex1) > To: > Sent: Tuesday, July 06, 1999 12:51 PM > Subject: RE: Jforum: "Pop J" > > > > It looks like an interesting article. However, when I try to print it, > page > > 15 won't print. Rather, it gives a "The document could not be printed. > > Read less image data than expected." > > > > Has anyone else had this problem? I haven't had success printing the > > Postscript version yet, either. > > > > -------------------------------------------------------------------------- ------ > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > ---------------------------------------------------------------------------- ---- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm ---------------------------------------------------------------------------- ---- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 22:48:28 1999 Date: Tue, 06 Jul 1999 22:47:15 -0400 From: Brian Bambrough Subject: Re: Jforum: "Pop J" References: <001101bec81a$9cd40c80$a898dbcf@godin.on.ca> Content-Transfer-Encoding: 8bit $159 for an extra 128 mb? When I go to computer shows in New Hampshire I can get memory for under $200 for 256 MB. "Ga�tan Godin" wrote: > Interesting comment in light of today's memory costs. When using a language > that allows full use of memory in manipulating its objects, spending $159 > for an extra 128 mb does not seem extravagant. > gg > > ----- Original Message ----- > From: Les Redman > To: > Sent: Tuesday, July 06, 1999 8:23 PM > Subject: Re: Jforum: "Pop J" > > Only 256mb -- that's awful! Shall we take up a collection? > > The illustrations on pages 16, 17, 19 can be printed by ticking "Print as > Image" on the print dialog. > > Les Redman > > >>> "Ga (F�tan Godin" 07/06 1:33 PM >>> > I also got the warning message but I got all images. I am using Acrobat > reader version 4.0 and my computer has 256mb RAM. > gg > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 6 22:49:27 1999 Date: Tue, 06 Jul 1999 22:48:17 -0400 From: Brian Bambrough Subject: Re: Jforum: Problems using Project Manager References: <37820C62.CBBA0BF6@worldnet.att.net> <008101bec7d2$77e4de60$d70114d1@Pcdburke> Chris; So it was a bug in Project Manager after all. Well, that's a relief. There is a saying: "that which does not kill us makes us stronger." I certainly feel that way. I have a much deeper understanding of how a run-time script should be built. Based on my experiences of the last week, I have some suggestions that you might want to consider: J documentation is terse. This is good, because there is so much of it (which is also good). I find the documentation to be complete, concise and useful. In the case of PM however, I found some of the documentation to be so terse as to be unintelligible. I feel that it needs to be rewritten from the perspective of a professional programmer who is coming to it for the first time. A good model would be the documentation for the form editor, which is terrific. I would assume that many of the people who use PM are writing serious commercial systems. It would be good if your test suite contained the same kinds of elements. These would include multiple locales, file handling, wdview. Include in the documentation the suggestion of putting cocurrent '...' at the start of each script. One thing I would dearly love to see at the top right hand corner of both PM and form editor is the little question mark icon. Click on it, then on one of the buttons, and find out what the button does. >> I, too, did not use Project Manager. I built a DOS batch file to >> assemble my run-time script. Just before I copied jview.ijs I >> inserted a file containing a single line: > It would probably be better to create a "postbuild" script to do all > this in PM, and have no target file. In this case, clicking "Build" > will just run the postbuild script. This should be more convenient > than running a DOS batch file, as well as being easier to program. I > sometimes do this myself, when the PM build does not do exactly what I > want. There are a bunch of reasons why I will stick with the DOS batch file approach for a while: For me it's very convenient to use a DOS batch file. This is because I use PC-Write to edit J scripts, so I'm in and out of DOS all the time. In DOS I hit two keys, then press Enter. In J I hit Ctrl B then click on "Build". I like to stay as "close to the silicon" as is practical. A DOS batch file is much more fundamental than an app like PM. I don't know what a postbuild script is. To find out, I would have to get back on the learning curve. Finally, right now DOS batch files work correctly but PM doesn't. PM is a great idea, and very convenient to have right in the J environment. I expect that, a few months from now, I will be using it exclusively. Brian -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 7 12:09:11 1999 Date: Wed, 07 Jul 1999 11:58:55 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Re: Jforum: "Pop J" References: "Printing as image" from Acrobat is a workaround -- but hardly a solution! The problem is that _text_ printed as image comes out looking quite inferior to text printed in the normal way, with many small breaks in the letters (the sort of thing an editor would mark up on galley proofs of a book or article). P.S. The "print as image" workaround is the same thing recommended for printing .pdf files created from TeX originals that otherwise refuse to properly recognize certain symbols; the results are equally, or more, unsatisfactory as in the present case. [There the cause is entirely different -- Acrobat's inability to deal with characters encoded in low-numbered positions.] Les Redman wrote: > > The illustrations on pages 16, 17, 19 can be printed by ticking "Print as > Image" on the print dialog. > -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 7 12:27:57 1999 Date: Wed, 7 Jul 1999 12:23:58 -0400 From: "Les Redman" Subject: Re: Jforum: "Pop J" Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id MAA06493 I agree, it's not a solution, it just happened to work for me, and did so without degrading the text, so I suppose it must depend on differences in one's "environment". Les Redman >>> Murray Eisenberg 07/07 11:58 AM >>> "Printing as image" from Acrobat is a workaround -- but hardly a solution! The problem is that _text_ printed as image comes out looking quite inferior to text printed in the normal way, with many small breaks in the letters (the sort of thing an editor would mark up on galley proofs of a book or article). P.S. The "print as image" workaround is the same thing recommended for printing .pdf files created from TeX originals that otherwise refuse to properly recognize certain symbols; the results are equally, or more, unsatisfactory as in the present case. [There the cause is entirely different -- Acrobat's inability to deal with characters encoded in low-numbered positions.] Les Redman wrote: > > The illustrations on pages 16, 17, 19 can be printed by ticking "Print as > Image" on the print dialog. > -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 7 15:04:13 1999 Date: Wed, 07 Jul 1999 20:59:01 +0200 From: "d.alis" Subject: Re: Jforum: "Pop J" References: <3783793F.F46C493C@math.umass.edu> Hi.. I had the same problems.. but i've generated a new pdf file from the postscript which printed fine. I can send you a copy if you want. David Murray Eisenberg wrote: > "Printing as image" from Acrobat is a workaround -- but hardly a > solution! The problem is that _text_ printed as image comes out looking > quite inferior to text printed in the normal way, with many small breaks > in the letters (the sort of thing an editor would mark up on galley > proofs of a book or article). > > P.S. The "print as image" workaround is the same thing recommended for > printing .pdf files created from TeX originals that otherwise refuse to > properly recognize certain symbols; the results are equally, or more, > unsatisfactory as in the present case. [There the cause is entirely > different -- Acrobat's inability to deal with characters encoded in > low-numbered positions.] > > Les Redman wrote: > > > > The illustrations on pages 16, 17, 19 can be printed by ticking "Print as > > Image" on the print dialog. > > > > -- > Murray Eisenberg murray@math.umass.edu > Mathematics & Statistics Dept. phone 413 549-1020 (H) > Univ. of Massachusetts 413 545-2859 (W) > Amherst, MA 01003-4515 > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 7 16:03:40 1999 Date: Wed, 07 Jul 1999 15:57:17 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Re: Jforum: "Pop J" References: <3783793F.F46C493C@math.umass.edu> <3783A375.6E86BA01@balcab.ch> Which version of Acrobat were you using? If it was the new one, 4.0, then I can do the same thing myself. "d.alis" wrote: > > Hi.. > I had the same problems.. but i've generated a new pdf file from the postscript > which printed fine. I can send you a copy if you want. > David > > Murray Eisenberg wrote: > > > "Printing as image" from Acrobat is a workaround -- but hardly a > > solution! The problem is that _text_ printed as image comes out looking > > quite inferior to text printed in the normal way, with many small breaks > > in the letters (the sort of thing an editor would mark up on galley > > proofs of a book or article). > > > > P.S. The "print as image" workaround is the same thing recommended for > > printing .pdf files created from TeX originals that otherwise refuse to > > properly recognize certain symbols; the results are equally, or more, > > unsatisfactory as in the present case. [There the cause is entirely > > different -- Acrobat's inability to deal with characters encoded in > > low-numbered positions.] > > > > Les Redman wrote: > > > > > > The illustrations on pages 16, 17, 19 can be printed by ticking "Print as > > > Image" on the print dialog. > > > > > > > -- > > Murray Eisenberg murray@math.umass.edu > > Mathematics & Statistics Dept. phone 413 549-1020 (H) > > Univ. of Massachusetts 413 545-2859 (W) > > Amherst, MA 01003-4515 > > > > -------------------------------------------------------------------------------- > > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 02:12:45 1999 From: "Randy MacDonald" References: <001101bec81a$9cd40c80$a898dbcf@godin.on.ca> <3782BFB3.56A59649@worldnet.att.net> Subject: Re: Jforum: "Pop J" Date: Thu, 8 Jul 1999 02:08:09 -0400 charset="iso-8859-1" Bambrough: >$159 for an extra 128 mb? When I go to computer shows in New Hampshire I can get >memory for under $200 for 256 MB. Note that that $159 is in Canadian Dollars, at .68 this is around $100USD... -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 05:49:39 1999 From: "Ondrus, Milan" Subject: Jforum: dde Date: Thu, 8 Jul 1999 11:45:37 +0200 X-MS-TNEF-Correlator: <9818339E731AD311AE5C00902715779C0F9324@szrh00313.tszrh.csfb.com> Content-Type: multipart/mixed; boundary="----_=_NextPart_000_01BEC926.A24C0305" This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_000_01BEC926.A24C0305 I get an error in trying that dde example with the hotlink - scripts : server : load'system\packages\dde\server1.ijs' NB. works ok client : load'system\packages\dde\client1.ijs' NB. ok but then if type in in the clients window : cmd'i.4 5' .. its correctly displayed on the servers dialog box but shortly after I get the following error message in the clients session: *************************************** domain error: wd wd'ddepoke jserver top item *',y. *************************************** Is this a real bug or did I overlook a feature in the scripts ? /Milan Ondrus Ps . I am using J 4.03b on an NT 4.0 platform ------_=_NextPart_000_01BEC926.A24C0305 Content-Type: application/ms-tnef Content-Transfer-Encoding: base64 eJ8+IigJAQaQCAAEAAAAAAABAAEAAQeQBgAIAAAA5AQAAAAAAADoAAEIgAcAGAAAAElQTS5NaWNy b3NvZnQgTWFpbC5Ob3RlADEIAQWAAwAOAAAAzwcHAAgACwAtACUABABGAQEggAMADgAAAM8HBwAI AAsALQAmAAQARwEBCYABACEAAAA3NzY5Qjg3MjBGMzVEMzExQUU1RjAwOTAyNzE1Nzc5QwD6BgEE gAEABAAAAGRkZQAtAQENgAQAAgAAAAIAAgABA5AGAGAHAAAtAAAAAwAAgAggBgAAAAAAwAAAAAAA AEYAAAAAUoUAAM4UAAAeAAGACCAGAAAAAADAAAAAAAAARgAAAABUhQAAAQAAAAUAAAA4LjAzAAAA AAMAAoAIIAYAAAAAAMAAAAAAAABGAAAAAAGFAAAAAAAACwADgAggBgAAAAAAwAAAAAAAAEYAAAAA A4UAAAAAAAALAASACCAGAAAAAADAAAAAAAAARgAAAAAOhQAAAAAAAAMABYAIIAYAAAAAAMAAAAAA AABGAAAAABCFAAAAAAAAAwAGgAggBgAAAAAAwAAAAAAAAEYAAAAAEYUAAAAAAAADAAiACCAGAAAA AADAAAAAAAAARgAAAAAYhQAAAAAAAB4ACYAIIAYAAAAAAMAAAAAAAABGAAAAADaFAAABAAAAAQAA AAAAAAAeAAqACCAGAAAAAADAAAAAAAAARgAAAAA3hQAAAQAAAAEAAAAAAAAAHgALgAggBgAAAAAA wAAAAAAAAEYAAAAAOIUAAAEAAAABAAAAAAAAAAIBCRABAAAAPQIAADkCAAB/AwAATFpGdSNlh1gD AAoAcmNwZzEyNRYyAPgLYG4OEDAzM50B9yACpAPjAgBjaArAYHNldDAgBxMCgH0ZCoF1YwBQCwN1 bG4dAiBlC6cKsQqAIEkg+mcRQCADkQSQA2AFwAuA0CB0cnkLgGcVIBEAjQVAZAEAFJB4YW0LUB8W EAPwFaAVkRYQaG90wmwLgGsgLSAE9AQgzjoTpBOlETBydgSQGFACIAkAYWQnc3lzEHRlbVwKoWNr YT8UMBDAAPABABqAGUQxLpBpanMnB7BCLhagpQWwawQgb2sTpWMXcP8J8AVAGb8ayx2UG/kdBhOl fGJ1BUAXAQOgBpAVIHn+cBYQFQEVAhcRHZQEIAPwmG5kbwfgGGxjbR5QIGkuNCA1HEAgLl8cgBbA BCAFoQlwYxdgefcV4AQAC1F5CYAc8CM0GUS3BCAnMAdAbxWABuB4IXnucxdAACAnAWEBgBmBFBT7 FwICEGwJACQRFYAUpAeB3nMa4SMeETAEEGkCIBhs/iovXzBvMLITpSRAAMAVAbsUox4AdwsxE8Mz ECcV8R5wHQAWEBwgGVR0b3CHJkEesC9AJyx5Li7v9za/MQ8YxkkEIBWgBAAUYP4gCXAHQCHRFYAF sScwJ7C9FBBvGXEJAB0AOdFmOhAudAhwLPcX5j8Yey9NywMQA5FPJDBydQQhGIreUAQgHIAUEBZQ ID7QFWLQSiA0Lg9QYifCA5H0TlRA4iALUQAwBbAewC8TtBh6QqsR8QBE4AAAAAsAAgABAAAAHgBw AAEAAAAEAAAAZGRlAAIBcQABAAAAFgAAAAG+ySah6joKlnQ09RHTgCQAUIstYHgAAEAAOQBgzO+h Jsm+AQMA8T8JBAAAHgAxQAEAAAAIAAAATU9ORFJVUwADABpAAAAAAB4AMEABAAAACAAAAE1PTkRS VVMAAwAZQAAAAAADAP0/5AQAAAMAJgAAAAAAAwA2AAAAAAADAIAQ/////wIBRwABAAAAOQAAAGM9 Q0g7YT0gO3A9Q1JFRElULVNVSVNTRTtsPVNaUkgwMDMxMy05OTA3MDgwOTQ1MzdaLTEyMTQwAAAA AAIB+T8BAAAAVAAAAAAAAADcp0DIwEIQGrS5CAArL+GCAQAAAAAAAAAvTz1DUkVESVQtU1VJU1NF L09VPVpVUklDSC1GQjEvQ049UkVDSVBJRU5UUy9DTj1NT05EUlVTAB4A+D8BAAAADgAAAE9uZHJ1 cywgTWlsYW4AAAAeADhAAQAAAAgAAABNT05EUlVTAAIB+z8BAAAAVAAAAAAAAADcp0DIwEIQGrS5 CAArL+GCAQAAAAAAAAAvTz1DUkVESVQtU1VJU1NFL09VPVpVUklDSC1GQjEvQ049UkVDSVBJRU5U Uy9DTj1NT05EUlVTAB4A+j8BAAAADgAAAE9uZHJ1cywgTWlsYW4AAAAeADlAAQAAAAgAAABNT05E UlVTAEAABzDHfUc4Jcm+AUAACDAFA0yiJsm+AR4APQABAAAAAQAAAAAAAAAeAB0OAQAAAAQAAABk ZGUAHgA1EAEAAABCAAAAPDk4MTgzMzlFNzMxQUQzMTFBRTVDMDA5MDI3MTU3NzlDMEY5MzI0QHN6 cmgwMDMxMy50c3pyaC5jc2ZiLmNvbT4AAAALACkAAAAAAAsAIwAAAAAAAwAGEJ43cDoDAAcQ9AEA AAMAEBAAAAAAAwAREAAAAAAeAAgQAQAAAGUAAABJR0VUQU5FUlJPUklOVFJZSU5HVEhBVERERUVY QU1QTEVXSVRIVEhFSE9UTElOSy1TQ1JJUFRTOlNFUlZFUjpMT0FEU1lTVEVNUEFDS0FHRVNEREVT RVJWRVIxSUpTTkJXT1JLAAAAAAIBfwABAAAAQgAAADw5ODE4MzM5RTczMUFEMzExQUU1QzAwOTAy NzE1Nzc5QzBGOTMyNEBzenJoMDAzMTMudHN6cmguY3NmYi5jb20+AAAAzoA= ------_=_NextPart_000_01BEC926.A24C0305-- -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 07:02:17 1999 From: "Chris Burke" References: <9818339E731AD311AE5C00902715779C0F9324@szrh00313.tszrh.csfb.com> Subject: Re: Jforum: dde Date: Thu, 8 Jul 1999 06:36:07 -0400 charset="iso-8859-1" This works fine for me in Win98. What did you enter before getting that error message? It looks like an invalid argument to cmd, which is defined as: cmd 3 : 'wd ''ddepoke jserver top item *'',y.' For example: cmd 1 2 3 domain error: cmd wd'ddepoke jserver top item *' ,y. What is the value of y. when you get the failure? > I get an error in trying that dde example with the hotlink - scripts : > > server : load'system\packages\dde\server1.ijs' NB. works ok > client : load'system\packages\dde\client1.ijs' NB. ok > > but then if type in in the clients window : > > cmd'i.4 5' .. its correctly displayed on the servers dialog box > but shortly after I get the following error message in the clients session: > > *************************************** > domain error: wd > wd'ddepoke jserver top item *',y. > *************************************** > > Is this a real bug or did I overlook a feature in the scripts ? > > /Milan Ondrus > > Ps . I am using J 4.03b on an NT 4.0 platform -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 07:37:08 1999 From: "Chris Burke" References: <9818339E731AD311AE5C00902715779C0F9324@szrh00313.tszrh.csfb.com> <00c901bec931$78e051e0$1f0114d1@Pcdburke> Subject: Re: Jforum: dde Date: Thu, 8 Jul 1999 07:38:13 -0400 charset="iso-8859-1" Careless of me. Please disregard the message below, which was a draft I thought I had discarded. I now suspect there is a problem with DDE under NT. ----- Original Message ----- From: Chris Burke Sent: Thursday, July 08, 1999 6:36 AM Subject: Re: Jforum: dde > This works fine for me in Win98. > > What did you enter before getting that error message? It looks like an > invalid argument to cmd, which is defined as: > > cmd > 3 : 'wd ''ddepoke jserver top item *'',y.' -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 11:54:43 1999 From: "Ondrus, Milan" Subject: Jforum: RE: MAPI/11 Date: Thu, 8 Jul 1999 17:44:19 +0200 X-MS-TNEF-Correlator: <9818339E731AD311AE5C00902715779C0F936E@szrh00313.tszrh.csfb.com> Content-Type: multipart/mixed; boundary="----_=_NextPart_000_01BEC958.BE629F62" This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_000_01BEC958.BE629F62 Forgot to say again: I'm using J 4.03b under NT 4.0. I use Excel 97 and Outlook-version 8.04. / > -----Original Message----- > From: Ondrus, Milan > Sent: Donnerstag, 8. Juli 1999 17:32 > To: 'jforum' > Subject: MAPI/11 > > > > There seems to be an annoying problem with sending attachments. > > All that I can say is the following: > > Despite the fact that the Excel-File I want to send and is existing I get > very often this DLL-error number > 11 back - which says that MAPI_E_ATTACHMENT_NOT_FOUND ..(see MSDN) as > Seymour Glass said before. > > Unfortunately I did not succeed in finding a logical rule(except maybe > the one below) what the conditions are when this happens. > Sometimes all files are sent ,sometimes only one. > > This code 11 is correct when the file does not exist ,fine,but there must > be an other condition when 11 is displayed > but the file IS there. > This behavior I found on both sorts of drives ,eg when the files are > sitting on the hard-drive or a network drive. > > Just one important trace I found that might me a hint. > > I have a process that creates excelfiles out of textfiles. > Lets say I create : > > file1.xls > file2.xls > file3.xls > > After the creation there is no Excel task in the background and no locks > are obvious. > > If I process them with the mapi-stuff in the opposite order ,then > everything is sent. > If I resent them all - then only the 1st is sent (file1.xls) - the rest > comes back with 11. > Only if I logoff from J and go back and reprocess it in the opposite > order again - all are sent but as before > resending them means again : only the 1st is sent. > > If I send them in the same order - then allways only the 1st one is sent > ,the rest comes back with error 11 > (Smashing down the J session does not help) > > One other important point.If I create a file directly under Excel ,this > behaviour I did not perceive.So there it looks all right. > > So ,from this point of view I can't even say that there is sth wrong with > the MAPI script ,but as it looks - there seems to be > a problem with mapi when I automatically create a spread-sheet - but this > creation of the excel-file itself works perfectly: > > fn =: 3 : 0 > > xlopen'' > xlcmd'wb add' > wbname=:xlget'temp name' > xlget'temp worksheets' > xlid'ws' > xlget'ws item 1' > wsname=:xlget'temp name' > y. xlwriter wbname,' ',wsname,' 1 1' > xlcmd 'temp saveas ',' xyzhhmmss.xls' > xlquit'' > ) > > As this is happening I can't release it into production. > > thanks for any help ...........Milan Ondrus > > > > > > > > > > > > > > > ------_=_NextPart_000_01BEC958.BE629F62 Content-Type: application/ms-tnef Content-Transfer-Encoding: base64 eJ8+IhUPAQaQCAAEAAAAAAABAAEAAQeQBgAIAAAA5AQAAAAAAADoAAEIgAcAGAAAAElQTS5NaWNy b3NvZnQgTWFpbC5Ob3RlADEIAQWAAwAOAAAAzwcHAAgAEQAsABMABAA5AQEggAMADgAAAM8HBwAI ABEALAAUAAQAOgEBCYABACEAAABCMDc3Qjg3MjBGMzVEMzExQUU1RjAwOTAyNzE1Nzc5QwD9BgEE gAEADAAAAFJFOiBNQVBJLzExAKkCAQ2ABAACAAAAAgACAAEDkAYA6AwAAC8AAAADAACACCAGAAAA AADAAAAAAAAARgAAAABShQAAzhQAAB4AAYAIIAYAAAAAAMAAAAAAAABGAAAAAFSFAAABAAAABQAA ADguMDMAAAAAAwACgAggBgAAAAAAwAAAAAAAAEYAAAAAAYUAAAAAAAALAAOACCAGAAAAAADAAAAA AAAARgAAAAADhQAAAAAAAAsABIAIIAYAAAAAAMAAAAAAAABGAAAAAA6FAAAAAAAAAwAFgAggBgAA AAAAwAAAAAAAAEYAAAAAEIUAAAAAAAADAAaACCAGAAAAAADAAAAAAAAARgAAAAARhQAAAAAAAAMA CIAIIAYAAAAAAMAAAAAAAABGAAAAABiFAAAAAAAAHgAJgAggBgAAAAAAwAAAAAAAAEYAAAAANoUA AAEAAAABAAAAAAAAAB4ACoAIIAYAAAAAAMAAAAAAAABGAAAAADeFAAABAAAAAQAAAAAAAAAeAAuA CCAGAAAAAADAAAAAAAAARgAAAAA4hQAAAQAAAAEAAAAAAAAAAgEJEAEAAABeBwAAWgcAALIMAABM WkZ1tEMF1gMACgByY3BnMTI1FjIA+Atgbg4QMDMznQH3IAKkA+MCAGNoCsDgc2V0MCAHEwKDAFCh EHZwcnEyEXZ9CoDZCMggOwlvDjA1AoAKgSx1YwBQCwNjAEF1bHpuAiBlC6cKsQqECzBsKGkzNhFg RgWwZ28BBUB0byBzYXkgdGFnC3E6CuMKhAqASTAnbSB1AJAPICBKqCA0Lg9QYhsQbgSBNQewVBuS LhqFGxFlIJBFeGNlAyA5NxmgIRwQIE91dAkAb2ucLXYEkACQAiAgOBuw+xugF7QvGioXtRaQF6ED YAp0BZB0EgQxNiAt9SKSTwUQZwuAB0AF0AeQ+RlwZ2Uikxe2IaQXowsxwSGmaS0xNDQBQBhweDE4 MAFADNAmMxvgRlUDYToMg2IRYE8cEHL9GyAsBdADEAOSF8Qb4AZgewIwJ8dEAiAXEBEgAZBnDyiw H2AbcBbQaSAxOWMr4CvANzozDlApRlQybyfHJ2oCECiAbScxKTh1Ymoh0SfHTUHwUEkvMRYjJH8h lhhz7xc/F+EaFSzwaASQHXARMOxlbQQgGUFiHXADkQBwdRbweRtCICGRAmA0kCC5A/B0aDRhHBAb QmECQG0A0GgHgAIwcxzFMzVBrmwDIDagN1AgHTBjA5G/GXIEADkRHXACEDjwbwPw1w8gJ8Ay6kQH kHA2kB1wPzpDANAZITkyOkIdky1G9wMQHXAdMHcAcBkkNuEeE+k6EWV4BAB0G0IdMCPA9wVAHuEZ kG8BgAnwORE6EXBETEwtBJADYAXAbn8uADUACtQzcTAQNPAA0Gv/IoA2cEFQEPAZYjoiOUEvwgBf RV9BVFRBQ4hITUUcYF9OT0WwAEZPVU5EIC4uRig0cQXRRE4pNeBh5wZCBsAIYSBHC2AEERlwLmke QDUALdFlOAxVbv0t0XQcADdQHcAZkB0wNwD3HkAW8AVAcxYAHbAJgDoA9wOgJgA29SAJACMQOZAD IHsogDZAKD+gHbAFMQDAef81ATpCFwE08TqxRyFDsD0l/wWgNvE/4AIgBCAKwB1wQ7DzQRYRAHBw CfA39gYAA3D/EUAHcwdAAyAmADZAUIQ20fkFQCxzUocCIEqhFwE4DP80EDoRBaABAEMCVnNB4CHR 51DWOmE+AmRvB5FLMj+j41QwTCFlLGIecDoyNEH+bRsgBUA1BBkQNCFP6FDU71b0NwA8EAtgeQmA MzVaBf1YVEkF8FpDUfZWUzUAEQD+dh8gBcAdMAIQHAFOkTTw71tRGWAJEVTRZkrgBRAe4PkEICxl G2BX21OVNpA/4+cfMTpCEQFkLWJDQNAFwPNMoBcQdHcFsENwYkM4DG8rgFlxTqIHcHAJET5zcq8A 0D4iYNQ5I20jAGhN4fc1EVFwC4B0OAwdMGBBapLfIZEdsAQROSMFAGVKcT+C/x2xU1QIYGfxYiAZ MD+gADB7U2JR9kwRQEhCSrJtVCDDOxxTUjEueGwQwHG4+jJybTNyaDh2QPEFwE+zf21iHyJaRDoR FvAdhQGQc/9DcEvxOkJDQgnAYOMeIncx/wkAQ2BQhDYgYGEbIGsdYiC/HTBsiDZWOkIAwDwgLSrw 1nUBIHgGb1Ggb2QhZaLtHCIsOkEDoGVAkkFBG1H/OhFT8lH2e5MJcFPzfGNTEm9DkH9jVPM6QjFZ cYBlIP4ochdHIIKjgZJtMVKBYAH/Q1I2gzAQUfYoUEqhBpA5Yf9MwUDgYiADUhtxHiIZAEM0/x4i CXBshjaQfe9+9BmzQ4H/UxJTt1oCR1FItDM1gaM3A/98YweABiKMJBoAgxuAXXsq/z7jfGN4FRlw aoF+9IKlUxH/PmBEMYMbaBOD9X9ShX82g9dB1DAWhFBTAMBzgBNYsP53WAQbgBEwBBAfIli3NCD4 bHApMtsoUE6BW2NoWL9ocGrie5NwxUygWGRpV4L/SqEcBB2Uf1Ff+EfRSthRsP5yHbBmslJwdpYF QB6SUvT/IvFqUFH2UgdZkYiSQUOek+9iAmBgB9E5cydZIR7gObTvPQZ21CrwNrB3A2AbUXy3/y/C GWAFBFnzR1GkN4KjNE2/MzVsYzY4fUJQ1B0wYR5w/wNxP+BNAUqhn0c8EG1hZUD/mhAJ4AVAQ5Ba BFZydiZiEfs6Qm3TLVhjNpARMG4QNnB3ZlGnAQSQZqBTcT0DoD39GgAzkFEXpTOWcnB+YAnwBicu JbixY21kJ3f/G+CyALnwMzW6ECMwB4C3YO9ycEBRqFA0kHBCIJQhuTj3u5e1Y7JCc7k4SIC6AL4p r7uSvwCKgTZRMbp3c7sf/bwqeStgcnCqMDwxBcC69PQsJy2gLMEUxCFDIMCH/7mzLaDB8xlwHuBH UcRQxDDweHl6aDegNKA38HJxebk4cXU2kLkmnHUzlkH/OiI6EVFXP/SoJAlwNkBHUO+kE2rhGVAh kWQWAFBCOAz9OSFuedEt0R4RGZCcMkZh/8/XKNQoVDLf0Y/TDjXg0h8v1M/UJBobE4EA2BAAAAMA LgAAAAAACwACAAEAAAAeAHAAAQAAAAgAAABNQVBJLzExAAIBcQABAAAAGwAAAAG+yVcHlpwBor41 PRHTgCQAUIstYHgAAFqdMAAeAEIQAQAAAEIAAAA8OTgxODMzOUU3MzFBRDMxMUFFNUMwMDkwMjcx NTc3OUMwRjkzNkRAc3pyaDAwMzEzLnRzenJoLmNzZmIuY29tPgAAAEAAOQAAHRq+WMm+AQMA8T8J BAAAHgAxQAEAAAAIAAAATU9ORFJVUwADABpAAAAAAB4AMEABAAAACAAAAE1PTkRSVVMAAwAZQAAA AAADAP0/5AQAAAMAJgAAAAAAAwA2AAAAAAADAIAQ/////wIBRwABAAAAOQAAAGM9Q0g7YT0gO3A9 Q1JFRElULVNVSVNTRTtsPVNaUkgwMDMxMy05OTA3MDgxNTQ0MTlaLTEyNzM4AAAAAAIB+T8BAAAA VAAAAAAAAADcp0DIwEIQGrS5CAArL+GCAQAAAAAAAAAvTz1DUkVESVQtU1VJU1NFL09VPVpVUklD SC1GQjEvQ049UkVDSVBJRU5UUy9DTj1NT05EUlVTAB4A+D8BAAAADgAAAE9uZHJ1cywgTWlsYW4A AAAeADhAAQAAAAgAAABNT05EUlVTAAIB+z8BAAAAVAAAAAAAAADcp0DIwEIQGrS5CAArL+GCAQAA AAAAAAAvTz1DUkVESVQtU1VJU1NFL09VPVpVUklDSC1GQjEvQ049UkVDSVBJRU5UUy9DTj1NT05E UlVTAB4A+j8BAAAADgAAAE9uZHJ1cywgTWlsYW4AAAAeADlAAQAAAAgAAABNT05EUlVTAEAABzAH yhe+WMm+AUAACDBin2K+WMm+AR4APQABAAAABQAAAFJFOiAAAAAAHgAdDgEAAAAIAAAATUFQSS8x MQAeADUQAQAAAEIAAAA8OTgxODMzOUU3MzFBRDMxMUFFNUMwMDkwMjcxNTc3OUMwRjkzNkVAc3py aDAwMzEzLnRzenJoLmNzZmIuY29tPgAAAAsAKQAAAAAACwAjAAAAAAADAAYQtMgQlwMABxBkBwAA AwAQEAAAAAADABEQDwAAAB4ACBABAAAAZQAAAEZPUkdPVFRPU0FZQUdBSU46SU1VU0lOR0o0MDNC VU5ERVJOVDQwSVVTRUVYQ0VMOTdBTkRPVVRMT09LLVZFUlNJT044MDQvLS0tLS1PUklHSU5BTE1F U1NBR0UtLS0tLUZST00AAAAAAgF/AAEAAABCAAAAPDk4MTgzMzlFNzMxQUQzMTFBRTVDMDA5MDI3 MTU3NzlDMEY5MzZFQHN6cmgwMDMxMy50c3pyaC5jc2ZiLmNvbT4AAAD7ew== ------_=_NextPart_000_01BEC958.BE629F62-- -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 11:56:45 1999 From: "Ondrus, Milan" Subject: Jforum: MAPI/11 Date: Thu, 8 Jul 1999 17:32:03 +0200 X-MS-TNEF-Correlator: <9818339E731AD311AE5C00902715779C0F936D@szrh00313.tszrh.csfb.com> Content-Type: multipart/mixed; boundary="----_=_NextPart_000_01BEC957.0886082C" This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_000_01BEC957.0886082C There seems to be an annoying problem with sending attachments. All that I can say is the following: Despite the fact that the Excel-File I want to send and is existing I get very often this DLL-error number 11 back - which says that MAPI_E_ATTACHMENT_NOT_FOUND ..(see MSDN) as Seymour Glass said before. Unfortunately I did not succeed in finding a logical rule(except maybe the one below) what the conditions are when this happens. Sometimes all files are sent ,sometimes only one. This code 11 is correct when the file does not exist ,fine,but there must be an other condition when 11 is displayed but the file IS there. This behavior I found on both sorts of drives ,eg when the files are sitting on the hard-drive or a network drive. Just one important trace I found that might me a hint. I have a process that creates excelfiles out of textfiles. Lets say I create : file1.xls file2.xls file3.xls After the creation there is no Excel task in the background and no locks are obvious. If I process them with the mapi-stuff in the opposite order ,then everything is sent. If I resent them all - then only the 1st is sent (file1.xls) - the rest comes back with 11. Only if I logoff from J and go back and reprocess it in the opposite order again - all are sent but as before resending them means again : only the 1st is sent. If I send them in the same order - then allways only the 1st one is sent ,the rest comes back with error 11 (Smashing down the J session does not help) One other important point.If I create a file directly under Excel ,this behaviour I did not perceive.So there it looks all right. So ,from this point of view I can't even say that there is sth wrong with the MAPI script ,but as it looks - there seems to be a problem with mapi when I automatically create a spread-sheet - but this creation of the excel-file itself works perfectly: fn =: 3 : 0 xlopen'' xlcmd'wb add' wbname=:xlget'temp name' xlget'temp worksheets' xlid'ws' xlget'ws item 1' wsname=:xlget'temp name' y. xlwriter wbname,' ',wsname,' 1 1' xlcmd 'temp saveas ',' xyzhhmmss.xls' xlquit'' ) As this is happening I can't release it into production. thanks for any help ...........Milan Ondrus ------_=_NextPart_000_01BEC957.0886082C Content-Type: application/ms-tnef Content-Transfer-Encoding: base64 eJ8+IgcPAQaQCAAEAAAAAAABAAEAAQeQBgAIAAAA5AQAAAAAAADoAAEIgAcAGAAAAElQTS5NaWNy b3NvZnQgTWFpbC5Ob3RlADEIAQWAAwAOAAAAzwcHAAgAEQAgAAMABAAdAQEggAMADgAAAM8HBwAI ABEAIAAFAAQAHwEBCYABACEAAAAyNzc2Qjg3MjBGMzVEMzExQUU1RjAwOTAyNzE1Nzc5QwDzBgEE gAEACAAAAE1BUEkvMTEAuAEBDYAEAAIAAAACAAIAAQOQBgBgCwAALQAAAAIBcQABAAAAFgAAAAG+ yVcHlpwBor41PRHTgCQAUIstYHgAAAMAAIAIIAYAAAAAAMAAAAAAAABGAAAAAFKFAADOFAAAHgAB gAggBgAAAAAAwAAAAAAAAEYAAAAAVIUAAAEAAAAFAAAAOC4wMwAAAAADAAKACCAGAAAAAADAAAAA AAAARgAAAAABhQAAAAAAAAsAA4AIIAYAAAAAAMAAAAAAAABGAAAAAAOFAAAAAAAACwAEgAggBgAA AAAAwAAAAAAAAEYAAAAADoUAAAAAAAADAAWACCAGAAAAAADAAAAAAAAARgAAAAAQhQAAAAAAAAMA BoAIIAYAAAAAAMAAAAAAAABGAAAAABGFAAAAAAAAAwAIgAggBgAAAAAAwAAAAAAAAEYAAAAAGIUA AAAAAAAeAAmACCAGAAAAAADAAAAAAAAARgAAAAA2hQAAAQAAAAEAAAAAAAAAHgAKgAggBgAAAAAA wAAAAAAAAEYAAAAAN4UAAAEAAAABAAAAAAAAAB4AC4AIIAYAAAAAAMAAAAAAAABGAAAAADiFAAAB AAAAAQAAAAAAAAACAQkQAQAAADYGAAAyBgAAuQoAAExaRnUfK3X3AwAKAHJjcGcxMjUWMgD4C2Bu DhAwMzOdAfcgAqQD4wIAY2gKwGBzZXQwIAcTAoB9GQqBdWMAUAsDdWxufQIgZQunCrEKgArjE+FU kmgEkGUgETBlbQQg8HRvIGIUwAORAHAS4MJ5C4BnICBwA2ACYCUVACAD8HRoFNFuZG0WImECQADQ aAeAAjBz5i4TpBOlQWwDIBcQF8AQIEkgYwORc2F5viAEABmBFMACEBlgbwPw1Q8gOhiLRAeQcBcA FMAfGrMA0AVAGZMaskV4Y3BlbC1GAxAUwBngd38AcB1RFVAXQhWRH4AagWVaeAQAdBYiGeBnEUAg 9nYEkBpgbwGACfAZgRqBcERMTC0EkANgBcBuznUG0ASQFAYxMRVgAND4ayAtFuAhwBDwGjIakgEZ sU1BUElfRV8AQVRUQUNITUUgTlRfTk8mIEZPgFVORCAuLigU4dEF0UROKRZQYQZCBsC5CGEgRwtg BBEaQGkfgE8VcAIQCXAYfFVuKUF0/HVuF8AeMBpgGeAXcB+A+xLgBUBzEjAeIAmAGnADoNZmC4AX dCAJAGcN4AdA5CByEsBlKCAQHiAFMf0AwHkVcRqyEvEVYRshJ5H/JCAdlQWgF2EgUAIgBCAKwM8U wCQgIYYRAHBwCfAYZv8GAANwEUAHcwdAAyAskBaw5zD0F0EFQCxzMvcCICsR/xLxGHwUgBqBBaAB ACNyNuM/IlAFkAVAMVUa0R5yZG8fB5EroiATNKAskWUsYuZ1HaMUsW11OeEVdCuwfxSRMFgxRDdk F3AcgAtgee8JgBOlOnU4xEkF8DqzMmb3NsMVcBEAdjDABcAZ4AIQfyrAH4A8kQbgFxIJETVBZs8r UAUQIQAEICxlFkA4Sz80BRcAIFM8kRqyEQFkLZ9CsyFABcAtEBMAdHcFsHcj4EKzGHxKOyIvEgdw cHsJER7jcgDQHpJBRBmTbXhpZ2guURWBMeALgHT/GHwZ4ECxSwIWcR4gBBEZk/0FAGUq4R/yHiEz xAhgSGG/QpAVMCAQADAz0jJmTBFANyiyKyJNxCAbjDPCMS6seGwQwFIoMlLdM1LY/xjmIWEFwDAj TdI8gjq0GoH/EuAd9BUwJ8Aj4CxhGrIjsv8JwEFTH6JXoQkAI9Aw9BaQ/0DROyBLjUKQGeBM+BbG GrJrAMAckC0gQHUBIFh2b/0yEG9EkUYSBIE0oBqxA6D+ZSECIbEWMRqBNGIyZlwD/wlwNGNc0zOC JABf0zVjGrLmMTnhYNUgKFKHJ5BjE/9iAk2hMvFAcSPCFvMjgDJm/k81cgaQGdEtMSFQQpADUv9I IB+TaKAjpB+iCXBM9hcA815fX2RhZwtxI/EzgjQn/zpyJ8EpJBOlYhMXc1zTB4D7BiJslDpjfGDN W5ofU1zT/1iFGkBK8V9kYxUzgR7QJKH/Y4tIg2RlX8Jl7xbzIkQjgGkTpShTAMBzYIM5IHd/OHRp QBEwBBA8gjknFJBs/HApGItn4C7xO9NIyEjg30tSXANRNS0QONRpN/IrEf9BYSLhV9RfwUBoKEEr SDIgfnIeIEciMuBXBgVACQBv/1pCGWEFEEqxMmYydzoBaQK/IbN/A0JyQNAH0RnjJzmRvyEAGiQd dldEIEAXIHcDYP8WMV0nJSIU0AUEOmMnwYSn/2MTFL0TpUzTFqhdsjFEGeD+YTqAA3EgUC1xKxF/ txyA/03RRbB6gAngBUAkADp0NuLvVpZCgRqyTkMtONMXABEw306AFuBGwYdxBJBmgMNRrfUDoD1w 0DNwwROdUuBe0AkJ8CcnmMdjbWQn9HdiFZBkmmATpZqAKtC/B4CX0FLgIMGIwBUAcCKQ33SRmaic B5XTkrJzmago8L+acJ6ZnAKfcGrxFsExmufmc5uPnJp5LpkhiqAcoYMFwJtkLCcgJyyhhP+kkSOQ oPeaI6SwomMaQCEAhyfBpMCkoHh5emgYEM8VEBhgUuGZqHF1FwCZlvt85RQGQRqSGoExxyBkiJR/ CXAWsCfAhINLURVQFnFk7xIwMLIYfBmRblpBKUEVke8aYHyiJtGwR00DEAORZ+D/QrA7IBiLsb8Y xhiLsn+1Lxe0hbW7EfEAuIAAAAsAAgABAAAAHgBwAAEAAAAIAAAATUFQSS8xMQADAP0/5AQAAEAA OQDQPJwHV8m+AQMA8T8JBAAAHgAxQAEAAAAIAAAATU9ORFJVUwADABpAAAAAAB4AMEABAAAACAAA AE1PTkRSVVMAAwAZQAAAAAADACYAAAAAAAMANgAAAAAAAwCAEP////8CAUcAAQAAADkAAABjPUNI O2E9IDtwPUNSRURJVC1TVUlTU0U7bD1TWlJIMDAzMTMtOTkwNzA4MTUzMjAzWi0xMjcyNAAAAAAC Afk/AQAAAFQAAAAAAAAA3KdAyMBCEBq0uQgAKy/hggEAAAAAAAAAL089Q1JFRElULVNVSVNTRS9P VT1aVVJJQ0gtRkIxL0NOPVJFQ0lQSUVOVFMvQ049TU9ORFJVUwAeAPg/AQAAAA4AAABPbmRydXMs IE1pbGFuAAAAHgA4QAEAAAAIAAAATU9ORFJVUwACAfs/AQAAAFQAAAAAAAAA3KdAyMBCEBq0uQgA Ky/hggEAAAAAAAAAL089Q1JFRElULVNVSVNTRS9PVT1aVVJJQ0gtRkIxL0NOPVJFQ0lQSUVOVFMv Q049TU9ORFJVUwAeAPo/AQAAAA4AAABPbmRydXMsIE1pbGFuAAAAHgA5QAEAAAAIAAAATU9ORFJV UwBAAAcw1PuZB1fJvgFAAAgwLAiGCFfJvgEeAD0AAQAAAAEAAAAAAAAAHgAdDgEAAAAIAAAATUFQ SS8xMQAeADUQAQAAAEIAAAA8OTgxODMzOUU3MzFBRDMxMUFFNUMwMDkwMjcxNTc3OUMwRjkzNkRA c3pyaDAwMzEzLnRzenJoLmNzZmIuY29tPgAAAAsAKQAAAAAACwAjAAAAAAADAAYQmjfu8gMABxC9 BgAAAwAQEAAAAAADABEQAAAAAB4ACBABAAAAZQAAAFRIRVJFU0VFTVNUT0JFQU5BTk5PWUlOR1BS T0JMRU1XSVRIU0VORElOR0FUVEFDSE1FTlRTQUxMVEhBVElDQU5TQVlJU1RIRUZPTExPV0lORzpE RVNQSVRFVEhFRkFDVFRIQVQAAAAAAgF/AAEAAABCAAAAPDk4MTgzMzlFNzMxQUQzMTFBRTVDMDA5 MDI3MTU3NzlDMEY5MzZEQHN6cmgwMDMxMy50c3pyaC5jc2ZiLmNvbT4AAACjDw== ------_=_NextPart_000_01BEC957.0886082C-- -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 11:59:10 1999 Date: Thu, 8 Jul 1999 11:51:44 -0400 From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= Subject: Jforum: OCX's charset=ISO-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id LAA28520 I am sending this note to find out who is using OCX's. Am I the only one? I use FarPoint's Spread, Calendar, and Input controls. I also use Olectra Chart, and XceedZip. What do you readers of this message use? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 12:58:42 1999 From: Andromeda Subject: Jforum: FW: MAPI/11 Date: Thu, 8 Jul 1999 18:54:05 +0200 Having thought twice about it when walking home - I'm more and more convinced that the MAPI stuff is ok - but it looks like a stack-problem somewhere when creating Excel-Spreadsheets the way described below (--> fn) By the way : each file below (file1.xls,file2.xls,file3.xls) should be sent different recepients. This is then done by: sess =: logonmail '' NB... send the 3 mails logoffmail sess //..sorry about sending the mail a 3rd time -----Original Message----- From: Ondrus, Milan [SMTP:milan.ondrus@csfb.com] Sent: Donnerstag, 8. Juli 1999 17:44 To: 'jforum' Subject: RE: MAPI/11 Forgot to say again: I'm using J 4.03b under NT 4.0. I use Excel 97 and Outlook-version 8.04. / -----Original Message----- From: Ondrus, Milan Sent: Donnerstag, 8. Juli 1999 17:32 To: 'jforum' Subject: MAPI/11 There seems to be an annoying problem with sending attachments. All that I can say is the following: Despite the fact that the Excel-File I want to send and is existing I get very often this DLL-error number 11 back - which says that MAPI_E_ATTACHMENT_NOT_FOUND ..(see MSDN) as Seymour Glass said before. Unfortunately I did not succeed in finding a logical rule(except maybe the one below) what the conditions are when this happens. Sometimes all files are sent ,sometimes only one. This code 11 is correct when the file does not exist ,fine,but there must be an other condition when 11 is displayed but the file IS there. This behavior I found on both sorts of drives ,eg when the files are sitting on the hard-drive or a network drive. Just one important trace I found that might me a hint. I have a process that creates excelfiles out of textfiles. Lets say I create : file1.xls file2.xls file3.xls After the creation there is no Excel task in the background and no locks are obvious. If I process them with the mapi-stuff in the opposite order ,then everything is sent. If I resent them all - then only the 1st is sent (file1.xls) - the rest comes back with 11. Only if I logoff from J and go back and reprocess it in the opposite order again - all are sent but as before resending them means again : only the 1st is sent. If I send them in the same order - then allways only the 1st one is sent ,the rest comes back with error 11 (Smashing down the J session does not help) One other important point.If I create a file directly under Excel ,this behaviour I did not perceive.So there it looks all right. So ,from this point of view I can't even say that there is sth wrong with the MAPI script ,but as it looks - there seems to be a problem with mapi when I automatically create a spread-sheet - but this creation of the excel-file itself works perfectly: fn =: 3 : 0 xlopen'' xlcmd'wb add' wbname=:xlget'temp name' xlget'temp worksheets' xlid'ws' xlget'ws item 1' wsname=:xlget'temp name' y. xlwriter wbname,' ',wsname,' 1 1' xlcmd 'temp saveas ',' xyzhhmmss.xls' xlquit'' ) As this is happening I can't release it into production. thanks for any help ...........Milan Ondrus -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 13:23:46 1999 From: Baker Stuart Subject: Jforum: Excel OLE: jsutil.xls wont load Date: Thu, 8 Jul 1999 15:21:48 -0000 ... just that, really: I get a message 'file format not valid' when I try: anyone else had this problem? Version info: j4.02a; jsutil.xls 23/09/98 20:53; Excel for Windows 95 vsn 7.0a; Windows NT 4.00.1381 Stuart Baker Hyder IT - Cardiff, UK Stuart.Baker@Hyder.com Tel +44 (0) 1222 444205 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 13:52:40 1999 From: lewis@cdcna.com Subject: RE: Jforum: Excel OLE: jsutil.xls wont load Date: Thu, 8 Jul 1999 13:49:12 -0400 I did have the same problem several months ago and found that it was due to the fact that jsutil.xls was an excel 97 workbook and could not be read by excel 7.0. My temporary fix was to use the jsutil.xls from j3.02, but I have subsequently upgraded. Mike Lewis ---------- From: Baker Stuart [SMTP:stuart.baker@hyder.com] Sent: Thursday, July 08, 1999 11:22 AM To: 'forum@jsoftware.com' Subject: Jforum: Excel OLE: jsutil.xls wont load ... just that, really: I get a message 'file format not valid' when I try: anyone else had this problem? Version info: j4.02a; jsutil.xls 23/09/98 20:53; Excel for Windows 95 vsn 7.0a; Windows NT 4.00.1381 Stuart Baker Hyder IT - Cardiff, UK Stuart.Baker@Hyder.com Tel +44 (0) 1222 444205 ---------------------------------------------------------------------------- ---- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 8 14:58:54 1999 Date: Thu, 08 Jul 1999 14:54:37 -0400 From: William Kindree Organization: Interlog Internet Services Subject: Jforum: Formal Description of J ? References: Anyone know of work being done on a formal description of J, along the lines of similar works that have been done for APL? Bill Kindree -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From ljdickey Thu Jul 8 16:05:51 1999 Subject: Re: Jforum: Formal Description of J ? Roger Hui has written a book on it. > Date: Thu, 08 Jul 1999 14:54:37 -0400 > From: William Kindree > To: forum@jsoftware.com > Subject: Jforum: Formal Description of J ? > > Anyone know of work being done on a formal description of J, along the lines of > similar works that have been done for APL? > > Bill Kindree > > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > From owner-jsoftware@lists.interlog.com Fri Jul 9 05:24:59 1999 From: Baker Stuart Subject: RE: Jforum: Excel OLE: jsutil.xls wont load Date: Fri, 9 Jul 1999 09:58:14 -0000 Thanks very much! I have managed to load successfully with the vsn 3.05 file: I notice the 4.02 vsn is a different size: does anyone know whether the contents have changed, or is that just the effect of being saved from a newer Excel vsn? > ---------- > From: lewis@cdcna.com[SMTP:lewis@cdcna.com] > Reply To: forum@jsoftware.com > Sent: 08 July 1999 17:49 > To: forum@jsoftware.com; stuart.baker@hyder.com > Subject: RE: Jforum: Excel OLE: jsutil.xls wont load > > I did have the same problem several months ago and found that it was due > to > the fact that jsutil.xls was an excel 97 workbook and could not be read by > excel 7.0. My temporary fix was to use the jsutil.xls from j3.02, but I > have subsequently upgraded. > > Mike Lewis > > ---------- > From: Baker Stuart [SMTP:stuart.baker@hyder.com] > Sent: Thursday, July 08, 1999 11:22 AM > To: 'forum@jsoftware.com' > Subject: Jforum: Excel OLE: jsutil.xls wont load > > ... just that, really: > > I get a message 'file format not valid' when I try: > anyone else had this problem? > > Version info: > j4.02a; > jsutil.xls 23/09/98 20:53; > Excel for Windows 95 vsn 7.0a; > Windows NT 4.00.1381 > > > Stuart Baker > > Hyder IT - Cardiff, UK > > Stuart.Baker@Hyder.com > Tel +44 (0) 1222 444205 > > > -------------------------------------------------------------------------- > -- > ---- > J Forum: for information about this list, see > http://www.jsoftware.com/forum.htm > > -------------------------------------------------------------------------- > ------ > J Forum: for information about this list, see > http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 9 13:18:08 1999 From: "Oleg Kobchenko" References: <199907081151_MC2-7C58-B3F3@compuserve.com> Subject: Re: Jforum: OCX's Date: Fri, 9 Jul 1999 20:12:16 +0300 charset="iso-8859-1" Content-Transfer-Encoding: 8bit I was fascinated by the VB6.0 new data connection/display facilities, esp. ADO Data Control and the Data Grid. However, VB utilizes special design-time techniques for mutual ActiveX controls awareness. E.g. one can be inserted into another (like tool bars) -- for the a Container Collection should be supported by the control, or just presence on the same form makes it possible to refer to another object. Some of these relations are made with the Control Dialog. The problem was how to make use of this technology in non-VB environments, like J. In particular I failed to replicate the simple link between ADO Data Control and the Data Grid on a J form. ----- Original Message ----- From: Alain Miville de Ch�ne Sent: Thursday, July 08, 1999 6:51 PM Subject: Jforum: OCX's > I am sending this note to find out who is using OCX's. Am I the only one? > > I use FarPoint's Spread, Calendar, and Input controls. I also use Olectra > Chart, and XceedZip. > > What do you readers of this message use? > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sat Jul 10 19:48:29 1999 Date: Sat, 10 Jul 1999 18:55:43 -0400 From: Daniel Torres Subject: Jforum: Runtime versus Locked files Content-Type: multipart/mixed; boundary="------------735774E5C346C3564390DD18" This is a multi-part message in MIME format. --------------735774E5C346C3564390DD18 I developed a short script to solve a math problem, but someone else is interested in having it. I want to give it to them, but I do not want them to know how the script works. I would have sent them the script in a runtime version if I hadn't come across the locked file option. I read in the documentation what I could find about locked files, but do not clearly understand how one might be different to the other, in practice. Are locked files more secure than runtime files? Why even have a runtime option if there is a locked file option? Are there other options I should be aware of? --------------735774E5C346C3564390DD18 name="torres.vcf" Content-Description: Card for Daniel Torres Content-Disposition: attachment; filename="torres.vcf" begin:vcard n:Torres;Daniel tel;fax:1-978.383-5817 tel;home:1-305.441-0369 tel;work:1-305.461-6829 x-mozilla-html:TRUE adr:;;600 Biltmore Way APT 412;Coral Gables;FL;33134-7529;USA version:2.1 email;internet:torres@SAmerica.com fn:Daniel Torres end:vcard --------------735774E5C346C3564390DD18-- -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 11 07:05:40 1999 From: "Chris Burke" References: <3787CF6F.BC92EDCF@SAmerica.com> Subject: Re: Jforum: Runtime versus Locked files Date: Sun, 11 Jul 1999 07:02:13 -0400 charset="iso-8859-1" > I developed a short script to solve a math problem, but someone else is > interested in having it. I want to give it to them, but I do not want them to > know how the script works. I would have sent them the script in a runtime > version if I hadn't come across the locked file option. I read in the > documentation what I could find about locked files, but do not clearly > understand how one might be different to the other, in practice. > > Are locked files more secure than runtime files? Why even have a runtime > option if there is a locked file option? Are there other options I should be > aware of? These are the only two encoding options. The script encoding is very similar, however there is a difference in the way they can be used. What you send depends largely on whether the recipient will be using a J development system or a runtime system. A runtime script can only be loaded by a runtime system, i.e. a system started with the /rt parameter. There is no way your user can poke around what has been defined by the script. If your user has only the runtime system, then there is no effective difference between a runtime script and a locked script. If your user has the development system, then sending a runtime script forces him to use the runtime system when using the script. A locked script can be also loaded by a development system. The user can see the names defined by the script, but not their definitions. Sending a locked script is more appropriate when your user has a J development system, and you have no objection to the development system being available when your script is loaded. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 11 12:05:52 1999 From: "Joan J. Ordinas" Subject: Jforum: Defining functions Date: Sun, 11 Jul 1999 18:01:09 +0200 charset="iso-8859-1" Importance: Normal In-Reply-To: <3787CF6F.BC92EDCF@SAmerica.com> I'm a bit confused with the mechanism for defining new functions (verbs) in J. Is there any text about this subject, preferably comparing the issue with other languages? Thanks in advance for your attention. JJOR -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 11 14:32:23 1999 Date: Sun, 11 Jul 1999 21:31:08 +0300 From: Anssi Seppala Subject: Re: Jforum: Defining functions In-Reply-To: <000401becbb6$97231de0$a14805c3@ariadna> References: <3787CF6F.BC92EDCF@SAmerica.com> Look at the Help/Primer/Part B/Verb defienition: The following is a definition of the verb centigrade that will convert its argument from a Fahrenheit value to a centigrade value. centigrade =. 3 : 0 t1 =. y. - 32 t2 =. t1 * 5 t3 =. t2 % 9 ) Anssi At 18:01 11.7.1999 +0200, you wrote: >I'm a bit confused with the mechanism for defining >new functions (verbs) in J. Is there any text about >this subject, preferably comparing the issue with >other languages? > >Thanks in advance for your attention. > >JJOR > > >---------------------------------------------------------------------------- >---- >J Forum: for information about this list, see >http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Sun Jul 11 22:08:50 1999 Date: Sun, 11 Jul 1999 22:01:25 -0400 From: Murray Eisenberg Organization: Mathematics & Statistics, Univ. of Mass./Amherst Subject: Re: Jforum: Defining functions References: <3787CF6F.BC92EDCF@SAmerica.com> <4.1.19990711212734.00a641e0@pop.kolumbus.fi> I always shudder a bit when I see that example from the Primer! If you were REALLY going to define it as an explicit function (instead of tacitly), surely it is preferable to eliminate the unnecessary intermediate results and use something like centrigrade=: monad define (y. - 32) * 5 % 9 ) or centigrade=: monad define 5r9 * (y. - 32) ) or even: centigrade=: monad define 5r9 * y. - 32 ) I worry about first impressions "sticking". Anssi Seppala wrote: > > Look at the Help/Primer/Part B/Verb defienition: > > The following is a definition of the verb centigrade that will convert its > argument from a Fahrenheit value to a centigrade value. > > centigrade =. 3 : 0 > t1 =. y. - 32 > t2 =. t1 * 5 > t3 =. t2 % 9 > ) > > Anssi > > At 18:01 11.7.1999 +0200, you wrote: > >I'm a bit confused with the mechanism for defining > >new functions (verbs) in J. Is there any text about > >this subject, preferably comparing the issue with > >other languages? > > > >Thanks in advance for your attention. > > > >JJOR > > > > > >---------------------------------------------------------------------------- > >---- > >J Forum: for information about this list, see > >http://www.jsoftware.com/forum.htm > > -------------------------------------------------------------------------------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 02:10:22 1999 From: "Roger Hui" References: <3784F3ED.E5B10360@interlog.com> Subject: Re: Jforum: Formal Description of J ? Date: Sun, 11 Jul 1999 23:02:20 -0400 charset="iso-8859-1" Bill Kindree writes on Thursday, July 8: > Anyone know of work being done on a formal description of J, along the lines of > similar works that have been done for APL? I don't know of any such current work. The closest thing along these lines is the book "An Implementation of J". In it are models in J of various primitives and interpreter processes. Other models are in J papers in various places. (e.g. a model of the rank operator is in the APL95 paper "Rank and Uniformity".) -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 08:49:25 1999 From: "Randy MacDonald" References: <000401becbb6$97231de0$a14805c3@ariadna> Subject: Re: Jforum: Defining functions Date: Mon, 12 Jul 1999 08:45:17 -0400 charset="iso-8859-1" Note that functions, or in J terminology, verbs, are defined much the same way as nouns are, with an assignment of an expression to a value. J is quite rich in the expressions which can result in verbs, but it also allows quoted strings to be used as expressions. For example: plus =: + 3 plus 4 7 defines a new verb, called "plus" whose definition is exactly the same as the primitive verb "+". An expression made up of verbs and adverbs which results in a verb can also be assigned to a name: sum =: +/ sum 1 2 3 6 mean =: +/%# mean 5 10 12 9 All of these are examples of "tacit" definition, a facility that very few languages share. If you are looking for something along the lines of plus(a,b) { return a + b } then J can do this kind of thing as well. This is called "explicit" definition. plus =: dyad : 'x. + y.' 3 plus 4 7 Note that you have no flexibility with the naming of your arguments: x. represents the left argument, y. represents the right argument. Where your programs get longer than one line, you can "include" lines of program as part of your input: In C-like languages: plus (a,b){ // a comment inside the function return (a + b) ) plus =: dyad : 0 NB. the 0 means "everything up to a line with ) starting it." NB. a comment inside the verb x. + y. ) 3 plus 4 7 The result of this kind of verb is the result of the last expression executed within the definition of the verb. Note as well that because the definitions of verbs are themselves expressions, you can wrap them in parentheses and use them directly in expressions. For example, with a previous definition of plus: 3 ( dyad : 'x. + y.') 4 7 or to get a little more complex: 3 ( dyad " '(2 * x.) + 3 * y.' ) 4 18 >From: Joan J. Ordinas > I'm a bit confused with the mechanism for defining > new functions (verbs) in J. Is there any text about > this subject, preferably comparing the issue with > other languages? > > Thanks in advance for your attention. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 11:14:02 1999 From: Ajith Prasad Subject: RE: Jforum: Defining functions Date: Mon, 12 Jul 1999 23:10:45 +0800 charset="windows-1252" The J documentation does not do full justice to J's ability to define verbs. For example, I did not see any reference to how so basic a requirement as computing a monadic function of, say, 3 parameters could be handled until I got the answer from Chris Burke. For example, if one wanted to compute a linear combination of of a,b,c where the a,b,c vary but the weights are fixed (say,3,4 and 5), one could define: f=: monad define 'a b c'=. y. result=. (3*a)+(4*b)+(5*c) ) and call f 1;2;3 to get 26 as required, etc. This particular example is trivial but the concept is of general applicability. I don't think the Primer covers examples of this nature. This is not meant as a criticism but simply to point out that the documetation does not cover some cases which are of quite general interest. -----Original Message----- From: Murray Eisenberg Sent: 7/12/99 10:01 AM Subject: Re: Jforum: Defining functions I always shudder a bit when I see that example from the Primer! If you were REALLY going to define it as an explicit function (instead of tacitly), surely it is preferable to eliminate the unnecessary intermediate results and use something like centrigrade=: monad define (y. - 32) * 5 % 9 ) or centigrade=: monad define 5r9 * (y. - 32) ) or even: centigrade=: monad define 5r9 * y. - 32 ) I worry about first impressions "sticking". Anssi Seppala wrote: > > Look at the Help/Primer/Part B/Verb defienition: > > The following is a definition of the verb centigrade that will convert its > argument from a Fahrenheit value to a centigrade value. > > centigrade =. 3 : 0 > t1 =. y. - 32 > t2 =. t1 * 5 > t3 =. t2 % 9 > ) > > Anssi > > At 18:01 11.7.1999 +0200, you wrote: > >I'm a bit confused with the mechanism for defining > >new functions (verbs) in J. Is there any text about > >this subject, preferably comparing the issue with > >other languages? > > > >Thanks in advance for your attention. > > > >JJOR > > > > > >----------------------------------------------------------------------- ----- > >---- > >J Forum: for information about this list, see > >http://www.jsoftware.com/forum.htm > > ------------------------------------------------------------------------ -------- > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -- Murray Eisenberg murray@math.umass.edu Mathematics & Statistics Dept. phone 413 549-1020 (H) Univ. of Massachusetts 413 545-2859 (W) Amherst, MA 01003-4515 ------------------------------------------------------------------------ -------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 11:22:57 1999 Date: Mon, 12 Jul 1999 09:22:05 -0400 From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= Subject: Re: Jforum: OCX's charset=ISO-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id LAA23097 Message text written by INTERNET:forum@jsoftware.com >I was fascinated by the VB6.0... < Does this mean that you don't use OCX's? It was your only attempt? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 11:24:21 1999 From: Andrew Nikitin Subject: Jforum: regexp Date: Mon, 12 Jul 1999 10:21:12 -0500 Does anybody know wether regular expressions are implemented in J manually, or it just calls some standard dll? If it calls dll, then what is the name of the dll? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 11:26:12 1999 From: "Randy MacDonald" Subject: Jforum: Typo... Date: Mon, 12 Jul 1999 11:22:11 -0400 charset="Windows-1252" ... a typo in the last example: a quote was typed where a colon belonged. (B=) On Mon, 12 Jul 1999, Randy MacDonald wrote: [snip] +or to get a little more complex: + + 3 ( dyad " '(2 * x.) + 3 * y.' ) 4 +18 This SHOULD read: 3 ( dyad : '(2 * x.) + 3 * y.' ) 4 18 -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 12:22:37 1999 From: "Randy MacDonald" References: Subject: Re: Jforum: Defining functions Date: Mon, 12 Jul 1999 12:16:27 -0400 charset="Windows-1252" This is all a by-product of J verbs being able to work on arguments and to give results of arbitrary (up to machine limits) complexity. You can build an argument any way you want, and you can take it apart within the verb. Perhaps those of us who came from APL backgrounds expect this, so we leave it unsaid. Remembering that J, like APL before it, is the first programming language for some people, suggests that more combinations of expressions deserve to be made explicit; the chance to discover them on one's own is lost, however. That said, the use of ; or "link" to make a list of arguments, and the use of "named" assignment are all included within the documentation as is. The case below could have been done as follows: f =: monad define a =. > 0 { y. b =. > 1 { y. c =. > 2 { y. (3*a) + (4*b) + (5*c) NB. note that last sentence passes for result. ) ----- Original Message ----- From: Ajith Prasad Sent: July 12, 1999 11:10 AM Subject: RE: Jforum: Defining functions > The J documentation does not do full justice to J's ability to define > verbs. For example, I did not see any reference to how so basic a > requirement as computing a monadic function of, say, 3 parameters could be > handled until I got the answer from Chris Burke. For example, if one wanted > to compute a linear combination of of a,b,c where the a,b,c vary but the > weights are fixed (say,3,4 and 5), one could define: > > f=: monad define > 'a b c'=. y. > result=. (3*a)+(4*b)+(5*c) > ) > > and call f 1;2;3 to get 26 as required, etc. This particular example is > trivial but the concept is of general applicability. I don't think the > Primer covers examples of this nature. This is not meant as a criticism but > simply to point out that the documetation does not cover some cases which > are of quite general interest. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 12:35:35 1999 From: gosi@centrum.is Subject: Re: Jforum: Runtime versus Locked files Date: Mon, 12 Jul 99 16:31:09 +0000 Chris Burke wrote: > A runtime script can only be loaded by a runtime system, > A locked script can be also loaded by a development system. The user can see > the names defined by the script, but not their definitions. Thanks for that explanations. I have often wondered about this. /Gosi -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 12:56:44 1999 From: "Oleg Kobchenko" References: <199907120922_MC2-7CAE-5153@compuserve.com> Subject: Re: Jforum: OCX's Date: Mon, 12 Jul 1999 19:53:38 +0300 charset="ISO-8859-1" Content-Transfer-Encoding: 8bit Actually, that meant that I had problems with a particular set of Data-aware ActiveX Controls. I have some examples of trivial other ActiveX controls: ListView, TreeView, MSFlexGrid, WebBrowser. Apart from that, some attempts of using Toolbar and Coolbar. However, those already required some inter-component awareness. Plus the behavior of sizing and positioning on the form in J is different form VB. ----- Original Message ----- From: Alain Miville de Ch�ne Sent: Monday, July 12, 1999 4:22 PM Subject: Re: Jforum: OCX's > Message text written by INTERNET:forum@jsoftware.com > >I was fascinated by the VB6.0... < > > Does this mean that you don't use OCX's? It was your only attempt? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 13:36:21 1999 Date: Mon, 12 Jul 1999 13:30:11 -0400 From: Eric Iverson Organization: Iverson Software Inc. Subject: Re: Jforum: regexp References: Andrew Nikitin wrote: > Does anybody know wether regular expressions are implemented in J > manually, or it just calls some standard dll? If it calls dll, then what > is the name of the dll? Currently the regex code used by the 17!:x foreigns is linked in as part of the J engine and is not available as a separate dll. In a future release this code will be broken out into a separate dll. Why do you ask? -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 13:54:59 1999 From: Andrew Nikitin Subject: RE: Jforum: regexp Date: Mon, 12 Jul 1999 12:52:10 -0500 > -----Original Message----- > From: Eric Iverson [SMTP:eiverson@interlog.com] > Sent: Monday, July 12, 1999 12:30 > > Currently the regex code used by the 17!:x foreigns is linked in as > part > of the J engine and is not available as a separate dll. In a future > release this code will be broken out into a separate dll. Why do you > ask? Some time ago there was file named regexp.dll (or something like this) in my w95 system directory. I was completely sure that it is standard and is being referenced from j like the previous case of odbc or curent case of sockets. Recently I decided to utilize this "standard" file for my own needs and did not found it in system directory (my old system directory in nine time zones from me and I cannot check it again). Now I suspect I was wrong thinking of this file as of standard. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 16:55:59 1999 Date: Mon, 12 Jul 1999 16:49:29 -0400 From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= Subject: Jforum: bug in debug charset=ISO-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id QAA01347 dbsq'' 45 *:*;Reduce *:*; dbs'' |dbs[0] | dbs'' | y.-:Empty |Reduce[:73] | 'p1 p2'=. References: Subject: Re: Jforum: Defining functions Date: Mon, 12 Jul 1999 21:15:53 +0300 charset="windows-1252" Continuing the thoughts of Randy MacDonald, in that J/APL treats their data differently that other (procedural) programming languages, I would recommend (again) the good-old Gilmann and Rose. There you will find what somehow escaped (for known reasons of different scope treatment than in APL) the documentation in J: the invocation context of a verb (function). There are only two possible arguments: right and left (left is optional). +-----------+ | L R | | +---+ | | | V | | | +---+ | +-----------+ What you do to them is up to you. BTW, in Perl there is only one _right_ argument, and you also have to do something like: my(a,b,c) = @_; to have 'named' parameters. Almost the same idea exists in some LISP dialects, where the whole list of arguments is accessilbe. There you operate on the lists with CARs and CDRs to get the same discrete atoms of the argument. In J there are additional techniques to allow for default values of the arguments and to foolproof errors: 'a b c'=. 3{.y. or 'a b c'=. 3{. boxopen y. or 'a b c'=. 3{.y.,_1,_1 ----- Original Message ----- From: Ajith Prasad Sent: Monday, July 12, 1999 6:10 PM Subject: RE: Jforum: Defining functions > The J documentation does not do full justice to J's ability to define > verbs. For example, I did not see any reference to how so basic a > requirement as computing a monadic function of, say, 3 parameters could be > handled until I got the answer from Chris Burke. For example, if one wanted > to compute a linear combination of of a,b,c where the a,b,c vary but the > weights are fixed (say,3,4 and 5), one could define: > > f=: monad define > 'a b c'=. y. > result=. (3*a)+(4*b)+(5*c) > ) > > and call f 1;2;3 to get 26 as required, etc. This particular example is > trivial but the concept is of general applicability. I don't think the > Primer covers examples of this nature. This is not meant as a criticism but > simply to point out that the documetation does not cover some cases which > are of quite general interest. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 22:32:48 1999 Date: Tue, 13 Jul 1999 11:26:18 -0700 From: Steven Phillips Organization: Electrotechnical Laboratory Subject: Jforum: regex builder lab Hi, There appear to be a small problem with the definition of "of" in: j403\system\packages\regex\regbuild.ijs. For example: 3 4 of 't' t(3,4} should be: t{3,4} The typo is in the line: of=: singlecomp@] , '('"_ , ":@{.@[ , ','"_ , ":@{:@[ , '}'"_ which I guess should read: of=: singlecomp@] , '{'"_ , ":@{.@[ , ','"_ , ":@{:@[ , '}'"_ Steve -- ------------------------------------------------------------------------ Steven Phillips, PhD Electrotechnical Laboratory Information Science Division 1-1-4 Umezono, Tsukuba Ibaraki 305-8568 Japan Phone: +81 298 543315 Fax : +81 298 545857 Email: stevep@etl.go.jp Home page: http://www.etl.go.jp/~stevep ------------------------------------------------------------------------ -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Mon Jul 12 22:47:02 1999 From: "Chris Burke" References: <378B84CA.3A75@etl.go.jp> Subject: Re: Jforum: regex builder lab Date: Mon, 12 Jul 1999 22:46:05 -0400 charset="iso-8859-1" Thanks, fixed now. ----- Original Message ----- From: Steven Phillips Sent: Tuesday, July 13, 1999 2:26 PM Subject: Jforum: regex builder lab > Hi, > > There appear to be a small problem with > the definition of "of" in: > j403\system\packages\regex\regbuild.ijs. > For example: > 3 4 of 't' > t(3,4} > should be: > t{3,4} > > The typo is in the line: > of=: singlecomp@] , '('"_ , ":@{.@[ , ','"_ , ":@{:@[ , '}'"_ > which I guess should read: > of=: singlecomp@] , '{'"_ , ":@{.@[ , ','"_ , ":@{:@[ , '}'"_ -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 13 02:15:58 1999 From: "Roger Hui" References: <005b01becc92$96f54320$0801a8c0@spin.kharkov.ua> Subject: Re: Jforum: Defining functions Date: Mon, 12 Jul 1999 23:03:55 -0400 charset="Windows-1252" Oleg Kobchenko writes on Monday, July 12: > Continuing the thoughts of Randy MacDonald, in that > J/APL treats their data differently that other (procedural) > programming languages, I would recommend (again) the good-old > Gilmann and Rose. There you will find what somehow escaped > (for known reasons of different scope treatment than in APL) > the documentation in J: the invocation context of a verb (function). > There are only two possible arguments: right and left (left is optional). Actually, the topic did not "escape the documentation in J". Please find in the Dictionary, section II Grammar, subsection E Parsing and Execution, under the summary of the parsing rules, 3. A verb is applied dyadically if possible; that is, if preceded by a noun that is not itself the right argument of a conjunction. and then immediately following the summary, One important consequence of these rules is that in an unparenthesized expression the right argument of any verb is the result of the entire phrase to its right. The sentence 3*p%q^|r-5 can therefore be read from left to right: the overall result is 3 times the result of the remaining phrase, which is the quotient of p and the part following the %, and so on. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 13 18:35:46 1999 From: dajoy@hoy.net (Ajoy Victor) Date: Tue, 13 Jul 1999 17:30:19 -0500 Content-type: text/plain; charset=US-ASCII Subject: Jforum: Genetic Algorithms I like the little J I know. I'm using the Freeware version. I'm doing some work with genetic algorithms. Has someone developed some tools or framework to program GAs (specially permutation crossover methods but any idea will do) ? I would appreciate any kind of advice. Daniel -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Tue Jul 13 20:22:07 1999 Date: Tue, 13 Jul 1999 20:21:52 -0400 From: Tom Chwastyk Organization: US Naval Research Laboratory Subject: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? References: <19990713222948.AAB41856@default> I would like to be able to capture the screen bitmap (as with the PrtSc key) so I can use J 4.01a to record data from applications that will display results on the screen but where I don't have access to results by DDE or OLE. (The current case is the "Battery Meter" on my laptop, but I've had occasion to use this with APL in the past). On Windows 3.1x, I could do this with the Windows Macro Recorder, pasting the resulting bitmap into Paint, saving the image, and reading the file in APL to recognize the text. Sure, there must be easier and faster ways to do this -- but not without documentation I don't have. With Win95 on the laptop, I no longer have macro recorder. I can no longer even tell what the name of the battery meter application is -- or whether it's something built into System/0008 (the only hit I could find in the registry). Suggestions, anyone? -- Tom Chwastyk -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From ljdickey Tue Jul 13 22:25:21 1999 Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? Is this a function of the screen resolution? From owner-jsoftware@lists.interlog.com Tue Jul 13 22:36:51 1999 Date: Tue, 13 Jul 1999 22:31:37 -0400 From: Eric Iverson Organization: Iverson Software Inc. Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? References: <19990713222948.AAB41856@default> <378BD820.CC2CDE52@cms.nrl.navy.mil> Tom Chwastyk wrote: > I would like to be able to capture the screen bitmap (as with the PrtSc key) so I > can use J 4.01a to record data from applications that will display results on the > screen but where I don't have access to results by DDE or OLE. (The current case is > the "Battery Meter" on my laptop, but I've had occasion to use this with APL in the > past). > > On Windows 3.1x, I could do this with the Windows Macro Recorder, pasting the > resulting bitmap into Paint, saving the image, and reading the file in APL to > recognize the text. Sure, there must be easier and faster ways to do this -- but not > without documentation I don't have. With Win95 on the laptop, I no longer have macro > recorder. I can no longer even tell what the name of the battery meter application > is -- or whether it's something built into System/0008 (the only hit I could find in > the registry). > > Suggestions, anyone? WOW! You're going to automate getting the battery power status from pixels. I don't want to discourage you and would love to see the final app. But you might want to consider using the DLL call facility. You could do this in J4.01, but upgrading to J4.03 has somewhat improved DLL call support and much better documentation on its use in the labs. The following info is extracted from the MSDN (with somewhat messy formatting). *** GetSystemPowerStatus The GetSystemPowerStatus function retrieves the power status of the system. The status indicates whether the system is running on AC or DC power, whether the battery is currently charging, and how much battery life remains. BOOL GetSystemPowerStatus( LPSYSTEM_POWER_STATUS lpSystemPowerStatus ); Parameters lpSystemPowerStatus Pointer to a SYSTEM_POWER_STATUS structure that receives status information. Return Values If the function succeeds, the return value is nonzero. If the function fails, the return value is zero. To get extended error information, call GetLastError. **** SYSTEM_POWER_STATUS The SYSTEM_POWER_STATUS structure contains information about the power status of the system. typedef struct _SYSTEM_POWER_STATUS { BYTE ACLineStatus; BYTE BatteryFlag; BYTE BatteryLifePercent; BYTE Reserved1; DWORD BatteryLifeTime; DWORD BatteryFullLifeTime; } SYSTEM_POWER_STATUS; typedef struct SYSTEM_POWER_STATUS *LPSYSTEM_POWER_STATUS; Members ACLineStatus AC power status. This parameter can be one of the following values. Value Meaning 0 Offline 1 Online 255 Unknown status. All other values are reserved. BatteryFlag Battery charge status. This parameter can be a combination of the following values. Value Meaning 1 High 2 Low 4 Critical 8 Charging 128 No system battery 255 Unknown status All other values are reserved. BatteryLifePercent Percentage of full battery charge remaining. This member can be a value in the range 0 to 100, or 255 if status is unknown. All other values are reserved. Reserved1 Reserved; must be zero. BatteryLifeTime Number of seconds of battery life remaining, or 0xFFFFFFFF if remaining seconds are unknown. BatteryFullLifeTime Number of seconds of battery life when at full charge, or 0xFFFFFFFF if full battery lifetime is unknown. Remarks The system is only capable of estimating BatteryFullTime based on calculations on BatteryLifeTime and BatteryLifePercent. Without smart battery subsystems, this value may not be accurate enough to be useful. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From tomC@cms.nrl.navy.mil Wed Jul 14 06:58:40 1999 Date: Wed, 14 Jul 1999 07:02:58 -0400 From: Tom Chwastyk Organization: US Naval Research Laboratory Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? References: <199907140225.WAA00594@pythagoras.math.uwaterloo.ca> Lee, I'm not sure what you mean. What I got before was an image on the clipboard which I would paste into an appropriate sized image in Paint and save as a bitmap. Since the bitmap was always the same size and type it was easy to locate the pixels that gave efficient signatures for the changes I was looking for. Interpreting the bitmap varies with the color depth (8, 16, or 24 bit) and the bitmap size (pixels high and wide). It does not vary with resolution (640, 800, 1024, ...) unless the application deliberately changes its display with screen dpi (e.g., to maintain image size in inches or cm). Does this answer your question? Tom Lee Dickey wrote: > Is this a function of the screen resolution? -- Tom Chwastyk (sounds like Fosdick; Polish: CH='H, W=F (here), A=AH, styk=stick, soften strict 'HFAH-stick) From owner-jsoftware@lists.interlog.com Wed Jul 14 07:24:40 1999 From: "John Tsolometes" References: <19990713222948.AAB41856@default> <378BD820.CC2CDE52@cms.nrl.navy.mil> <378BF689.5AD5@interlog.com> Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? Date: Wed, 14 Jul 1999 06:20:14 -0400 charset="iso-8859-1" This would be nessesary if J had a reflective or self-referential ability. It would then be a trivial matter to Code a message sniffer to get the output of the program Or use any number of established techniques that work in C++ or assembly language. Many things just don't work in API's. They are anything but flexible. ----- Original Message ----- From: Eric Iverson Sent: Tuesday, July 13, 1999 10:31 PM Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results... How? > Tom Chwastyk wrote: > > I would like to be able to capture the screen bitmap (as with the PrtSc key) so I > > can use J 4.01a to record data from applications that will display results on the > > screen but where I don't have access to results by DDE or OLE. (The current case is > > the "Battery Meter" on my laptop, but I've had occasion to use this with APL in the > > past). > > > > On Windows 3.1x, I could do this with the Windows Macro Recorder, pasting the > > resulting bitmap into Paint, saving the image, and reading the file in APL to > > recognize the text. Sure, there must be easier and faster ways to do this -- but not > > without documentation I don't have. With Win95 on the laptop, I no longer have macro > > recorder. I can no longer even tell what the name of the battery meter application > > is -- or whether it's something built into System/0008 (the only hit I could find in > > the registry). > > > > Suggestions, anyone? > > WOW! You're going to automate getting the battery power status from > pixels. I don't want to discourage you and would love to see the final > app. > > But you might want to consider using the DLL call facility. You could do > this in J4.01, but upgrading to J4.03 has somewhat improved DLL call > support and much better documentation on its use in the labs. The > following info is extracted from the MSDN (with somewhat messy > formatting). > > *** > GetSystemPowerStatus > > The GetSystemPowerStatus function retrieves the power status of the > system. The status indicates whether the system is running on AC or DC > power, whether the battery is currently charging, and how much battery > life remains. > > BOOL GetSystemPowerStatus( > LPSYSTEM_POWER_STATUS lpSystemPowerStatus > ); > > Parameters > lpSystemPowerStatus > Pointer to a SYSTEM_POWER_STATUS structure that receives status > information. > > Return Values > If the function succeeds, the return value is nonzero. > > If the function fails, the return value is zero. To get extended error > information, call GetLastError. > > **** > > SYSTEM_POWER_STATUS > The SYSTEM_POWER_STATUS structure contains information about the power > status of the system. > > typedef struct _SYSTEM_POWER_STATUS { > BYTE ACLineStatus; > BYTE BatteryFlag; > BYTE BatteryLifePercent; > BYTE Reserved1; > DWORD BatteryLifeTime; > DWORD BatteryFullLifeTime; > } SYSTEM_POWER_STATUS; > typedef struct SYSTEM_POWER_STATUS *LPSYSTEM_POWER_STATUS; > > Members > ACLineStatus > AC power status. This parameter can be one of the following values. > Value Meaning > 0 Offline > 1 Online > 255 Unknown status. > > > All other values are reserved. > > BatteryFlag > Battery charge status. This parameter can be a combination of the > following values. Value Meaning > 1 High > 2 Low > 4 Critical > 8 Charging > 128 No system battery > 255 Unknown status > > > All other values are reserved. > > BatteryLifePercent > Percentage of full battery charge remaining. This member can be a value > in the range 0 to 100, or 255 if status is unknown. All other values are > reserved. > Reserved1 > Reserved; must be zero. > BatteryLifeTime > Number of seconds of battery life remaining, or 0xFFFFFFFF if remaining > seconds are unknown. > BatteryFullLifeTime > Number of seconds of battery life when at full charge, or 0xFFFFFFFF if > full battery lifetime is unknown. > Remarks > The system is only capable of estimating BatteryFullTime based on > calculations on BatteryLifeTime and BatteryLifePercent. Without smart > battery subsystems, this value may not be accurate enough to be useful. > > > -------------------------------------------------------------------------- ------ > J Forum: for information about this list, see http://www.jsoftware.com/forum.htm > -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 14 15:31:25 1999 Date: Wed, 14 Jul 1999 21:27:42 +0200 From: Martin Neitzel Subject: Re: Jforum: regexp In-Reply-To: Organization: Gaertner Datensysteme, Braunschweig, Germany >Recently I decided to utilize this "standard" file for my own needs and >did not found it in system directory (my old system directory in nine >time zones from me and I cannot check it again). Now I suspect I was >wrong thinking of this file as of standard. The regexp.dll you saw was not from the J system. In case you are wondering about the exact regexp version used with J: It's the GNU rx library. This one is comparatively slow and big, but it does implement the Posix standard. Personally, I'm open-minded about relaxing this arrangement a bit to use whatever is the native regexp on a platform. The J binary could then be a bit smaller, but not all systems have Posix-compliant regex-libraries natively. Any opinions on this? Martin -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Wed Jul 14 23:08:01 1999 Date: Wed, 14 Jul 1999 23:02:53 -0400 From: Fraser Jackson Subject: RE: Jforum: Defining functions charset=ISO-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id XAA25430 One of the things we need to do much more of in J is to define sequences of boxed objects which together convey essential features of some domain of discourse. Users of other languages define classes of objects with such structures, and build extensive libraries to manipulate them. By restricting function arguments to unboxed array objects we seriously limit the brilliant expressive power of J. As one who routinely uses boxed arguments for functions, Murray's response on defining functions is interesting. Where is the behaviour of opening the boxes when using the copula indirectly specified in the Dictionary? Since there are three boxed items in the list 1;2;3 one would have expected that 'a b c' =. 1;2;3 would give three boxed items. It does not. It opens them. Fraser -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 15 09:06:59 1999 Date: Thu, 15 Jul 1999 08:58:50 -0400 From: David Ness Subject: Re: Jforum: Defining functions References: <199907142303_MC2-7D0E-18CA@compuserve.com> Fraser Jackson wrote: > > ... Where is the behaviour of opening the boxes when using the > copula indirectly specified in the Dictionary? > > Since there are three boxed items in the list > > 1;2;3 > > one would have expected that > > 'a b c' =. 1;2;3 > > would give three boxed items. It does not. It opens them. > > Fraser > In the `Copula' definition we have: x =: 'abc';'c' (x) =: 3 4; 5 6 7 where the implicit unboxing seems much more natural, as the left side structure parallels the right side structure. Of course in: x =: 'abc';'c' (x) =: 10 20 we don't break the rule, as (>10) -: 10, but it is probably this that is the more `special case'. I guess that leaves implicit the interpretation when the left hand side is: 'abc c' =: as ('abc';'c') =: -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Thu Jul 15 09:21:58 1999 From: dajoy@hoy.net (Ajoy Victor) Date: Thu, 15 Jul 1999 08:19:38 -0500 Content-type: text/plain; charset=US-ASCII Subject: Jforum: Newer version of FAQ? Does anyone has a newer version of this FAQ? Subject: J Programming FAQJ Programming FAQ Compiled by: Ehud Lamm Updated : 3/6/95 Posted to : comp.lang.apl Changes : example of under, obverse, trains, rank error. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 04:03:22 1999 From: "eesuk" Subject: Jforum: JForum: equvalent tacit form Date: Fri, 16 Jul 1999 16:57:18 +0900 charset="euc-kr" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from base64 to 8bit by plus.interlog.com id DAA26022 Is there any equivalent tacit form of F such that function G appears only one time? F =: G , -.@G + [ % G + (-~>:) F,G : dyadic function -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 05:20:48 1999 Date: Fri, 16 Jul 1999 12:02:13 +0300 From: Anssi Seppala Subject: Jforum: Why Amend ( } ) is so difficult in J? In-Reply-To: <000701becf60$dc9c6680$4f460693@trffpc2.kotel.co.kr> J is my favourite programming language. However I have allways been wondering the diffculties to set values into arrays with Amend ( } ). Well an interesteing point is that I actually need it rather rarely. Most of the programming is other transformations with Ravel ( , ). Today I again met a case of Amend that I do not understand directly. Here is the example m=. i. 4 4 NB. Let's have a matrix m (4 2$100) 1 2}"1 m NB. Amending of two columns with rank ("1) works fine 0 100 100 3 4 100 100 7 8 100 100 11 12 100 100 15 (4$100) 1}"1 m NB. But amending one column with rank ("1) is an error ????? |rank error | (4$100) 1}"1 m (4 1$100) 1}"1 m NB. I tried this too with same result. |rank error | (4 1$100) 1}"1 m (4$100) ( Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? Date: Fri, 16 Jul 1999 11:44:07 +0200 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id FAA03279 Try this: (4 1$100)(,1)}"1 m Regards/Paul >-----Message d'origine----- >De: Anssi Seppala [SMTP:anssi.seppala@enease.fi] >Date: vendredi 16 juillet 1999 11:02 >�: forum@jsoftware.com >Objet: Jforum: Why Amend ( } ) is so difficult in J? > >J is my favourite programming language. However I have allways been >wondering the diffculties to set values into arrays with Amend ( } ). Well >an interesteing point is that I actually need it rather rarely. Most of the >programming is other transformations with Ravel ( , ). > >Today I again met a case of Amend that I do not understand directly. Here >is the example > > m=. i. 4 4 NB. Let's have a matrix m > > (4 2$100) 1 2}"1 m NB. Amending of two columns with rank ("1) works fine > 0 100 100 3 > 4 100 100 7 > 8 100 100 11 >12 100 100 15 > > (4$100) 1}"1 m NB. But amending one column with rank ("1) is an >error ????? >|rank error >| (4$100) 1}"1 m > > (4 1$100) 1}"1 m NB. I tried this too with same result. >|rank error >| (4 1$100) 1}"1 m > > (4$100) (someone tell why the previous fails? > 0 100 2 3 > 4 100 6 7 > 8 100 10 11 >12 100 14 15 > > >Anssi > >----------------------------------------------------------------------------- >--- >J Forum: for information about this list, see >http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 06:58:51 1999 Date: Fri, 16 Jul 1999 13:42:43 +0300 From: Anssi Seppala Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? In-Reply-To: Thanks, but it is still a mystery why (4$100) 1}"1 m fails. Anssi At 11:44 16.7.1999 +0200, you wrote: >Try this: > >(4 1$100)(,1)}"1 m > >Regards/Paul -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 08:04:21 1999 From: Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? Date: Fri, 16 Jul 1999 14:01:40 +0200 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id IAA14469 Try this: (4$100) 1}"0 1 m Regards/Paul >-----Message d'origine----- >De: Anssi Seppala [SMTP:anssi.seppala@enease.fi] >Date: vendredi 16 juillet 1999 12:43 >�: forum@jsoftware.com >Objet: RE: Jforum: Why Amend ( } ) is so difficult in J? > >Thanks, but it is still a mystery why (4$100) 1}"1 m fails. Anssi > >At 11:44 16.7.1999 +0200, you wrote: >>Try this: >> >>(4 1$100)(,1)}"1 m >> >>Regards/Paul > > >----------------------------------------------------------------------------- >--- >J Forum: for information about this list, see >http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 08:41:45 1999 From: "Oleg Kobchenko" References: <000701becf60$dc9c6680$4f460693@trffpc2.kotel.co.kr> Subject: Re: Jforum: JForum: equvalent tacit form Date: Fri, 16 Jul 1999 15:40:24 +0300 charset="euc-kr" It might not be what you wanted, but in this form G by itself need not be present in F at all: F =: 11 : 'x. , -.@x. + [ % x. + (-~>:)' F ]: ([. , ([. (]: (([. (-.@)) + ([. ((]: + (-~ >:)) ([`%`) (`:6))))))) G F G , -.@G + [ % G + (-~ >:) Here, what puzzles me, is why the adverb definition insists on x. rather than u. ??? ----- Original Message ----- From: eesuk Sent: Friday, July 16, 1999 10:57 AM Subject: Jforum: JForum: equvalent tacit form > Is there any equivalent tacit form of F such that function G appears only one time? > > F =: G , -.@G + [ % G + (-~>:) > > F,G : dyadic function -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 10:08:43 1999 From: Andrew Nikitin Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? Date: Fri, 16 Jul 1999 09:04:22 -0500 > -----Original Message----- > From: Anssi Seppala [SMTP:anssi.seppala@enease.fi] > Sent: Friday, July 16, 1999 05:43 > > Thanks, but it is still a mystery why (4$100) 1}"1 m fails. Anssi It should fail. It would be very strange if it doesn't. Each cell specified in rank conjunction (in your case -- 1-cell, i.e. line) should agree in size to fit the left operand of the (1}) verb. This verb (1}) accepts left operands of the same shape as index expression. In your case -- atoms. So if you really meant "1 (and not "0 1) then you should write (100) 1}"1 m 0 100 2 3 4 100 6 7 8 100 10 11 12 100 14 15 Or in the case of (1 2}) verb you should write (2$100) 1 2}"1 m 0 100 100 3 4 100 100 7 8 100 100 11 12 100 100 15 nsg -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 10:11:34 1999 From: Andrew Nikitin Subject: RE: Jforum: JForum: equvalent tacit form Date: Fri, 16 Jul 1999 09:08:57 -0500 > -----Original Message----- > From: Oleg Kobchenko [SMTP:gccinc@usa.net] > Sent: Friday, July 16, 1999 07:40 > > Here, what puzzles me, is why the adverb definition > insists on x. rather than u. ??? u. is verb argument of explicitly defined adverbs. (More correct expression should be "adverbs that produce explicit, rather then tacit verbs".) In this case x. and y. are noun arguments of the entire verb phrase, and u. is verb argument of the adverb. nsg -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 16:52:56 1999 Date: Fri, 16 Jul 1999 23:51:14 +0300 From: Anssi Seppala Subject: Re: Jforum: Why Amend ( } ) is so difficult in J? In-Reply-To: <003401becfb1$b9c418f0$0801a8c0@spin.kharkov.ua> References: You have right. Most of my J skills are remembering patterns of sequences of J + some understanding. Because of this compact notation the patterns are easy to remember and re use. The way how J is presented as a language of "verbs","adverbs" and "nouns" is a bit strange for me, but I will not let it to disturb (ref. Thy Fly of the Bumblebee). :) Anssi PS. With Amend and From I am also remembering APL's selective assignment that is something that J is missing. For example (v#a)=. 1 2 3 works with APL, but not in J. Is this something that we will never see in J? At 20:36 16.7.1999 +0300, you wrote: >And this is a vivid example of the fact that J/APL have >incremental learning curve. I.e. to start doing something >you don't have to know everything. Moreover, you can use >certain features without understanding how it works. > >In this example, you intuitively guess that } for that >case requires a list argument: (,1) instead of just atom (1). >And you remember it and do so in other future instances. -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 17:16:31 1999 Date: Fri, 16 Jul 1999 16:45:02 +0300 From: Anssi Seppala Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? In-Reply-To: You must have a Black Belt in J to write such rank definitions :) At 14:01 16.7.1999 +0200, you wrote: >Try this: >(4$100) 1}"0 1 m >Regards/Paul -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 17:16:34 1999 From: Andrew Nikitin Subject: RE: Jforum: JForum: equvalent tacit form Date: Fri, 16 Jul 1999 11:11:09 -0500 Oops, no matter. One can use either x. or u. in "usual" adverb definition (but not both at a time!). But for some mysterious reasons 11 : convertor accepts only x., but not u.. My guess this is hard heritage from those dark ages when there were not such things as u. at all. > -----Original Message----- > From: Andrew Nikitin [SMTP:anikitin@fastenal.com] > Sent: Friday, July 16, 1999 09:09 > > -----Original Message----- > > From: Oleg Kobchenko [SMTP:gccinc@usa.net] > > Sent: Friday, July 16, 1999 07:40 > > Here, what puzzles me, is why the adverb definition > > insists on x. rather than u. ??? > u. is verb argument of explicitly defined adverbs... -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 17:16:35 1999 From: Subject: RE: Jforum: Why Amend ( } ) is so difficult in J? Date: Fri, 16 Jul 1999 16:41:14 +0200 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id KAA24240 Actually, Andrew Nikitin has a black belt in J, I am merely a J practitioner, but I have been exposed for a while, since 1990 actually. J is beyond my capacities but it's part of it's attractiveness and I am a curious person. I always enjoy Andrew Nikitin's explanations... If I could explain like him then I would raise my belt but I am not yet there :-) >-----Message d'origine----- >De: Anssi Seppala [SMTP:anssi.seppala@enease.fi] >Date: vendredi 16 juillet 1999 15:45 >�: forum@jsoftware.com >Objet: RE: Jforum: Why Amend ( } ) is so difficult in J? > >You must have a Black Belt in J to write such rank definitions :) >At 14:01 16.7.1999 +0200, you wrote: >>Try this: >>(4$100) 1}"0 1 m >>Regards/Paul > > >----------------------------------------------------------------------------- >--- >J Forum: for information about this list, see >http://www.jsoftware.com/forum.htm -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 17:21:10 1999 Date: Fri, 16 Jul 1999 12:45:00 -0400 From: David Ness Subject: Jforum: J in HTML? Can anyone tell me any easy way to display J output (in this case it happens to be arrays of boxes of numbers) in an HTML document, particularly when the recipient/reader of the document is not likely to have ISIJ on their machine? My examples look just fine on my machine (I can use either ISIJ or `Focal Point On-Line') but when I look at them on other people's machines, the nice little boxes become accented characters and superscripts and the like... -------------------------------------------------------------------------------- J Forum: for information about this list, see http://www.jsoftware.com/forum.htm From owner-jsoftware@lists.interlog.com Fri Jul 16 17:21:11 1999 From: "Oleg Kobchenko" References: <378F618C.E10B2AA4@Home.Com> Subject: Re: Jforum: J in HTML? Date: Fri, 16 Jul 1999 20:05:50 +0300 charset="iso-8859-1" 1. Use am OEM code-page for the HTML document, e.g. cp437 or cp866. But this only works in IE. Donno about latest NS. -- Plus a with   and
, or 
.

2. Can try the okhtml 'boxes to tables' converter.
It takes a J noun and returns a piece of HTML.
This works everywhere with very close visual resemblence
of boxes.

----- Original Message -----
From: David Ness 
Sent: Friday, July 16, 1999 7:45 PM
Subject: Jforum: J in HTML?

> Can anyone tell me any easy way to display J output
> (in this case it happens to be arrays of boxes of numbers) in an
> HTML document, particularly when the recipient/reader of the document
> is not likely to have ISIJ on their machine? My examples look just
> fine on my machine (I can use either ISIJ or `Focal Point On-Line') but
> when I look at them on other people's machines, the nice little boxes
> become accented characters and superscripts and the like...
>
> --------------------------------------------------------------------------------
> J Forum: for information about this list, see http://www.jsoftware.com/forum.htm
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 16 17:21:56 1999
Date: Fri, 16 Jul 1999 20:28:53 +0300
From: Anssi  Seppala 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
In-Reply-To: 

Thanks for the help. Obviously I expected too much correspondense bteween
Amend ( } ) and From ( { ).
Anssi

At 09:04 16.7.1999 -0500, you wrote:
>Each cell specified in rank conjunction (in your case -- 1-cell, i.e.
>line) should agree in size to fit the left operand of the (1}) verb.
>This verb (1}) accepts left operands of the same shape as index
>expression. In your case -- atoms. So if you really meant "1 (and not "0
>1) then you should write
>    (100) 1}"1 m
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 16 17:21:57 1999
From: "Oleg Kobchenko" 
References: 
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Fri, 16 Jul 1999 20:36:23 +0300
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit

I was enthusiastically watching the development of this thread,
bacause I sometimes stumble over certain cavities of J logic.

And this is a vivid example of the fact that J/APL have
incremental learning curve. I.e. to start doing something
you don't have to know everything. Moreover, you can use
certain features without understanding how it works.

In this example, you intuitively guess that } for that
case requires a list argument: (,1) instead of just atom (1).
And you remember it and do so in other future instances.

For a change you might want to get into the underlying
details of k-cells behavior and relationship. But even if
once this will become clear to you, this does not mean
you will be aware of the intrinsics, every time you use
the feature. Your memory will make up for that.
It is very much like you use 1/x as the derivative
of ln(x) not always ready to tell why it is so. BTW, many
people successfully do this, without having experienced
the moment of truth ever in their life.

----- Original Message -----
From: 
Sent: Friday, July 16, 1999 5:41 PM
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?

> Actually, Andrew Nikitin has a black belt in J, I am merely a J
> practitioner, but I have been exposed for a while, since 1990 actually.
> J is beyond my capacities but it's part of it's attractiveness and I am
> a curious person. I always enjoy Andrew Nikitin's explanations... If I
> could explain like him then I would raise my belt but I am not yet there
> :-)
>
> >-----Message d'origine-----
> >De: Anssi  Seppala [SMTP:anssi.seppala@enease.fi]
> >Date: vendredi 16 juillet 1999 15:45
> >�: forum@jsoftware.com
> >Objet: RE: Jforum: Why Amend ( } ) is so difficult in J?
> >
> >You must have a Black Belt in J to write such rank definitions  :)
> >At 14:01 16.7.1999 +0200, you wrote:
> >>Try this:
> >>(4$100) 1}"0 1 m
> >>Regards/Paul
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 16 17:55:24 1999
From: "Roger Hui" 
References: <000701becf60$dc9c6680$4f460693@trffpc2.kotel.co.kr> <001901becf88$6140ef60$0801a8c0@spin.kharkov.ua>
Subject: Re: Jforum: equvalent tacit form
Date: Fri, 16 Jul 1999 11:36:19 -0400
	charset="euc-kr"

EESuk writes on Friday, July 16:

> Is there any equivalent tacit form of  F  such that function G appears only one time?
>
> F =: G , -.@G + [ % G + (-~>:)
>
>  F,G : dyadic function

One technique is to write F1 in one the forms
 G H ]
 G H [
 G H ,&<
 G H ,
where H has the basic form of the original F, with [ substituting for G .
For the rest, depending on F, one'd substitute ] for [ , or if the rest
of F uses both the original left and right arguments, substitute
0&{::@] for [ and 1&{::@] for ]  , in the original F.  Thus:

F =: G , -.@G  + [ % G + (-~>:)
F1=: G ([ , -.@[  + 0&{::@] % [ + 0&{::@] (-~>:) 1&{::@]) ,&<

The phrase 0&{::@] (-~>:) 1&{::@] is equivalent to (-~>:)&>/@]  . So:

F2=: G ([ , -.@[  + 0&{::@] % [ + (-~>:)&>/@]) ,&<

The phrase -~>: is equivalent to -.@-  .  Also, if the arguments are
scalar numbers, boxing can be eliminated.  So:

F3 =: G ([ , -.@[  + 0&{@] % [ + -.@-/@]) ,

If F gets more complicated (e.g. having G1 G2 G3 ... that needs
to be mentioned just once each), the result gets quite unwieldy,
and I'd use explicit rather than tacit.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 02:05:11 1999
Date: Sat, 17 Jul 1999 01:42:45 -0400
From: Tom Chwastyk 
Subject: Re: Jforum: PrtSc (print screen) key to capture Battery Meter results...  How?
References: <19990713222948.AAB41856@default> <378BD820.CC2CDE52@cms.nrl.navy.mil> <378BF689.5AD5@interlog.com>

Eric Iverson wrote:

> WOW! You're going to automate getting the battery power status from
> pixels. I don't want to discourage you and would love to see the final
> app.
>
> But you might want to consider using the DLL call facility. You could do
> this in J4.01, but upgrading to J4.03 has somewhat improved DLL call
> support and much better documentation on its use in the labs. The
> following info is extracted from the MSDN (with somewhat messy
> formatting).
>
> ***
> GetSystemPowerStatus

[rest of helpful reply snipped]

Thanks for the lead. Just what I was looking for. I have it running in
4.01a with the DLL. The call takes about 14 msec., possibly for serial
checks with the smart battery.

I think the "WOW!" must have been like "WOW! You're going to find how
many liters of soda
you have by having J count drops with an eyedropper and divide by
20,000. I wouldn't want
to discourage you, and would love to see the final app. But you might
want to consider
looking on the bottle for the manufacturer's claimed volume, or getting
a more appropriate
measuring tool..."  :-) MIGHT want to consider? How diplomatic!

I must admit that my distrust of Microsoft and reluctance to clutter my
memory with myriad
changable details of the system calls sure had me set to go the long way
around the barn
on this one. The last time I used pixel copying like this was 1995; a 33
MHz, 4MB 486
captured minefield layouts from MS Minesweeper at the expert level.
(This was part of a
final project in a Bayesian decision theory course.)  There was *no* DLL
available for that,
I assume. The pixel analysis gave 3,300 layouts in an 15 hour run; most
of that time was spent outside APL. This sampling was enough to show
some interesting things ("significance" below was by an appropriate
chi-square test with p~0.122, although my prof was _very_ disappointed I
didn't use the equivalent Bayesian calculation):

- the upper right corner is a cold spot (significantly lower probability
  p_cold ~ 357 mines / 3300 games or p_cold~0.108 that a mine is
present)
- there is a p_hot ~ 1042/3300 ~ 0.316 hot spot one square to the right
of the
  upper left corner; however, the presence of a mine at the hot spot
does not
  depend significantly on the absence of a mine at the cold spot
- all the other locations are the same "temperature" (i.e., no
significant
  differences in mine probability (3300x99-(1042+357))/(3300x(480-2)) or
  p_avg~0.206226; note this is not significantly different from
99/480~0.20625)
- even if all the spots had had equal probability of being mined, the
fact that
  inside cells have 8 potentially mined neighbors while edge cells have
only 5
  and corner cells have only 3 means that corner cells, then edge cells,
are
  more likely to automatically "clear around" (a cell with a mine count
of zero,
  upon uncovering, automatically uncovers all its neigbors) than are
interior
  cells. The hot and cold spots make only minor differences in this.

I note from MSDN that MS has now made the one key that _can't_ be sent
by the SendKeys statement in VB etc. the PrtSc key. How very
interesting. Not surprising, though.

The pixel deciphering was trivial once I found a way to capture the
screen and find my way around a clipboard file. There turned out to be
16x30=480 fixed offsets into the file that gave pixels which, if black,
marked a mine or flag. All of 10 non-comment lines to read the .clp
file, check for validity (status frown), extract the mine map, track the
cell sums, and flush the mine maps in groups of 50 to a file (to keep
the ws small enough that everything ran in 4MB). There's an MS Word
write-up of the project, but at 1.6 MB (mostly two large screen shots)
it unsuitable for anything but e-mail to interested parties. If you want
it, send to my home e-mail, fosdick@erols.com . I'm not going to put any
code having to do with "research into playing MineSweeper" onto my work
computer.

--
Tom Chwastyk (sounds like Fosdick;
Polish: CH='H, W=F (here), A=AH,
styk=stick, soften strict 'HFAH-stick)

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 02:20:05 1999
From: gosi@centrum.is
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Sat, 17 Jul 99 06:18:18 +0000

Anssi.seppala@enease.fi writes:
> PS. With Amend and From I am also remembering APL's selective assignment
> that is something that J is missing. For example (v#a)=. 1 2 3 works with
> APL, but not in J. Is this something that we will never see in J?

Is this what you are asking for?

 ('v e a')=. 1 2 3
   v
1
   e
2
   a
3

/Gosi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 07:37:27 1999
From: "Chris Burke" 
References:  <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi>
Subject: Jforum: Why Amend ( } ) is so difficult in J?
Date: Sat, 17 Jul 1999 07:36:47 -0400
	charset="iso-8859-1"

Anssi writes:
> PS. With Amend and From I am also remembering APL's selective assignment
> that is something that J is missing. For example (v#a)=. 1 2 3 works with
> APL, but not in J. Is this something that we will never see in J?

You have this functionality in J, but in a different way; just as "indexed
assignment" is different in J.

Indexed assignment:

   A[i] <- B           APL
   A=. B i}A           J

Selective assignment:

   (F A) <- B          APL
   A=. B (F A)} A      J

In APL, assignment is a special syntax, in J it is just part of the ordinary
language.

For example, in APL you need a named object to be assigned into, but not in
J. Suppose you want to use assignment to create a list of the first 5
integers, but with item 2 of the list set to 100:

APL:

   A=: iota 5
   A[2] <- 100

J:

   100 (2)} i.5

Now suppose you have a mask v=. 0 0 1 0 0 that indicates the position to
replace with 100:

APL:

   A=: iota 5
   (v/A) <- 100

J:

   100 (bx v)} i.5

See also "J for the APL programmer" at
http://www.jsoftware.com/pubs/j4apl.html, and Vector Article Vol 13 No. 2
"APL&J - Indexing". The latter is not yet on our Web site - I'll try to get
it up in the next day or so.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 07:37:29 1999
From: "Chris Burke" 
References: 
Subject: Jforum: Why Amend ( } ) is so difficult in J?
Date: Sat, 17 Jul 1999 07:35:49 -0400
	charset="iso-8859-1"

Some comments on this thread:

1. In the following, note that scalars of the left argument are used to
amend in rows of the right argument, hence the rank is 0 1.

   m=. i.4 4

   (4$100) (1)}"0 1 m
 0 100  2  3
 4 100  6  7
 8 100 10 11
12 100 14 15

If the left argument is itself a matrix, then rows of the left argument are
used to amend in rows of the right argument, hence the rank is 1 1 (or, more
simply, just 1):

   (4 2$100) (1 3)}"1 1 m
 0 100  2 100
 4 100  6 100
 8 100 10 100
12 100 14 100

The following fails, because the rank specification of 1 is applied to both
left and right arguments, hence it tries to use the rows of the left
argument (all 4 numbers) to replace in each row of the right argument, but
it will not fit.

   (4$100) (1)}"1 m
rank error

2.  You can guess that this suggestion:

   (4 1$100)(,1)}"1 m

is not optimal, since the left argument is being reshaped into a matrix, and
the index is being ravelled. In general, rank largely eliminates the need,
that that you often see in APL, to reshape or ravel arguments to fit.

3.
>Thanks for the help. Obviously I expected too much correspondense bteween
>Amend ( } ) and From ( { ).

I dont understand this comment. There is a very close correspondence between
amend and from. Where might the above examples contradict this?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 08:19:20 1999
From: "Chris Burke" 
References: <378F618C.E10B2AA4@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Sat, 17 Jul 1999 08:21:06 -0400
	charset="iso-8859-1"

> Can anyone tell me any easy way to display J output
> (in this case it happens to be arrays of boxes of numbers) in an
> HTML document, particularly when the recipient/reader of the document
> is not likely to have ISIJ on their machine? My examples look just
> fine on my machine (I can use either ISIJ or `Focal Point On-Line') but
> when I look at them on other people's machines, the nice little boxes
> become accented characters and superscripts and the like...

One way in J403 is to save the output in a file (you can just save the ijx
session), then use menu Run|HTML Publish...

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 08:34:15 1999
Date: Sat, 17 Jul 1999 15:01:53 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
In-Reply-To: <3790202a1ab4001@blik.skima.is>

No, I meant selective assignment  (at least in APL2 and APL/W) where you
can combine the selection and assignment . For example using # or { operator :
a=. 1 1 1 1 1 1 1 1
v=.  1 0 0 0 1 0 1 0
(v#a)=. 1 2 3          NB. This is not working in today's J
(0 4 6{a) =. 1 2 3   NB. This is not working in J either
a
1 1 1 1 2 1 3 1

Anssi

At 06:18 17.7.1999 +0000, you wrote:
>Anssi.seppala@enease.fi writes:
> > PS. With Amend and From I am also remembering APL's selective assignment
> > that is something that J is missing. For example (v#a)=. 1 2 3 works with
> > APL, but not in J. Is this something that we will never see in J?
>
>Is this what you are asking for?
>
>  ('v e a')=. 1 2 3
>    v
>1
>    e
>2
>    a
>3
>
>/Gosi
>
>
>---------------------------------------------------------------------------
>-----
>J Forum: for information about this list, see
>http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 09:34:25 1999
Date: Sat, 17 Jul 1999 16:30:09 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
In-Reply-To: <005e01bed048$b9dc5540$9545fea9@t500>
References: 
 <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi>

Thanks Chris,
bx idiom was new for me.  Now I can do it by myself:

One of the things I have been missing in J is the possibility also include
ravel in selective assignment in any shape of array in APL:

((v='a')/,v)<- 'b'

In J this can now be done using your example but the result must be then
reshaped

($v)$'b' (bx 'a'=,v)} ,v

Anssi

PS. One of the difficulties learning J is to understant Amend. For example
I have tried many times to find such examples from the Phrases.

At 07:36 17.7.1999 -0400, you wrote:
>Now suppose you have a mask v=. 0 0 1 0 0 that indicates the position to
>replace with 100:
>
>APL:
>
>    A=: iota 5
>    (v/A) <- 100
>
>J:
>
>    100 (bx v)} i.5

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 09:34:27 1999
Date: Sat, 17 Jul 1999 16:09:06 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
In-Reply-To: <005d01bed048$b90b49a0$9545fea9@t500>
References: 

Chris,
Your examples are good and right. This was the original confusion:  {"1
works fine

    1 {"1 m=. i.4 4
1 5 9 13

but the changing of { to }  in the statement fails:
        a=. 10 50 90 13
        a 1}"1 m
|rank error
|   a     1}"1 m

OK. Now I know that defining Rank better I get the correspondence between 1
{"0 1 and 1}"0 1
    1 {"0 1 m=. i.4 4
1 5 9 13
         a 1}"0 1 m
  0 10  2  3
  4 50  6  7
  8 90 10 11
12 13 14 15

Anssi

At 07:35 17.7.1999 -0400, you wrote:
>3.
> >Thanks for the help. Obviously I expected too much correspondense bteween
> >Amend ( } ) and From ( { ).
>
>I dont understand this comment. There is a very close correspondence between
>amend and from. Where might the above examples contradict this?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 10:03:25 1999
From: "Chris Burke" 
References: <4.2.0.58.19990717145459.00aa15c0@pop.kolumbus.fi>
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Sat, 17 Jul 1999 10:02:25 -0400
	charset="iso-8859-1"

> No, I meant selective assignment  (at least in APL2 and APL/W) where you
> can combine the selection and assignment . For example using # or
 operator :
> a=. 1 1 1 1 1 1 1 1
> v=.  1 0 0 0 1 0 1 0
> (v#a)=. 1 2 3          NB. This is not working in today's J
> (0 4 6{a) =. 1 2 3   NB. This is not working in J either
> a
> 1 1 1 1 2 1 3 1

Well, I would say that the following do combine selection and assignment:

  a=. 1 2 3 (bx v)} a
  a=: 1 2 3 (0 4 6)} a

But I think what you are asking is whether the syntax supported by APL for
selective assignment is now or will be sometime supported by J. Right now,
it is not. I suppose it could be, though I think the existing J syntax works
well enough.

Note that the APL syntax is special. For example, consider (in Dyalog APL):

      a {gets} 1 2 3
      1 0 1/a
1 3
      (1 0 1/a) {gets} 10 20
      a
10 2 20

This is as expected. However, try to extend it to a list of names:

      a {gets} b {gets} c {gets} 1 2 3
      (({enclose} 1 0 1)/ {each} a b c)
 1 3  1 3  1 3
      (({enclose}1 0 1)/{each} a b c) {gets} (10 20)(30 40)(50 60)
SYNTAX ERROR

Why should this be a syntax error? It is a perfectly good extension of the
APL syntax, but not supported, almost certainly because the syntax was
special to start off with.

In J however, this is OK:

   'a b c'=: (10 20;30 40;50 60) (bx 1 0 1)}each a;b;c

   a;b;c
+-------+-------+-------+
|10 2 20|30 2 40|50 2 60|
+-------+-------+-------+

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 12:44:15 1999
Date: Sat, 17 Jul 1999 19:44:10 +0300
From: Anssi  Seppala 
Subject: Jforum: J Plot key with linepatterns
In-Reply-To: <003b01bed05d$66ee9400$9545fea9@t500>
References: <4.2.0.58.19990717145459.00aa15c0@pop.kolumbus.fi>

Can someone tell if it is possible to have linepatterns in plot key? I try
to make plots for b/w presentation and colour printing is not possible.

Trying this
    style=:'type line;color black;penstyle 0 1 2 3 4;key as1 as2 as3 as4'
    style plot i.3 4

The keys are black boxes, but I want samples of each line pattern with key
text.

Similar problem seems to be with bar charts. The only fill pattern is solid
with different colours. Of course I can use Excel + other tools, but hope
to hear how far the plot-utilites can be utilised .

Anssi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 16:33:45 1999
From: dajoy@hoy.net (Ajoy Victor)
Date: Sat, 17 Jul 1999 15:31:15 -0500
Content-type: text/plain; charset=US-ASCII
Subject: Jforum: adyacence to incidence matrix transformation

Is there any primitive to do this?
Is there any known phrase to do this?

I would first look at the archive but I can't find it. Is there an archive?

Second try:

Does anyone has a newer version of this FAQ?

Subject: J Programming FAQJ Programming FAQ
Compiled by: Ehud Lamm
Updated    : 3/6/95
Posted to  : comp.lang.apl
Changes    : example of under, obverse, trains, rank error.

--

I'm doing some work with genetic algorithms. Has someone developed some
tools or framework to program GAs (specially permutation crossover methods
but any idea will do) ?

I would appreciate any kind of advice.

Daniel

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 17:33:36 1999
Date: Sat, 17 Jul 1999 17:31:38 -0400
From: Kirk B Iverson 
Subject: Re: Jforum: J in HTML?
In-Reply-To: <008e01bed04e$dde12960$9545fea9@t500>
References: <378F618C.E10B2AA4@Home.Com>

At 08:21 1999-07-17 -0400, you wrote:
>
>> Can anyone tell me any easy way to display J output
>> (in this case it happens to be arrays of boxes of numbers) in an
>> HTML document, particularly when the recipient/reader of the document
>> is not likely to have ISIJ on their machine? My examples look just
>> fine on my machine (I can use either ISIJ or `Focal Point On-Line') but
>> when I look at them on other people's machines, the nice little boxes
>> become accented characters and superscripts and the like...
>
>One way in J403 is to save the output in a file (you can just save the ijx
>session), then use menu Run|HTML Publish...

 Holding the shift key down when you Run|HTML Publish will publish the active
session manager window.  The default schemes translate box-drawing characters
to ascii characters - | +.

/K

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 18:42:12 1999
From: Andromeda 
Subject: Jforum: FW: MAPI/11
Date: Sun, 18 Jul 1999 00:42:19 +0200

      I did some investigations with J403b and Windows 98 - and it seems to fall back
      against the mapi again - and it appears to be far worse as with my NT machine
      in the office.

     Only under some very wierd conditions (can't even reproduce them)  I can send
     attachments using the mapi script.
     Sending normal messages is not a problem ,but the attachments seem to be a real pain.

     Sending simple & existing Excel - Attachments is returned with error code 11
     (Attachment not found).
     But it does not only happen with Excel-Attachments - also simple text-attachments seem to be a problem.
     (same reason .. code 11)

      /Milan

-----Original Message-----
From:	Andromeda [SMTP:milan.ice@bluewin.ch]
Sent:	Donnerstag, 8. Juli 1999 18:54
To:	'Jforum'
Subject:	FW: MAPI/11

      Having thought twice about it when walking home - I'm more and more convinced that the MAPI
      stuff is ok - but it looks like a stack-problem somewhere when creating Excel-Spreadsheets the way
      described below (--> fn)

      By the way : each file below (file1.xls,file2.xls,file3.xls) should be sent different recepients.

      This is then done by:

      sess =: logonmail ''
      NB...  send the 3 mails
      logoffmail sess

      //..sorry about sending the mail a 3rd time

-----Original Message-----
From:	Ondrus, Milan [SMTP:milan.ondrus@csfb.com]
Sent:	Donnerstag, 8. Juli 1999 17:44
To:	'jforum'
Subject:	RE: MAPI/11

Forgot to say again:

I'm using J 4.03b under NT 4.0.
I use Excel 97 and Outlook-version 8.04.
/

-----Original Message-----
From:	Ondrus, Milan
Sent:	Donnerstag, 8. Juli 1999 17:32
To:	'jforum'
Subject:	MAPI/11

 There seems to be an annoying  problem with sending attachments.

 All that I can say is the following:

 Despite the fact that the Excel-File I want to send and is existing I get very often this DLL-error number
 11 back - which says that MAPI_E_ATTACHMENT_NOT_FOUND ..(see MSDN)  as Seymour Glass said before.

 Unfortunately I did not succeed in finding a logical rule(except maybe the one below)  what the conditions are when this happens.
 Sometimes all files are sent ,sometimes only one.

 This code 11 is correct when the file does not exist ,fine,but there must be an other condition when 11 is displayed
 but the file IS there.
 This behavior I found on both sorts of drives ,eg when the files are sitting on the hard-drive or a network drive.

 Just one important trace I found that might me a hint.

 I have a process that creates excelfiles out of  textfiles.
 Lets say I create :

 file1.xls
 file2.xls
 file3.xls

 After the creation there is no Excel task in the background and no locks are obvious.

 If I process them with the mapi-stuff in the opposite order ,then everything is sent.
 If I resent them all - then only the 1st is sent (file1.xls) - the rest comes back with 11.
 Only if I logoff from J and go back and reprocess it in the opposite order again - all are sent but as before
 resending them means again : only the 1st  is sent.

 If I send them in the same order - then allways only the 1st one is sent ,the rest comes back with error 11
 (Smashing down the J session does not help)

 One other important point.If I create a file directly under Excel ,this behaviour I did not perceive.So there it looks all right.

 So ,from this point of view I can't even say that there is sth wrong with the MAPI script ,but as it looks - there seems to be
 a problem with mapi when I automatically create a spread-sheet - but this creation of the excel-file itself works perfectly:

 fn =: 3 : 0

 xlopen''
 xlcmd'wb add'
 wbname=:xlget'temp name'
 xlget'temp worksheets'
 xlid'ws'
 xlget'ws item 1'
 wsname=:xlget'temp name'
 y. xlwriter wbname,' ',wsname,' 1 1'
 xlcmd 'temp saveas ',' xyzhhmmss.xls'
 xlquit''
)

 As this is happening I can't release it into production.

 thanks for any help ...........Milan Ondrus

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 17 20:24:42 1999
From: Eemcd@aol.com
Date: Sat, 17 Jul 1999 20:21:48 EDT
Subject: Re: Jforum: adyacence to incidence matrix transformation

dajoy asked for primitive or defined functions to convert from adjacence to
incidence matrix -- and though I don't recognize the terminology I think
these might do, from Section 4.B of the Phrase Book (available under Help):

m38=: <"1
m39=: $ #: (# i.@$)@,

m85=: 1: [`(m38@])`(m93@])} ]
m93=: 0: $~ [: >: >./

Given an incidence matrix q:

   q
0 1 0 1 0
0 1 1 1 0
1 1 0 0 1
1 0 0 0 0

A 2-column table giving the row/column indices of the 1s in q can be obtained
with m39:

   m39 q
0 1
0 3
1 1
1 2
1 3
2 0
2 1
2 4
3 0

and inversely, the incidence table can be derived from the index table by m85:

   m85 m39 q
0 1 0 1 0
0 1 1 1 0
1 1 0 0 1
1 0 0 0 0

Eugene McDonnell
Palo Alto

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 07:52:08 1999
From: Andromeda 
Subject: Jforum: FW: MAPI/11
Date: Sun, 18 Jul 1999 13:50:48 +0200

     ...was trying to find some conditions in Win98
     It looks like as it as soon it finds that code 11 it does not behave correctly.

     Example :  I did a test with a text-file:
                      The file exists on 2 drives:  e:\yyy.txt and d:\j403b\exp\july15\yyy.txt

                      the usual paranoia-check..

                      fread 'e:\yyy.txt' and fread 'd:\j403b\exp\july15\yyy.txt' gives in both cases the correct result

                      sess sendmail "the file under e" gives 0 -> attachm. sent
                      sess sendmail "another file under e" .. gives also 0 -> attachm. sent

                      but then ..
                      sess sendmail "the file under d" gives 11 --> attachm not sent
                      sess sendmail "the same file under e" gives 11 --> attachm not send.

                      logoffmail sess or reloading J does not help ...It always returns 11 back
                      even if I try   :   sess sendmail "the file under e" which I did at the beginning does not work
                      any longer.

                      Healthy logic would ask why it did work at all..but this is exactly the dilemma..
                      Its even not possible to say why it worked.

                     /Milan

      -----Original Message-----
From:	Andromeda [SMTP:milan.ice@bluewin.ch]
Sent:	Sonntag, 18. Juli 1999 00:42
To:	'Jforum'
Subject:	FW: MAPI/11

      I did some investigations with J403b and Windows 98 - and it seems to fall back
      against the mapi again - and it appears to be far worse as with my NT machine
      in the office.

     Only under some very wierd conditions (can't even reproduce them)  I can send
     attachments using the mapi script.
     Sending normal messages is not a problem ,but the attachments seem to be a real pain.

     Sending simple & existing Excel - Attachments is returned with error code 11
     (Attachment not found).
     But it does not only happen with Excel-Attachments - also simple text-attachments seem to be a problem.
     (same reason .. code 11)

      /Milan

-----Original Message-----
From:	Andromeda [SMTP:milan.ice@bluewin.ch]
Sent:	Donnerstag, 8. Juli 1999 18:54
To:	'Jforum'
Subject:	FW: MAPI/11

      Having thought twice about it when walking home - I'm more and more convinced that the MAPI
      stuff is ok - but it looks like a stack-problem somewhere when creating Excel-Spreadsheets the way
      described below (--> fn)

      By the way : each file below (file1.xls,file2.xls,file3.xls) should be sent different recepients.

      This is then done by:

      sess =: logonmail ''
      NB...  send the 3 mails
      logoffmail sess

      //..sorry about sending the mail a 3rd time

-----Original Message-----
From:	Ondrus, Milan [SMTP:milan.ondrus@csfb.com]
Sent:	Donnerstag, 8. Juli 1999 17:44
To:	'jforum'
Subject:	RE: MAPI/11

Forgot to say again:

I'm using J 4.03b under NT 4.0.
I use Excel 97 and Outlook-version 8.04.
/

-----Original Message-----
From:	Ondrus, Milan
Sent:	Donnerstag, 8. Juli 1999 17:32
To:	'jforum'
Subject:	MAPI/11

 There seems to be an annoying  problem with sending attachments.

 All that I can say is the following:

 Despite the fact that the Excel-File I want to send and is existing I get very often this DLL-error number
 11 back - which says that MAPI_E_ATTACHMENT_NOT_FOUND ..(see MSDN)  as Seymour Glass said before.

 Unfortunately I did not succeed in finding a logical rule(except maybe the one below)  what the conditions are when this happens.
 Sometimes all files are sent ,sometimes only one.

 This code 11 is correct when the file does not exist ,fine,but there must be an other condition when 11 is displayed
 but the file IS there.
 This behavior I found on both sorts of drives ,eg when the files are sitting on the hard-drive or a network drive.

 Just one important trace I found that might me a hint.

 I have a process that creates excelfiles out of  textfiles.
 Lets say I create :

 file1.xls
 file2.xls
 file3.xls

 After the creation there is no Excel task in the background and no locks are obvious.

 If I process them with the mapi-stuff in the opposite order ,then everything is sent.
 If I resent them all - then only the 1st is sent (file1.xls) - the rest comes back with 11.
 Only if I logoff from J and go back and reprocess it in the opposite order again - all are sent but as before
 resending them means again : only the 1st  is sent.

 If I send them in the same order - then allways only the 1st one is sent ,the rest comes back with error 11
 (Smashing down the J session does not help)

 One other important point.If I create a file directly under Excel ,this behaviour I did not perceive.So there it looks all right.

 So ,from this point of view I can't even say that there is sth wrong with the MAPI script ,but as it looks - there seems to be
 a problem with mapi when I automatically create a spread-sheet - but this creation of the excel-file itself works perfectly:

 fn =: 3 : 0

 xlopen''
 xlcmd'wb add'
 wbname=:xlget'temp name'
 xlget'temp worksheets'
 xlid'ws'
 xlget'ws item 1'
 wsname=:xlget'temp name'
 y. xlwriter wbname,' ',wsname,' 1 1'
 xlcmd 'temp saveas ',' xyzhhmmss.xls'
 xlquit''
)

 As this is happening I can't release it into production.

 thanks for any help ...........Milan Ondrus

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 10:31:47 1999
Date: Sun, 18 Jul 1999 10:29:18 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: OCXs in J
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id KAA14913

Hi everyone,

A few days ago I asked on this forum who uses OCXs and which ones. So far,
only Oleg responded. Does this mean that we are the only two users of OCXs?
Are there any more? Anyone willing to state that they do NOT use them, and
why?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 10:32:00 1999
Date: Sun, 18 Jul 1999 10:29:18 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: OCXs in J
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id KAA14906

Hi everyone,

A few days ago I asked on this forum who uses OCXs and which ones. So far,
only Oleg responded. Does this mean that we are the only two users of OCXs?
Are there any more? Anyone willing to state that they do NOT use them, and
why?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 11:43:04 1999
Date: Sun, 18 Jul 1999 17:53:22 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: OCXs in J
In-Reply-To: <199907181029_MC2-7D60-69E3@compuserve.com>

I do not use OCX with J. I have done instead several J DLL  OLEServers for
other applications. If I need to extend my own analysis I could use Excel &
PowerPoint in Windows. I suppose that if I want to extend my application
with OCX I loose the portability that pure J has over many operating systems.
Anssi

At 10:29 18.7.1999 -0400, you wrote:
>A few days ago I asked on this forum who uses OCXs and which ones. So far,
>only Oleg responded. Does this mean that we are the only two users of OCXs?
>Are there any more? Anyone willing to state that they do NOT use them, and
>why?
>
>---------------------------------------------------------------------------
>-----
>J Forum: for information about this list, see
>http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 13:19:48 1999
From: "John D. Baker" 
Subject: Re: Jforum: OCXs in J
Date: Sun, 18 Jul 1999 13:19:30 -0400
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit

I'll take the bait on that one.

I've tried OCX's in J and other environments, mostly VB and Delphi,
but I've pretty much given up on them for the time being.

I've given them a rest until they:

a) shrink  - many OCX's are larger than
J and almost always MUCH larger than anything you write in J.
There's something wrong with distributing a multi-megabyte beast
with your 30K J app.

b) become easiar to use - the effort required to learn the OCX command
language or all the damm.deep.object.calls.to.do.diddly is often
more tiresome that just writing a few more lines of J.

c) install with fewer problems - just because you managed to
get an OCX to work on one machine is no guarantee that it
will work on the next machine.  I've seen this very problem
on my basement network.

d) remain stable from version to version - components that
are embedded in other software have much higher compatibility
constraints than most software.  Many developers freeze their
OCX's at one version because bitter experience has shown that
the next version often breaks the app you've just sweated blood over.

e) become better suited to array languages - I haven't looked in
sometime but the last graph OCX I tried insisted on getting
it's data points one at call at a time.  It was only a mere two orders
of magnitude faster to use the little J plot package.  I'm sure
more current OCX's do a better job.

Despite my mostly negative experiences with OCX's I do see
their value when they are done well.  An excellent example of
a nice OCX is the the HTML rendering tool that the J4.03b publish
tool uses to display J scripts after they've been HTML'ized.   As time
go by OCX's, and other software modules, will get better but for now
use them with caution and only when the gain exceeds the pain.

John

-----Original Message-----
From: Alain Miville de Ch�ne 
Date: Sunday, July 18, 1999 10:33 AM
Subject: Jforum: OCXs in J

>Hi everyone,
>
>A few days ago I asked on this forum who uses OCXs and which ones. So far,
>only Oleg responded. Does this mean that we are the only two users of OCXs?
>Are there any more? Anyone willing to state that they do NOT use them, and
>why?
>
>-------------------------------------------------------------------------------
-
>J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 13:22:50 1999
From: "Oleg Kobchenko" 
References: <199907181029_MC2-7D60-69E3@compuserve.com>
Subject: Re: Jforum: OCXs in J
Date: Sun, 18 Jul 1999 20:22:30 +0300
	charset="ISO-8859-1"

Lately I have been working with various ActiveX controls
for user applications in VB environment. I came to the
conclusion that currently client companies tend to emphasize
user interface presentation in addition to functionality.
Such things as MS Office look-and-feel, toolboxes, property
editors are important. With OLAP brought to scene, user controls
with sofisticated presentation logic are required.

J has most of the prerequisites to handle ActiveX controls.
However, since the time this possibility was introduced to
J system, qualities of ActiveX controls largely evolved and
now exceed J's capacity to handle their certain features.
Most essential are these:

1. Positional binding with possible contraction of the form
   client area, e.g.
   - status bar glued to the bottom of form
   - tool bar at the top
   - panels that stick to sides of form
   - coolbar et al, can even change size (height) dynamically
     at run time, hence client area adjustment is required

2. Control position and sizing. J has build-in layout mechanism
   that can define behavior of controls positions with the change
   of form size, making it easy to maintain layouts. However, this
   inhibits programatical posional adjustments at run time. Plus,
   there isn't such event and form_resize.

3. Mutual awareness of controls: incapsulation, hierarchy, referencing
   E.g. many controls know whether there is an ImageList control on
   the form at design time. Panels or tabs can have their own client area
   and position inner controls appropriately.

4. Property sheets, property dialogs, designers for ActiveX controls.
   These are common ways to set-up controls at design time. Existence
   of each of these 3 methods varies between controls.

I have not specifically looked into the issue of component container
design. It might indeed be not a trivial task. But it's a shame that
such powerful system as J looses the competition over such matters as
mere presentation. There is nothing exciting about VB, except for the
user interface. But it acts as the dress code for the party of contemporary
business applications.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 13:46:36 1999
From: dajoy@hoy.net (Ajoy Victor)
Date: Sun, 18 Jul 1999 12:43:42 -0500
Content-type: text/plain; charset=US-ASCII
Subject: Re: Jforum: adyacence to incidence matrix transformation

 _____ _                 _
|_   _| |__   __ _ _ __ | | _____
  | | | '_ \ / _` | '_ \| |/ / __|
  | | | | | | (_| | | | |   <\__ \_
  |_| |_| |_|\__,_|_| |_|_|\_\___(_)

Daniel

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 13:46:38 1999
From: dajoy@hoy.net (Ajoy Victor)
Date: Sun, 18 Jul 1999 12:43:42 -0500
Content-type: text/plain; charset=US-ASCII
Subject: Re: Jforum: adyacence to incidence matrix transformation

> dajoy asked for primitive or defined functions to convert from adjacence to
> incidence matrix -- and though I don't recognize the terminology I think
> these might do, from Section 4.B of the Phrase Book (available under Help):
>
> m38=: <"1
> m39=: $ #: (# i.@$)@,
>
> m85=: 1: [`(m38@])`(m93@])} ]
> m93=: 0: $~ [: >: >./
>
> Given an incidence matrix q:
>
>    q
> 0 1 0 1 0
> 0 1 1 1 0
> 1 1 0 0 1
> 1 0 0 0 0

I don't see how q can be an incidence matrix since

+-+---------+
| |0 1 2 3 4|
+-+---------+
|a|0 1 0 1 0|
|b|0 1 1 1 0|
|c|1 1 0 0 1|
|d|1 0 0 0 0|
+-+---------+

would mean that:
side a touches node 1 and node 4 [ok]
side b touches node 1, node 2, and node 3 [??, at least weird]
...

So, for my peace of mind, lets use this q

   [q=:0 1 0 1 0 , 0 1 1 0 0, 0 1 0 0 1,:1 0 0 0 0
0 1 0 1 0
0 1 1 0 0
0 1 0 0 1
1 0 0 0 0

Side d is OK, it is a loop in node 0 so it touches only node 0.

I'll look like this:

[2]--b--[1]--a--[3]--c--[4]

[0]--+
 |   |
 +---d

+-+---------+
| |0 1 2 3 4|
+-+---------+
|a|0 1 0 1 0|
|b|0 1 1 0 0|
|c|0 1 0 0 1|
|d|1 0 0 0 0|
+-+---------+

But the adjacence matrix for thi graph would be

+-+---------+
| |0 1 2 3 4|
+-+---------+
|0|1 0 0 0 0|
|1|0 0 1 1 0|
|2|0 0 0 0 0|
|3|0 0 0 0 1|
|4|0 0 0 0 0|
+-+---------+

there is a path between:
node 0 and node 0
node 1 and node 2
node 1 and node 3
node 4 and node 4

The PROBLEM follows:

How do I transform this:

1 0 0 0 0
0 0 1 1 0
0 0 0 0 0
0 0 0 0 1
0 0 0 0 0

into this:

0 1 0 1 0
0 1 1 0 0
0 1 0 0 1
1 0 0 0 0

Maybe using creating somehow:

0 1
0 3
1 1
1 2
2 1
2 4
3 0

but not using 'm39 q' because q is what I want to get.

so that

> and inversely, the incidence table can be derived from the index table by m85:
>
>    m85 m39 q

0 1 0 1 0
0 1 1 0 0
0 1 0 0 1
1 0 0 0 0

Anyway, I only wanted to know if there was something already done. I thought
adjacence-to-incidence was a general (useful?) enough transformation to be
subject to the developers' thoughts. :)

Now that I notice the phrase book doesn't have a section about graphs.

shortest path
maximal flux
critical path
etc.

Daniel

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 15:27:25 1999
Date: Sun, 18 Jul 1999 15:23:40 -0400
From: Eric Iverson 
Organization: Iverson Software Inc.
Subject: Jforum: hybrid applications - VB/C++/Java and J
References: <199907181029_MC2-7D60-69E3@compuserve.com> <003601bed142$1e70d3b0$0801a8c0@spin.kharkov.ua>

Oleg Kobchenko wrote:
> ... There is nothing exciting about VB, except for the UI.
> But it acts as the dress code for the party of contemporary
> business applications.

The J UI is adequate for applications with modest UI requirements. It is
not adequate for applications that require the latest UI sizzle. We've
put a lot of effort into providing a 90% GUI solution. The cost of
providing the remaining 10% remains too high. The problem is that VB
won't stand still. We could put in the significant effort to catch up to
where they are today. But by then they'd have moved and we'd still be
behind. I don't like being in a race I am guaranteed to lose.

If your applications really require the latest GUI, then you should use
the best GUI tools available (C++, VB, OCX,...). But you probably don't
want to use them to do serious data processing (DP). If your application
needs sizzling GUI and serious DP, it is fortunate indeed that you can
use the J COM server to get the best of both worlds. Learning to program
the GUI in tools like VB is actually fairly easy and J and APL
programmers should not be so reluctant to take that bull by the horns.
And developing and debugging hybrid applications that use VB (or
whatever) for the GUI and J for DP works very well. In fact, I would
argue that for big applications it imposes a rigourous line between GUI
and DP that is to the benefit of all, and allows the use of the best
tool for each part of the job.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 16:15:27 1999
Date: Sun, 18 Jul 1999 16:13:34 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <378F618C.E10B2AA4@Home.Com> <3.0.5.32.19990717173138.00919b90@mail.interlog.com>

I appreciate the attempts to help, but am still having a problem for
a couple of reasons.

(1) Apparently Run|HTML Publish must be something new in J of some
relatively recent origin, so I don't have access to it, and thus
don't really have any idea of what it does.

(2) I'm actually not so much interested in solutions that involve
having J generate things differently (if this is what HTML Publish does)
as for expository purposes I want to be able to generate quite `literal'
`in' and `out' scripts. As a result I need to be able to solve the
problem of processing a perfectly normal J output file into some form
that will reproduce decently in NetScape, Explorer, Opera, Amaya and any
 other `normal' browsers, as it does, by the way, if you do have
access to `ISIJ' or `Focal Point On-Line' fonts.

(3) I am beginning to guess that I'll probably need to interpose PERL
into the process anyway, so I suppose I could have PERL rewrite the
J output into some form that would use  table facilities. That
wouldn't be pretty, but I suppose it would work, and be a general
solution, not requiring any `hand treatment' during the editing and
refinement process.

Kirk B Iverson wrote:
>
> At 08:21 1999-07-17 -0400, you wrote:
> >
> >> Can anyone tell me any easy way to display J output
> >> (in this case it happens to be arrays of boxes of numbers) in an
> >> HTML document, particularly when the recipient/reader of the document
> >> is not likely to have ISIJ on their machine? My examples look just
> >> fine on my machine (I can use either ISIJ or `Focal Point On-Line') but
> >> when I look at them on other people's machines, the nice little boxes
> >> become accented characters and superscripts and the like...
> >
> >One way in J403 is to save the output in a file (you can just save the ijx
> >session), then use menu Run|HTML Publish...
>
>  Holding the shift key down when you Run|HTML Publish will publish the active
> session manager window.  The default schemes translate box-drawing characters
> to ascii characters - | +.
>
> /K
>
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 17:23:38 1999
From: "Randy MacDonald" 
References: <378F618C.E10B2AA4@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Sun, 18 Jul 1999 17:21:05 -0400
	charset="iso-8859-1"

----- Original Message -----
From: David Ness 
Sent: July 16, 1999 12:45 PM
Subject: Jforum: J in HTML?

> Can anyone tell me any easy way to display J output
> (in this case it happens to be arrays of boxes of numbers) in an
> HTML document, particularly when the recipient/reader of the document

If the problem is merely with the display of the line-draw characters
around boxed arrays, what is the problem with changing the display
characters
to ASCII, via the 9!:6 mechanism.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 17:32:58 1999
Date: Sun, 18 Jul 1999 17:31:01 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <378F618C.E10B2AA4@Home.Com> <001101bed163$722be780$9de27018@lndn1.on.wave.home.com>

I must miss the point of your question. At the moment, things
get written into files and look OK if you pick them up in any
character oriented (monospaced, of coure) editor. They also look fine
inserted into  if the user has ISIJ or "Focal Point On-Line" (to
name two that I have found) fonts. However, if you do not have either
of those fonts, then the  display will produce ugly looking
lines of accented capital characters, obviously assuming that the
characters in the file are from some (Latin, I suppose) Character Set.

I am trying to find some way to tell the browsers interpreting the
HTML about the Character set of the data in the inserted boxed tables.
I don't see how the 9!:6 mechanism has anything to do with my problem.

Randy MacDonald wrote:
>
> If the problem is merely with the display of the line-draw characters
> around boxed arrays, what is the problem with changing the display
> characters
> to ASCII, via the 9!:6 mechanism.
>
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 18:14:33 1999
From: "Chris Burke" 
References: <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi> <4.2.0.58.19990717161202.00ab1680@pop.kolumbus.fi>
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Sun, 18 Jul 1999 18:14:54 -0400
	charset="iso-8859-1"

> One of the things I have been missing in J is the possibility also include
> ravel in selective assignment in any shape of array in APL:
>
> ((v='a')/,v)<- 'b'
>
> In J this can now be done using your example but the result must be then
> reshaped
>
> ($v)$'b' (bx 'a'=,v)} ,v
>
> Anssi

Try:

   [ v=: 3 4$'adamannemary'
adam
anne
mary

   f=: bx @: ('a'&=) @ , @ ]

   f v
0 2 4 9

   'Z' f} v
ZdZm
Znne
mZry

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 18:27:49 1999
From: "Randy MacDonald" 
References: <378F618C.E10B2AA4@Home.Com> <001101bed163$722be780$9de27018@lndn1.on.wave.home.com> <37924794.F001A5AA@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Sun, 18 Jul 1999 18:25:20 -0400
	charset="iso-8859-1"

You speak of "ugly looking lines of accented capital characters." My
impression is that with the ISIJ font in place, the characters appear
as line draw characters.  These line draw characters are the default
characters used to display boxed arrays. This leaves you with two options
you can use immediately:

1. Make the ISIJ font available to those that need it. This may not
   be possible due to copyright issues, I agree.
2. Change the characters that are used to display boxed arrays to ones
   that are more universal.  This is what the 9!:6 would be used to do:

   (9!:7) 1 0 2 3 4 5 { 9!:6 ''   NB. Switch first two rows.

   At this point all boxed arrays will display with plus and dash symbols.
   1 2 $ < 1 2
+-----+-----+
| 1 2 | 1 2 |
+-----+-----+

My question, again, is: what is the problem with doing this?

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 5:31 PM
Subject: Re: Jforum: J in HTML?

> I must miss the point of your question. At the moment, things
> get written into files and look OK if you pick them up in any
> character oriented (monospaced, of coure) editor. They also look fine
> inserted into  if the user has ISIJ or "Focal Point On-Line" (to
> name two that I have found) fonts. However, if you do not have either
> of those fonts, then the  display will produce ugly looking
> lines of accented capital characters, obviously assuming that the
> characters in the file are from some (Latin, I suppose) Character Set.
>
> I am trying to find some way to tell the browsers interpreting the
> HTML about the Character set of the data in the inserted boxed tables.
> I don't see how the 9!:6 mechanism has anything to do with my problem.
>
> Randy MacDonald wrote:
> >
> > If the problem is merely with the display of the line-draw characters
> > around boxed arrays, what is the problem with changing the display
> > characters
> > to ASCII, via the 9!:6 mechanism.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 18:35:32 1999
Date: Sun, 18 Jul 1999 18:33:29 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <378F618C.E10B2AA4@Home.Com> <001101bed163$722be780$9de27018@lndn1.on.wave.home.com> <37924794.F001A5AA@Home.Com> <001401bed16c$6bc50620$9de27018@lndn1.on.wave.home.com>

The problem is that it is too ugly. This particular output is used in
a context where if it looks lousy, it might as well not be there.
It would be easier and simpler just to leave J out of the description.
While showing boxes with '-', '+' ... may be ok with programmers,
I find it too ugly to be useful for others.

Thanks, anyway, for trying to help.

Randy MacDonald wrote:
>
> You speak of "ugly looking lines of accented capital characters." My
> impression is that with the ISIJ font in place, the characters appear
> as line draw characters.  These line draw characters are the default
> characters used to display boxed arrays. This leaves you with two options
> you can use immediately:
>
> 1. Make the ISIJ font available to those that need it. This may not
>    be possible due to copyright issues, I agree.
> 2. Change the characters that are used to display boxed arrays to ones
>    that are more universal.  This is what the 9!:6 would be used to do:
>
>    (9!:7) 1 0 2 3 4 5 { 9!:6 ''   NB. Switch first two rows.
>
>    At this point all boxed arrays will display with plus and dash symbols.
>    1 2 $ < 1 2
> +-----+-----+
> | 1 2 | 1 2 |
> +-----+-----+
>
> My question, again, is: what is the problem with doing this?
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 18:45:44 1999
From: "Randy MacDonald" 
References: <378F618C.E10B2AA4@Home.Com> <001101bed163$722be780$9de27018@lndn1.on.wave.home.com> <37924794.F001A5AA@Home.Com> <001401bed16c$6bc50620$9de27018@lndn1.on.wave.home.com> <37925639.840E9EE2@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Sun, 18 Jul 1999 18:43:37 -0400
	charset="iso-8859-1"

This leaves the other solution.  Is this acceptable to interested
parties?

----- Original Message -----
From: David Ness 

> > 1. Make the ISIJ font available to those that need it. This may not
> >    be possible due to copyright issues, I agree.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 19:52:08 1999
From: gosi@centrum.is
Subject: Re: Jforum: J in HTML?
Date: Sun, 18 Jul 99 23:48:56 +0000

I prefer PDF to HTML when distributing text and pictures.
Then they look the same everywhere and no need to
distribute fonts or anything else along with it.

All you need is a reader which is free and used by millions
of people.

/Gosi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 20:39:22 1999
Date: Sun, 18 Jul 1999 20:37:25 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <379267e800bd001@blik.skima.is>

I was a beta tester for Adobe back in the early days of PDF. While
I use it, I don't like it, and, of course, it doesn't work to serve
web pages anyway, so isn't useful in that regard.

gosi@centrum.is wrote:
>
> I prefer PDF to HTML when distributing text and pictures.
> Then they look the same everywhere and no need to
> distribute fonts or anything else along with it.
>
> All you need is a reader which is free and used by millions
> of people.
>
> /Gosi
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 20:41:59 1999
Date: Sun, 18 Jul 1999 20:39:51 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <378F618C.E10B2AA4@Home.Com> <001101bed163$722be780$9de27018@lndn1.on.wave.home.com> <37924794.F001A5AA@Home.Com> <001401bed16c$6bc50620$9de27018@lndn1.on.wave.home.com> <37925639.840E9EE2@Home.Com> <003601bed16e$f9ae25a0$9de27018@lndn1.on.wave.home.com>

It may be. I have to figure out how to deliver a font in an HTML
document, and then check with my user community to see if anyone will
resent getting fonts downloaded. Then I will have to see if I can find
a font which can be used without copyright getting in the way.

Randy MacDonald wrote:
>
> This leaves the other solution.  Is this acceptable to interested
> parties?
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 21:19:46 1999
Date: Sun, 18 Jul 1999 21:17:40 -0400
From: David Ness 
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
References: <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi> <4.2.0.58.19990717161202.00ab1680@pop.kolumbus.fi> <006c01bed16b$2c686ea0$c0c7fea9@t500>

Chris Burke wrote:
>
> ...
>
>    f=: bx @: ('a'&=) @ , @ ]
>

Chris, I see that this works, but it would be very helpful to me if
you could say a word or two about `why' the `At' and why the
`Same-Right'. I find I can often follow your elegant constructs, but
I still have great trouble generating them myself, and I'm looking for
clues on how you `see' things that results in such nice encodings...

Thanks for any insights you can offer...

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 23:02:52 1999
Date: Sun, 18 Jul 1999 22:59:45 -0400
From: Murray Eisenberg 
Organization: Mathematics & Statistics, Univ. of Mass./Amherst
Subject: Re: Jforum: J in HTML?
References: <379267e800bd001@blik.skima.is>

PDF is hardly infallible:  Just try to get Acrobat Reader (3.0.1 or 4.0)
to correctly PRINT to an HP PCL printer any .pdf documents created from
Postscript output of TeX that include symbols coming from the code range
of control characters.  For example, TeX's large integral and summation
signs don't print correctly there.

Adobe is aware of this problem but has not taken steps to fix it.  They
offer a supposed workaround, namely, to Print As Image; unfortunately,
this gives a grossly inferior result, with ordinary letters being
broken.

Pardon me if I've posted such comments before to this newsgroup.

gosi@centrum.is wrote:
>
> I prefer PDF to HTML when distributing text and pictures.
> Then they look the same everywhere and no need to
> distribute fonts or anything else along with it.
>
> All you need is a reader which is free and used by millions
> of people.
>
> /Gosi
>
> --------------------------------------------------------------------------------
> J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

--
Murray Eisenberg                     murray@math.umass.edu
Mathematics & Statistics Dept.       phone 413 549-1020 (H)
Univ. of Massachusetts                     413 545-2859 (W)
Amherst, MA 01003-4515

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 23:33:30 1999
Date: Sun, 18 Jul 1999 23:31:33 -0400
From: David Ness 
Subject: Re: Jforum: J in HTML?
References: <379267e800bd001@blik.skima.is> <379294A1.91813C5F@math.umass.edu>

I'd go substantially further than `hardly infallible'. I did a study
for some major NYC financial institutions around the time that PDF
was a `new vogue' (perhaps `new vague' would be better). The results
weren't pretty. Most of Adobe's claims about PDF were, to be _very_
charitable, `misleading'. It was trivially easy to produce documents
that were _horrible_ when rendered in PDF, and there were few
circumstances that we could find where PDF was superior to simple
`compressed' PostScript. unless people bothered to use the document
structuring and `snapshot' capabilities of PDF, which virtually no one
ever did.

Any notion that PDF offered a `simple better solution' to the document
exchange problem was best sold, IMO, to the `simple minded'...

Murray Eisenberg wrote:
>
> PDF is hardly infallible:  Just try to get Acrobat Reader (3.0.1 or 4.0)
> to correctly PRINT to an HP PCL printer any .pdf documents created from
> Postscript output of TeX that include symbols coming from the code range
> of control characters.  For example, TeX's large integral and summation
> signs don't print correctly there.
>
> Adobe is aware of this problem but has not taken steps to fix it.  They
> offer a supposed workaround, namely, to Print As Image; unfortunately,
> this gives a grossly inferior result, with ordinary letters being
> broken.
>
> Pardon me if I've posted such comments before to this newsgroup.
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 23:38:40 1999
Date: Sun, 18 Jul 1999 23:35:58 -0400
From: Fraser Jackson 
Subject: Jforum: J Plot key with linepatterns
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id XAA05979

I have found I get lines when I use the keystyle option with 2 in the first
location. The second parameter seems to modify the location of the key.
Perhaps the documentation could be clarified.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 23:46:41 1999
Date: Sun, 18 Jul 1999 23:44:02 -0400
From: David Ness 
Subject: Jforum: Does it strike anyone that ...

... not mentioning what x and y might stand for in, for example, the
definition of `x 17!:1 y' as `Matches. Find all matches' is somewhat
less than adequate?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sun Jul 18 23:55:46 1999
From: "Randy MacDonald" 
References: <37929F02.2BC07C2F@Home.Com>
Subject: Re: Jforum: Does it strike anyone that ...
Date: Sun, 18 Jul 1999 23:53:10 -0400
	charset="iso-8859-1"

The script \SYSTEM\MAIN\REGEX.IJS details the calling convention
for this verb, renamed "rxmatches" in the script...

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 11:44 PM
Subject: Jforum: Does it strike anyone that ...

> ... not mentioning what x and y might stand for in, for example, the
> definition of `x 17!:1 y' as `Matches. Find all matches' is somewhat
> less than adequate?
>
> --------------------------------------------------------------------------
------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 00:12:29 1999
Date: Mon, 19 Jul 1999 00:10:04 -0400
From: David Ness 
Subject: Re: Jforum: Does it strike anyone that ...
References: <37929F02.2BC07C2F@Home.Com> <00ab01bed19a$37eb7860$9de27018@lndn1.on.wave.home.com>

I thought, perhaps mistakenly, that whenever I read descriptions of
the `Foreign Conjunction', when x and/or y appeared in the definition
but not in the text, its value was `irrelevant' as in `6!:0 y' being
defined `Current ...' with no further reference to `y'. When one saw
`[x] 6!:2 y' the definition was something like `seconds to exectute
sentence y (mean of x times with default once).'

I would have thought that the purpose of help was to be helpful.
But then, maybe its purpose is to make you track down things that
could otherwise be said quite simply. Otherwise, why bother with the
`help'?

In either event, I'd guess it was sensible to be consistent, and
either have one page covering all of the foreigns, just pointing
at the scripts that use them (17!: ... see regex.ijs, 18!: ... see ...)
or, alternatively, actually bother to define the foreigns thus making
a rather complicated (at least for some of us) loading and searching
of a script unnecessary to establish a simple fact.

I guess I'd choose the latter. It seems simpler to me. But then, as
usual, YMMV.

Randy MacDonald wrote:
>
> The script \SYSTEM\MAIN\REGEX.IJS details the calling convention
> for this verb, renamed "rxmatches" in the script...
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 00:29:44 1999
From: "Randy MacDonald" 
References: <37929F02.2BC07C2F@Home.Com> <00ab01bed19a$37eb7860$9de27018@lndn1.on.wave.home.com> <3792A51C.944E385D@Home.Com>
Subject: Re: Jforum: Does it strike anyone that ...
Date: Mon, 19 Jul 1999 00:27:06 -0400
	charset="iso-8859-1"

Did looking at the specific script not give you the answer you sought?
A windows "Find" with "text=17!:..." was sufficient to find the half-dozen
scripts containing the phrase, and an eyeball of the files was enough
to find a documentary reference, in the third one I looked at, in fact.

Did the running of the "regular expression" lab give you no insight
either? (I'm assuming, of course that you were aware that the (17!:)
functions were all related to regular expressions.)

Perhaps your apparent and overarching problem of "these J help files
suck" should have been stated earlier.  Given that they represent
a repurposing of a printed document into the Windows Help file format,
their currency is not expected, at least not by me.  The Labs seem to
have taken over a bit of the tutorial aspects of J, expecially for
bleeding-edge packages.

----- Original Message -----
From: David Ness 
Sent: July 19, 1999 12:10 AM
Subject: Re: Jforum: Does it strike anyone that ...

> I thought, perhaps mistakenly, that whenever I read descriptions of
> the `Foreign Conjunction', when x and/or y appeared in the definition
> but not in the text, its value was `irrelevant' as in `6!:0 y' being
> defined `Current ...' with no further reference to `y'. When one saw
> `[x] 6!:2 y' the definition was something like `seconds to exectute
> sentence y (mean of x times with default once).'
>
> I would have thought that the purpose of help was to be helpful.
> But then, maybe its purpose is to make you track down things that
> could otherwise be said quite simply. Otherwise, why bother with the
> `help'?
>
> In either event, I'd guess it was sensible to be consistent, and
> either have one page covering all of the foreigns, just pointing
> at the scripts that use them (17!: ... see regex.ijs, 18!: ... see ...)
> or, alternatively, actually bother to define the foreigns thus making
> a rather complicated (at least for some of us) loading and searching
> of a script unnecessary to establish a simple fact.
>
> I guess I'd choose the latter. It seems simpler to me. But then, as
> usual, YMMV.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 00:33:32 1999
Date: Mon, 19 Jul 1999 00:31:29 -0400
From: David Ness 
Subject: Re: Jforum: Does it strike anyone that ...
References: <37929F02.2BC07C2F@Home.Com> <00ab01bed19a$37eb7860$9de27018@lndn1.on.wave.home.com> <3792A51C.944E385D@Home.Com> <00b801bed19e$f5862600$9de27018@lndn1.on.wave.home.com>

Since you obviously can't help with this, why not butt out?

Randy MacDonald wrote:
>
> Did looking at the specific script not give you the answer you sought?
> 
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 00:55:28 1999
From: "Randy MacDonald" 
References: <378F618C.E10B2AA4@Home.Com> <3.0.5.32.19990717173138.00919b90@mail.interlog.com> <3792356E.916AC6A7@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Mon, 19 Jul 1999 00:53:05 -0400
	charset="iso-8859-1"

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 4:13 PM
Subject: Re: Jforum: J in HTML?

> I appreciate the attempts to help, but am still having a problem for
> a couple of reasons.
>
> (1) Apparently Run|HTML Publish must be something new in J of some
> relatively recent origin, so I don't have access to it, and thus
> don't really have any idea of what it does.

If, as Kirk stated, you don't have J4.03 then you may not have this
feature.

> (2) I'm actually not so much interested in solutions that involve
> having J generate things differently (if this is what HTML Publish does)
> as for expository purposes I want to be able to generate quite `literal'
> `in' and `out' scripts. As a result I need to be able to solve the
> problem of processing a perfectly normal J output file into some form
> that will reproduce decently in NetScape, Explorer, Opera, Amaya and any
>  other `normal' browsers, as it does, by the way, if you do have
> access to `ISIJ' or `Focal Point On-Line' fonts.

One solution is to , wherever you need boxed output is to use a verb
to convert the output into html, and to manually edit out the use of that
verb
in that file.

    NB. untested code, but close to what you want.

    htdisp 1 2 $ < 1 2

1 21 2


    tabletag =: (''&,)@:(,&'
')
    rowtag =: (''&,)"1
    celltag =: (''&,each)"0

    htdisp =:   tabletag@: ; @: rowtag @: celltag @: ":

    NB. and if your boxes have boxes, you'll have to recurse.

> (3) I am beginning to guess that I'll probably need to interpose PERL
> into the process anyway, so I suppose I could have PERL rewrite the
> J output into some form that would use  table facilities.

the Perl would, of course be much harder to write, and just another
thing that could go wrong.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 01:12:42 1999
From: "Randy MacDonald" 
References: <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi> <4.2.0.58.19990717161202.00ab1680@pop.kolumbus.fi> <006c01bed16b$2c686ea0$c0c7fea9@t500> <37927CB4.361760AC@Home.Com>
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Mon, 19 Jul 1999 01:10:12 -0400
	charset="iso-8859-1"

Explicitly

     f x ==  bx 'a'= , x
         ==  bx ('a'&=) , x
         ==  ( bx @: ('a' & = ) @ , @ ] ) x
     f   ==  ( bx @: ('a' & = ) @ , @ ]

using the idea that

     f g x == (f @: g @: ]) x
     f g h x == (f @: g @: h @: ]) x

transforms a composition of functions into a list with an unconventional
connector and a placeholder.

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 9:17 PM
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?

> Chris Burke wrote:
> >
> > ...
> >
> >    f=: bx @: ('a'&=) @ , @ ]
> >
>
> Chris, I see that this works, but it would be very helpful to me if
> you could say a word or two about `why' the `At' and why the
> `Same-Right'. I find I can often follow your elegant constructs, but
> I still have great trouble generating them myself, and I'm looking for
> clues on how you `see' things that results in such nice encodings...
>
> Thanks for any insights you can offer...

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 05:57:09 1999
From: gosi@centrum.is
Subject: Re: Jforum: J in HTML?
Date: Mon, 19 Jul 99 07:53:08 +0000

David Ness wrote:
> Any notion that PDF offered a `simple better solution' to the document
> exchange problem was best sold, IMO, to the `simple minded'...

Yes that is it. We simple people like it. And there are so
many of us.

I started to use PDF because I was asked to find a solution to publishing a
paper about y2k progress. It was being
printed in colour and distributed on paper to a lot of
people every month at a considereable expense.

I was asked to see if I could get it to the web using
HTML. I tried to use MSWord and it produced several
hundred BMP files and various other stuff. Most of all
it was not very good quality when browsed in IE or NS
even if the pictures looked ok in a BMP editor.

I discovered Adobe and we are very happy with it and
there are no more need for distributing printed copies.
It takes a few minutes to create and only one file and
it is easy to send it in mail as well as putting it on the web.

I have used it in several similar instances since. One
recent was when I was invited to a meeting where there
were people discussing creating advanced web structures
for a project which needed to distribute text quickly out
to all the citicens af Reykjavik and get responses back.

The web creators wanted to do some fancy stuff which
would occupy several people for some weeks. The
guy presenting it got absolutely furious when I suggested
trying PDF as a start. As it happens the whole web
effort was more or less dropped besides publishing
the PDF files and asking for e-mail about its contents.

I can understand why the web expert was furious. He has
obviously seen several projects go to PDF.

The simple solutions to simple people are probably
easier to sell than the symbol solutions.

/Gosi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 08:01:40 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
Date: Mon, 19 Jul 1999 07:59:15 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <006c01bed16b$2c686ea0$c0c7fea9@t500>

I have nothing to add to Chris's explanations, but
I'll offer this anyway:

} mystified me for a long time too.  The hardest thing
for me was figuring out how to apply a verb to selected
cells of a list, in place.  (Often that's what you
really want, rather than a scattered assignment)

I now use the following conjunction to do it, and I
don't have to think about it any more:

NB. Conjunction.  u. is verb, n. (or [x.] v. y.) is arg to {
NB.   to select s = desired portion of y.
NB. The result of x. u. s (if dyad) or u. s (if monad) replaces s
xuvy_z_ =: ([.`].) `:6  NB. x u v y
applytoselectionm_z_ =: ([. @ (].&{)) (]: ]: ]) (].})
applytoselectiond_z_ =: ([. xuvy (].&{)) (]: ]: ]) (].})

You pick the m form for monadic, or d for dyadic.  So:
   - applytoselectionm (<2 2) i. 3 3
0 1  2
3 4  5
6 7 _8

   20 - applytoselectiond (<2 2) i. 3 3
0 1  2
3 4  5
6 7 12

If you don't want to be troubled to decide whether the
verb is monadic or dyadic, you can use the bivalent form:
bivalent_z_ =: [. ^: (1:`((]`].)`:6))  NB. u. v. y. if monad, x. u. (v. y.)
if dyad
applytoselection_z_ =: ([. bivalent (].&{)) (]: ]: ]) (].})

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 10:45:06 1999
From: Andrew Nikitin 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
Date: Mon, 19 Jul 1999 09:42:08 -0500

> -----Original Message-----
> From:	Chris Burke [SMTP:cdburke@interlog.com]
> Sent:	Sunday, July 18, 1999 17:15
>
>    f=: bx @: ('a'&=) @ , @ ]
>
>    f v
> 0 2 4 9
>
>    'Z' f} v
> ZdZm
> Znne
> mZry
>
Though I am aware of this special case of }, but I am still curious how
J distinguishes this "linear" addressing from usual case? What if I want
the function argument opf the } adverb to return line numbers instead of
linear indices? Is the sophisticated boxed syntax my only option?

BTW, Dictionary says about } :
	u} is defined in terms of the noun case m}, the verb u applying
to the argument or arguments to provide the numeric indices required by
it.

This is at least not true. Consider

   100 (1 2)} i. 3 3
  0   1   2
100 100 100
100 100 100
   100 (1 2"_)} i. 3 3
0 100 100
3   4   5
6   7   8

Probably I have misinterpreted the word 'numeric indices'. Does it mean
the same as just 'indices'?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 11:23:52 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
Date: Mon, 19 Jul 1999 11:20:49 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

The linear-addressing of } was in J7, wasn't it?  I thought
it had been superseded by the new form in J2.

The Dictionary suggests that x u} y is equivalent to
x (x u y)} y.  (At least that's how I read it).  But
Andrew's example shows that this is not the case.
The example under u} shows this too, but I hadn't read
that closely until this thread came along.

Perhaps the documentation of u}, which matches the J7
form (if I remember rightly), was not updated when m}
was changed.

Is the current implementation of u} the way it's supposed
to work?  I think that it is better design to have
x u} y   equivalent to  x (x u y)} y .  Then, if I want
to support linear addressing, I can do it by
x ($@] <@#: u)} y .  It's odd to have linear addressing
supported in just this one variant of one primitive.

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
> Sent: Monday, July 19, 1999 10:42
> To: 'forum@jsoftware.com'
> Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
>
>
> > -----Original Message-----
> > From:	Chris Burke [SMTP:cdburke@interlog.com]
> > Sent:	Sunday, July 18, 1999 17:15
> >
> >    f=: bx @: ('a'&=) @ , @ ]
> >
> >    f v
> > 0 2 4 9
> >
> >    'Z' f} v
> > ZdZm
> > Znne
> > mZry
> >
> Though I am aware of this special case of }, but I am still curious how
> J distinguishes this "linear" addressing from usual case? What if I want
> the function argument opf the } adverb to return line numbers instead of
> linear indices? Is the sophisticated boxed syntax my only option?
>
> BTW, Dictionary says about } :
> 	u} is defined in terms of the noun case m}, the verb u applying
> to the argument or arguments to provide the numeric indices required by
> it.
>
> This is at least not true. Consider
>
>    100 (1 2)} i. 3 3
>   0   1   2
> 100 100 100
> 100 100 100
>    100 (1 2"_)} i. 3 3
> 0 100 100
> 3   4   5
> 6   7   8
>
> Probably I have misinterpreted the word 'numeric indices'. Does it mean
> the same as just 'indices'?
>
>
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 13:40:44 1999
From: Andrew Nikitin 
Subject: RE: Jforum: adyacence to incidence matrix transformation
Date: Mon, 19 Jul 1999 12:36:51 -0500

> -----Original Message-----
> From:	dajoy@hoy.net [SMTP:dajoy@hoy.net]
> Sent:	Sunday, July 18, 1999 12:44
	...
> But the adjacence matrix for thi graph would be
>
> +-+---------+
> | |0 1 2 3 4|
> +-+---------+
> |0|1 0 0 0 0|
> |1|0 0 1 1 0|
> |2|0 0 0 0 0|
> |3|0 0 0 0 1|
> |4|0 0 0 0 0|
> +-+---------+
>
> there is a path between:
> node 0 and node 0
> node 1 and node 2
> node 1 and node 3
> node 4 and node 4
Probably, you mean node 3 and node 4.
> The PROBLEM follows:
>
> How do I transform this:
>
> 1 0 0 0 0
> 0 0 1 1 0
> 0 0 0 0 0
> 0 0 0 0 1
> 0 0 0 0 0
>
> into this:
>
> 0 1 0 1 0
This is 1 and 3
> 0 1 1 0 0
This is 1 and 2
> 0 1 0 0 1
This is 1 and four. It is noth neither 3 and 4 nor even 4 and 4 as you
posted in previous paragraph.
> 1 0 0 0 0
>
> Anyway, I only wanted to know if there was something already done. I
> thought
> adjacence-to-incidence was a general (useful?) enough transformation
> to be
> subject to the developers' thoughts. :)
>
All these transformstions are trivial enough and this is one of possible
reasons why they are not included into phrase book. For example I invent
them from the scratch any time I deal with graphs. And usually it takes
less time than it would take if I search and decypt "standard" ones.
In addition it highly depends on the task being solved wether or not
allow loops, directed/nondirected edges, multiple edges between nodes,
weighted edges and so on. In all these cases treatment of adjacency and
incidency will slightly differ.

My current solution for you case is:

NB. This one produces (nondirected) edges list from adjacency matrix.
   edgesfromadj=:[: ~. [: /:~"1 , # (#: i.@*/)@$

NB. This one produces incidence matrix from edges list
   incfromedges=:(4 : '1 x.} 0,y. # 0')"1 >./@,

   incfromadj=:incfromedges@edgesfromadj
   adj=: ".;._2 (0 : 0)
1 0 0 0 0
0 0 1 1 0
0 0 0 0 0
0 0 0 0 1
0 0 0 0 0
)
   adj
1 0 0 0 0
0 0 1 1 0
0 0 0 0 0
0 0 0 0 1
0 0 0 0 0
   ]edg=:edgesfromadj adj
0 0
1 2
1 3
3 4
   incfromedges edg
1 0 0 0 0
0 1 1 0 0
0 1 0 1 0
0 0 0 1 1

> Now that I notice the phrase book doesn't have a section about graphs.
This is true. As it does not contain section on cellular automatas,
image processing, pixel graphics, expert systems and a lot of other
thing that seem natural to be solved in J. Probably all of them are
subjects of another book(s).

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 14:08:33 1999
From: "Joan J. Ordinas" 
Subject: Jforum: Functional capabilities of J
Date: Mon, 19 Jul 1999 20:00:57 +0200
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <000701becf60$dc9c6680$4f460693@trffpc2.kotel.co.kr>

I'm impressed with the functional capabilities of J.
I've a long background with LISP and Scheme, and I will
found interesting any paper comparing J with other functional
languages, in the LISP or ML family.

If this document exists, please, send a link to the list.

I think that the vocabulary describing J is not easy for all of us
with background in other languages. I'm continuously translating
between the J words and the "traditional" ones. And I've a lot
of questions, for example:

1. Vectors and arrays are always of a homogeneous type?
2. Are boxed vectors the solution for 1?
3. How many types do exist? Integer, rational...
4. Are vectors implemented as linked lists or as expandable
   blocks of memory?

And one more question. Is there anybody working with music and sound
with J?

Thanks in advance for your attention.

JJOR

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 14:40:27 1999
From: "Roger Hui" 
References: <000101bed210$a70b5500$f74e05c3@ariadna>
Subject: Re: Jforum: Functional capabilities of J
Date: Mon, 19 Jul 1999 11:36:10 -0400
	charset="iso-8859-1"

Joan Ordinas writes on Monday, July 19:

> I think that the vocabulary describing J is not easy for all of us
> with background in other languages. I'm continuously translating
> between the J words and the "traditional" ones.

The J terminology is more approachable to the uninitiated,
for example, "verb" and "adverb" as opposed to "function"
and "operator".  I expect that many more people can more
quickly grasp that the adverb "quickly" quickly adds a lot
of expressive power ("run quickly", "grasp quickly",
"add quickly" etc.) than will understand "operator" or
"function that takes a function argument".

> And I've a lot of questions, for example:
>
> 1. Vectors and arrays are always of a homogeneous type?

Yes, all the elements of an array are of the same internal type.

> 2. Are boxed vectors the solution for 1?

Yes.

> 3. How many types do exist? Integer, rational...

There are many internal types, and new types are being added
all the time:

Literal (bytes), boxed, booleans (bytes of 0 or 1), integer
(4 bytes), floating point (8 bytes), complex, extended integer,
rational.  Sparse types (not yet generally available): boolean,
integer, floating point, complex.

The transformation between various types are meant to
be transparent.  For example, 2+3.5 works, so does
2+3j5, so does 2+3r5, etc. Also, %:_2 works (and is
0j1.414...) Also, 2=2.0 is true (a boolean 1), as are
2=(2j2-0j2) and 2=(5r2-1r2). And so on.

> 4. Are vectors implemented as linked lists or as expandable
>    blocks of memory?

Vectors are implemented as a fixed sized block of memory.
The potential exists for up to doubling the size of a vector
before a new block is required, but this is not yet implemented.

> And one more question. Is there anybody working with music and sound
> with J?

Yes there are. I expect that they will respond to your msg.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 17:53:15 1999
Date: Mon, 19 Jul 1999 17:50:49 -0400 (EDT)
From: "Brian M. Schott" 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
In-Reply-To: <000201bed1fa$486413e0$a580f7a5@glasss>

I have been puzzled by this distinction between m} and u},
too. I wish there were a way to make them more consistent.
But looking at your characterization (x (x u y)} y) makes me
wonder if such a characterization would fail from the need
for three or four separate ranks, one pair for u and another
pair for ()} ? Perhaps this is the reason for the u} version
requiring the linear form only.

(B=)

On Mon, 19 Jul 1999, Seymour Glass wrote:

+The linear-addressing of } was in J7, wasn't it?  I thought
+it had been superseded by the new form in J2.
+
+The Dictionary suggests that x u} y is equivalent to
+x (x u y)} y.  (At least that's how I read it).  But
+Andrew's example shows that this is not the case.
+The example under u} shows this too, but I hadn't read
+that closely until this thread came along.
+
+Perhaps the documentation of u}, which matches the J7
+form (if I remember rightly), was not updated when m}
+was changed.
+
+Is the current implementation of u} the way it's supposed
+to work?  I think that it is better design to have
+x u} y   equivalent to  x (x u y)} y .  Then, if I want
+to support linear addressing, I can do it by
+x ($@] <@#: u)} y .  It's odd to have linear addressing
+supported in just this one variant of one primitive.
+
+Henry Rich
+
+> -----Original Message-----
+> From: owner-jsoftware@lists.interlog.com
+> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
+> Sent: Monday, July 19, 1999 10:42
+> To: 'forum@jsoftware.com'
+> Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
+>
+>
+> > -----Original Message-----
+> > From:	Chris Burke [SMTP:cdburke@interlog.com]
+> > Sent:	Sunday, July 18, 1999 17:15
+> >
+> >    f=: bx @: ('a'&=) @ , @ ]
+> >
+> >    f v
+> > 0 2 4 9
+> >
+> >    'Z' f} v
+> > ZdZm
+> > Znne
+> > mZry
+> >
+> Though I am aware of this special case of }, but I am still curious how
+> J distinguishes this "linear" addressing from usual case? What if I want
+> the function argument opf the } adverb to return line numbers instead of
+> linear indices? Is the sophisticated boxed syntax my only option?
+>
+> BTW, Dictionary says about } :
+> 	u} is defined in terms of the noun case m}, the verb u applying
+> to the argument or arguments to provide the numeric indices required by
+> it.
+>
+> This is at least not true. Consider
+>
+>    100 (1 2)} i. 3 3
+>   0   1   2
+> 100 100 100
+> 100 100 100
+>    100 (1 2"_)} i. 3 3
+> 0 100 100
+> 3   4   5
+> 6   7   8
+>
+> Probably I have misinterpreted the word 'numeric indices'. Does it mean
+> the same as just 'indices'?
+>
+>
+>
+> ------------------------------------------------------------------
+> --------------
+> J Forum: for information about this list, see
+http://www.jsoftware.com/forum.htm
+
+
+--------------------------------------------------------------------------------
+J Forum: for information about this list, see http://www.jsoftware.com/forum.htm
+
+

Brian Schott/Decision Sciences Dept.  (404) 651-4070    bschott@gsu.edu
J Mack Robinson College of Business        /         Georgia State Univ
Atlanta, Georgia USA 30303-3083           interests: approx. reasoning,
http://www.Gsu.EDU/~dscbms/         (B=)       decision support systems

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 19 23:19:12 1999
Date: Mon, 19 Jul 1999 23:16:03 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: OCXs
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id XAA25313

I use  OCXs is to provide functionalities that are considered essential in
modern software: spreadsheet, graphics, calendar. I tried using the J
spreadsheet. I tried using the J graphics. Extraordinary code implementing
these things. Read it and learn how to program in J.

However, many things I need in the functionalities are not offered in J and
have been programmed by specialized teams who devote their entire
livelyhood to perfecting these data interfaces or views. The cost of
selecting and buying these objects is much less than that of either  going
without my desires or programming them in J. They offer many more features
than I use, but I'll live with that.

It is very important for me to stay away from Microsoft's stuff. I can not
conceive of making a front end in VB without having to write in VB Basic.
And having to learn all that environment. And distancing myself from the
power of J. The front end is in J - that is its strength and simplicity.

J has classes allowing encapsulation ot the actual method calls. It is
rather straightforward to use OCXs. They are less of a weight on the
registry and the operating system's own proprietary stuff  than anything
coming from Microsoft. Installation is fully understandable.

I see OCXs as more complex controls - beefed up edit fields. Just something
on the periphery of what is really going on. I use only a few but there is
a vast world of OCXs out there. It is quite surprising. Still, I use only
as little as possible. Each is a small environment in itself.

That is why I agree that J already offers enough in terms of spreadsheet
and graphics. It should have no ambition to match some of the OCXs. It
interfaces rather cleanly with them.

J is not just as a back end.It is a respectable front end - as good as any
other.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 13:57:44 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
Date: Tue, 20 Jul 1999 13:49:36 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

I don't see that shortage of ranks is a problem.  ()} has
infinite rank, and u has whatever its rank is.  If
I define

   boxlinear =. ]: <@#:~ $@]

then I could write

   x (u boxlinear)} y

and get identical results (with the proposed }) that
I get now with  x u} y  using the current } .

But really, who cares?  x u} y   strikes me as not
very useful: it'll be rare that an x can be used
both as the data that modifies y and as a participant
in the index calculation.

So, I guess (without evidence) that u} is the way it
is as a vestige of earlier use, that it has been
superseded by gerund m}, and that it's just not worth
anybody's time to change it.

Henry Rich

Brian Shott wrote:
>
>
> I have been puzzled by this distinction between m} and u},
> too. I wish there were a way to make them more consistent.
> But looking at your characterization (x (x u y)} y) makes me
> wonder if such a characterization would fail from the need
> for three or four separate ranks, one pair for u and another
> pair for ()} ? Perhaps this is the reason for the u} version
> requiring the linear form only.
>
> (B=)
>
> On Mon, 19 Jul 1999, Seymour Glass wrote:
>
> +The linear-addressing of } was in J7, wasn't it?  I thought
> +it had been superseded by the new form in J2.
> +
> +The Dictionary suggests that x u} y is equivalent to
> +x (x u y)} y.  (At least that's how I read it).  But
> +Andrew's example shows that this is not the case.
> +The example under u} shows this too, but I hadn't read
> +that closely until this thread came along.
> +
> +Perhaps the documentation of u}, which matches the J7
> +form (if I remember rightly), was not updated when m}
> +was changed.
> +
> +Is the current implementation of u} the way it's supposed
> +to work?  I think that it is better design to have
> +x u} y   equivalent to  x (x u y)} y .  Then, if I want
> +to support linear addressing, I can do it by
> +x ($@] <@#: u)} y .  It's odd to have linear addressing
> +supported in just this one variant of one primitive.
> +
> +Henry Rich
> +
> +> -----Original Message-----
> +> From: owner-jsoftware@lists.interlog.com
> +> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
> +> Sent: Monday, July 19, 1999 10:42
> +> To: 'forum@jsoftware.com'
> +> Subject: RE: Jforum: Why Amend ( } ) is so difficult in J?
> +>
> +>
> +> > -----Original Message-----
> +> > From:	Chris Burke [SMTP:cdburke@interlog.com]
> +> > Sent:	Sunday, July 18, 1999 17:15
> +> >
> +> >    f=: bx @: ('a'&=) @ , @ ]
> +> >
> +> >    f v
> +> > 0 2 4 9
> +> >
> +> >    'Z' f} v
> +> > ZdZm
> +> > Znne
> +> > mZry
> +> >
> +> Though I am aware of this special case of }, but I am still curious how
> +> J distinguishes this "linear" addressing from usual case? What
> if I want
> +> the function argument opf the } adverb to return line numbers
> instead of
> +> linear indices? Is the sophisticated boxed syntax my only option?
> +>
> +> BTW, Dictionary says about } :
> +> 	u} is defined in terms of the noun case m}, the verb u applying
> +> to the argument or arguments to provide the numeric indices required by
> +> it.
> +>
> +> This is at least not true. Consider
> +>
> +>    100 (1 2)} i. 3 3
> +>   0   1   2
> +> 100 100 100
> +> 100 100 100
> +>    100 (1 2"_)} i. 3 3
> +> 0 100 100
> +> 3   4   5
> +> 6   7   8
> +>
> +> Probably I have misinterpreted the word 'numeric indices'. Does it mean
> +> the same as just 'indices'?
> +>
> +>
> +>
> +> ------------------------------------------------------------------
> +> --------------
> +> J Forum: for information about this list, see
> +http://www.jsoftware.com/forum.htm
> +
> +
> +-----------------------------------------------------------------
> ---------------
> +J Forum: for information about this list, see
> http://www.jsoftware.com/forum.htm
> +
> +
>
> Brian Schott/Decision
> Sciences Dept.  (404) 651-4070    bschott@gsu.edu
> J Mack Robinson College of Business        /         Georgia State Univ
> Atlanta, Georgia USA 30303-3083           interests: approx. reasoning,
> http://www.Gsu.EDU/~dscbms/         (B=)       decision support systems
>
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
> http://www.jsoftware.com/forum.htm
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 14:48:27 1999
Date: Tue, 20 Jul 1999 19:40:47 +0100
From: Mike Day 
CC: centre@cwcom.net
Subject: Jforum: [Fwd: greetings]
Content-Type: multipart/mixed; boundary="------------3431EAA292555FCF25565E17"

This is a multi-part message in MIME format.
--------------3431EAA292555FCF25565E17

Anybody able to help Stephen?

Stephen Wynn wrote to me (M Day):

>In your J exploits have you come across "semi-circulant"
> matrices?
>  M2=(b a)
>     (c b)
>
>     (c  b  a)
>  M3=(d  c  b)
>     (e  d  c)
> (Diagonals have the same value,  xij = x(i+k)(j+k))

Mike
20 vii 99
--------------3431EAA292555FCF25565E17
Content-Type: message/rfc822
Content-Disposition: inline

Date: Sat, 17 Jul 1999 19:48:41 +0100
From: Stephen Wynn 
Subject: greetings

> Hello,
>
> It is great to hear your news. Cw is faster and cheaper than my previous
> provider. In your J exploits have you come across "semi-circulant"

> matrices?

>  M2= (b a)                   (c  b  a)
>           (c b)        M3 =  (d  c  b)     etc
>                                     (e  d  c)
>
> (Diagonals have the same value,  xij = x(i+k)(j+k))
>
> Regards,
>
> Steve.

--------------3431EAA292555FCF25565E17--

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 16:26:38 1999
From: "Oleg Kobchenko" 
References: <199907192316_MC2-7D8A-9CC3@compuserve.com>
Subject: Re: Jforum: OCXs
Date: Tue, 20 Jul 1999 23:25:24 +0300
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit

I can't but agree with all these arguments, which for
a moment made me doubt in the rationale of my original
message. Indeed, in it there was one important point implied
but still missing and ought to be replaced. The logic behind
J GUI is really extraordinary. A single example is the position
bound control behavior, which is unmached in VB and requires
painstaking layout objects in Java and only approached certain
simplicity in VJ++6.0. Another one is event model and control
properties as variables.

What makes it 90% is the fact that a compact neet
model rules the complicated world of controls and events.
If fact, it is rather 90+1%, where 1 is repetedly emerging
increments, which eventually mould with the 90 and make most
of the solutions possible, like grid, sm enhancements, etc.

It is certainly out of common sense to figure another
implementation of Visual Studio, for J. However, economical
and logically non-contradictory investments in the small
increments should yield in significant added value for the 90%.
Here are a few examples.

1. There is a wd command to programmatically change the position
and size of a control,
   setxywh[x] id x y w h ;  set child position and size in units/pixels
which works fine for fixed size forms. Whenever the form changes
size, the set values are forgotten and the control receives its
predefined position and then is moved/scaled according to its settings.
This would not happen should the original position be adjusted with
a setxywh call giving a new 'original position'.

2. To handle this situation as is and for possible other applications
a useful event would be form_size, for absence of which I do
not see any serious reasons.

An add-hoc present solution may be to place a (hidden) isigraph
with [sides]move property and capture the _size event. But it will
lead to re-implementing of the layout logic in expensive J control
flow structures.

3. There is a mechanism already implemented in J GUI to use
multiple child forms for a parent form: the wd 'creategroup tabs',
where the child forms have their own form locales and event handling.
Why not extend this into MDI, which has very similar functionality?

----- Original Message -----
From: Alain Miville de Ch�ne 
Sent: Tuesday, July 20, 1999 6:16 AM
Subject: Jforum: OCXs

> I use  OCXs is to provide functionalities that are considered essential in
> modern software: spreadsheet, graphics, calendar. I tried using the J
> spreadsheet. I tried using the J graphics. Extraordinary code implementing
> these things. Read it and learn how to program in J.^^^^^^^^^^^^^^^^^^^^^^
  ^^^^^^^^^^^^

> It is very important for me to stay away from Microsoft's stuff. I can not
> conceive of making a front end in VB without having to write in VB Basic.
> And having to learn all that environment. And distancing myself from the
> power of J. The front end is in J - that is its strength and simplicity.
>
> J has classes allowing encapsulation ot the actual method calls. It is
> rather straightforward to use OCXs. They are less of a weight on the
> registry and the operating system's own proprietary stuff  than anything
> coming from Microsoft. Installation is fully understandable.

> J is not just as a back end.It is a respectable front end - as good as any
> other.

----- Original Message -----
From: Eric Iverson 
Sent: Sunday, July 18, 1999 10:23 PM
Subject: Jforum: hybrid applications - VB/C++/Java and J

> Oleg Kobchenko wrote:
> > ... There is nothing exciting about VB, except for the UI.
> > But it acts as the dress code for the party of contemporary
> > business applications.
>
> The J UI is adequate for applications with modest UI requirements. It is
> not adequate for applications that require the latest UI sizzle. We've
> put a lot of effort into providing a 90% GUI solution. The cost of
> providing the remaining 10% remains too high. The problem is that VB
> won't stand still. We could put in the significant effort to catch up to
> where they are today. But by then they'd have moved and we'd still be
> behind. I don't like being in a race I am guaranteed to lose.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 17:52:02 1999
From: Andrew Nikitin 
Subject: Jforum: more on amendment
Date: Tue, 20 Jul 1999 16:48:25 -0500

Here is a small j problem.

Suppose you have character matrix of arbitrary size. You want to "build
a fence" inside of it. This means:
1) put horisontal bars in place of top and bottom rows,
2) put vertical bars in place of left and right column,
3) put plus signs (or borderdraw corners) in place of corner elements.

These steps can be implemented as follows:

M=:3 5$'*'
M1=:M

NB. --- step 1)
HORZ=:<(<0 _1),<<0 _1
M1=:'-' HORZ} M1

NB. --- step 2)
VERT=:<(<<0 _1),<0 _1
M1=:'|' VERT} M1

NB. --- step 3)
CORNERS=:(<"1 -#:i.4)
M1=:'+' CORNERS} M1

Note that nouns VERT, CORNERS and HORZ do not depend on M size.

Is it possible to produce noun BORDER (that is also does not depend on
M) such that
M1=: '+-|' BORDER} M
will produce the same result as previos sequenece of steps?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 21:19:48 1999
From: "Seymour Glass" 
Subject: RE: Jforum: more on amendment
Date: Tue, 20 Jul 1999 21:14:08 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

Are you counting a gerund as a noun (then it's easy)?  If not,
I'd like to see the answer.

The question reminds me of something that threw me recently:

When you have an amend m} where m is not an atom, it's a
good idea to make all the atoms select subarrays of the
same shape; otherwise element 0 of the array gets modified.
So, if you do an amend like

   ( (<(<0 _1),<<0 _1) , (<(<<0 _1),<0 _1) )}

you will modify the (apparently unselected) top-left
element as well.

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
> Sent: Tuesday, July 20, 1999 17:48
> To: 'forum@jsoftware.com'
> Subject: Jforum: more on amendment
>
>
> Here is a small j problem.
>
> Suppose you have character matrix of arbitrary size. You want to "build
> a fence" inside of it. This means:
> 1) put horisontal bars in place of top and bottom rows,
> 2) put vertical bars in place of left and right column,
> 3) put plus signs (or borderdraw corners) in place of corner elements.
>
> These steps can be implemented as follows:
>
> M=:3 5$'*'
> M1=:M
>
> NB. --- step 1)
> HORZ=:<(<0 _1),<<0 _1
> M1=:'-' HORZ} M1
>
> NB. --- step 2)
> VERT=:<(<<0 _1),<0 _1
> M1=:'|' VERT} M1
>
> NB. --- step 3)
> CORNERS=:(<"1 -#:i.4)
> M1=:'+' CORNERS} M1
>
> Note that nouns VERT, CORNERS and HORZ do not depend on M size.
>
> Is it possible to produce noun BORDER (that is also does not depend on
> M) such that
> M1=: '+-|' BORDER} M
> will produce the same result as previos sequenece of steps?
>
>
>
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 22:44:00 1999
Date: Tue, 20 Jul 1999 22:40:43 -0400
From: David Ness 
Subject: Jforum: Even more questions about amend
References: <4.2.0.58.19990716114829.00aa1e00@pop.kolumbus.fi>

There is a problem that I don't quite know how to characterize that
I find very easy to solve in K, but not so easy to solve in J.
Perhaps I haven't looked at the right part of `Phrases', or I maybe
don't know the right words to characterize the problem. I think it
relates to the nature of `amend' in J and K, and may thus fit into
the current discussion of the nature of amend.

Essentially the problem is to do a frequency count of the occurrence
of values [0,...,m] in a n-long vector.

I am curious about how one should use `J' to deal with the kind of
amend that `K' accomplishes with a construct like:
    a[v] +: 1
where `a' starts as a vector of 0's (a:1000#0) and `v' is a random
vector of [0...999] (v:1000000 _draw 1000).

My first try was to do something awful like:
     +/"1 (i.1000) =/ v
which ran, although slowly, for short v, but blew up with v that was
100,000 long much less 1,000,000. This had the `advantage' that it
seemed straightforward, but obviously generating the intermediate
`table' strains any reasonable memory availability.

I then found the `key' operation and ended up gradually evolving to:
    (]((<@#)/.)##1:)/:~.
which ran faster for small v, and would work for the 1,000,000 case.
However, it also seems `artificial', building up the vectors of 1s, and
then counting them, rather than tabulating directly.

Also, the K construct not only runs leagues faster than this, it
does seem conceptually `simpler'. I'm not sure how much of the speed
differential is attributable to the general speed advantage of K, and
how much is due to the fact that I have an unnecessarily complex
formulation of the problem in J.

Is there another approach in `J' that I should be seeing? Is this
another form of `amend' or is it already in `J' hidden some way I am not
seeing it.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 23:17:40 1999
From: Eemcd@aol.com
Date: Tue, 20 Jul 1999 23:12:07 EDT
Subject: Re: Jforum: Even more questions about amend

David Ness:

f you know what the nub of the data q is you can do #/.~q. If you need to
know the nub, you can use nub and count phrases m0 or m1 from Section 15.A of
the Phrase Book:

m0=: ~. ,. #/.~
m1=: ({.,#)/.~

Look also at Section 8.A m46, m47, m48,m49 and others nearby. There is some
overlap between 15.A and 8.A.

Eugene McDonnell

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 20 23:18:19 1999
Date: Tue, 20 Jul 1999 23:15:10 -0400
From: David Ness 
Subject: Re: Jforum: Even more questions about amend
References: 

As usual, many thanks. It will be a `productive' evening investigating
your help.

Eemcd@aol.com wrote:
>
> David Ness:
>
> f you know what the nub of the data q is you can do #/.~q. If you need to
> know the nub, you can use nub and count phrases m0 or m1 from Section 15.A of
> the Phrase Book:
>
> m0=: ~. ,. #/.~
> m1=: ({.,#)/.~
>
> Look also at Section 8.A m46, m47, m48,m49 and others nearby. There is some
> overlap between 15.A and 8.A.
>
> Eugene McDonnell
>
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 02:13:14 1999
Date: Wed, 21 Jul 1999 09:09:48 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: OCXs
In-Reply-To: <199907192316_MC2-7D8A-9CC3@compuserve.com>

This OCX thread has been interesting. Can someone kindly tell what are the
best OCX controls for J and give examples how to use them?
Anssi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 02:29:58 1999
Date: Wed, 21 Jul 1999 09:29:12 +0300
From: Anssi  Seppala 
Subject: Re: Jforum: J Plot key with linepatterns
In-Reply-To: <4.2.0.58.19990717192951.0094bde0@pop.kolumbus.fi>
References: <003b01bed05d$66ee9400$9545fea9@t500>
 <4.2.0.58.19990717145459.00aa15c0@pop.kolumbus.fi>

Indeed the JPlot utility has undocumented features. keystyle is documented
values 0 0 to 1 2. However there seems to be implementation for 2  0, 3 0 ,
etc. giving also linepatterns to the key.

What about the idea to put ... int the documentation to tell that there is
more, but no one has time to write it? Then we enthusiaists can start our
own eperiments :)

Anssi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 02:39:16 1999
From: "Randy MacDonald" 
References: <4.2.0.58.19990716232641.00ab5100@pop.kolumbus.fi> <4.2.0.58.19990717161202.00ab1680@pop.kolumbus.fi> <006c01bed16b$2c686ea0$c0c7fea9@t500> <37927CB4.361760AC@Home.Com>
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?
Date: Mon, 19 Jul 1999 01:10:12 -0400
	charset="iso-8859-1"

Explicitly

     f x ==  bx 'a'= , x
         ==  bx ('a'&=) , x
         ==  ( bx @: ('a' & = ) @ , @ ] ) x
     f   ==  ( bx @: ('a' & = ) @ , @ ]

using the idea that

     f g x == (f @: g @: ]) x
     f g h x == (f @: g @: h @: ]) x

transforms a composition of functions into a list with an unconventional
connector and a placeholder.

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 9:17 PM
Subject: Re: Jforum: Why Amend ( } ) is so difficult in J?

> Chris Burke wrote:
> >
> > ...
> >
> >    f=: bx @: ('a'&=) @ , @ ]
> >
>
> Chris, I see that this works, but it would be very helpful to me if
> you could say a word or two about `why' the `At' and why the
> `Same-Right'. I find I can often follow your elegant constructs, but
> I still have great trouble generating them myself, and I'm looking for
> clues on how you `see' things that results in such nice encodings...
>
> Thanks for any insights you can offer...

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 04:05:44 1999
From: "Randy MacDonald" 
References: <378F618C.E10B2AA4@Home.Com> <3.0.5.32.19990717173138.00919b90@mail.interlog.com> <3792356E.916AC6A7@Home.Com>
Subject: Re: Jforum: J in HTML?
Date: Mon, 19 Jul 1999 00:53:05 -0400
	charset="iso-8859-1"

----- Original Message -----
From: David Ness 
Sent: July 18, 1999 4:13 PM
Subject: Re: Jforum: J in HTML?

> I appreciate the attempts to help, but am still having a problem for
> a couple of reasons.
>
> (1) Apparently Run|HTML Publish must be something new in J of some
> relatively recent origin, so I don't have access to it, and thus
> don't really have any idea of what it does.

If, as Kirk stated, you don't have J4.03 then you may not have this
feature.

> (2) I'm actually not so much interested in solutions that involve
> having J generate things differently (if this is what HTML Publish does)
> as for expository purposes I want to be able to generate quite `literal'
> `in' and `out' scripts. As a result I need to be able to solve the
> problem of processing a perfectly normal J output file into some form
> that will reproduce decently in NetScape, Explorer, Opera, Amaya and any
>  other `normal' browsers, as it does, by the way, if you do have
> access to `ISIJ' or `Focal Point On-Line' fonts.

One solution is to , wherever you need boxed output is to use a verb
to convert the output into html, and to manually edit out the use of that
verb
in that file.

    NB. untested code, but close to what you want.

    htdisp 1 2 $ < 1 2

1 21 2


    tabletag =: (''&,)@:(,&'
')
    rowtag =: (''&,)"1
    celltag =: (''&,each)"0

    htdisp =:   tabletag@: ; @: rowtag @: celltag @: ":

    NB. and if your boxes have boxes, you'll have to recurse.

> (3) I am beginning to guess that I'll probably need to interpose PERL
> into the process anyway, so I suppose I could have PERL rewrite the
> J output into some form that would use  table facilities.

the Perl would, of course be much harder to write, and just another
thing that could go wrong.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 08:32:36 1999
From: "Alex Kornilovski" 
Subject: Re: Jforum: OCXs
Date: Wed, 21 Jul 1999 08:37:00 -0400
	charset="iso-8859-1"

Anssi wrote:
>This OCX thread has been interesting. Can someone kindly tell what are the
>best OCX controls for J and give examples how to use them?
>Anssi
If you need to add a cool inteface to your programs look at
http://webhome.idirect.com/~kornal
or go to : http://members.xoom.com/olegyk/jpage.html - lots of examples

P.S. Also take a look at J and APL WebRing and add your site if you have
one.

Alex

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 08:38:12 1999
From: "Seymour Glass" 
Subject: Jforum: Need help accessing the Windows Registry
Date: Wed, 21 Jul 1999 08:38:31 -0400
	charset="Windows-1252"
Importance: Normal
In-Reply-To: 

I would like to save some configuration information in
the Windows Registry.

Does anybody have some J that will read and write the
Registry?

Also, does anybody know what the 'shell environment' is
in J under Windows, for the purposes of 2!:5 ?  Is that
the Registry, and if so do you know the syntax for
naming a variable?

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 08:56:40 1999
Date: Wed, 21 Jul 1999 08:54:21 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Re: Jforum: OCXs
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id IAA15765

Message text written by INTERNET:forum@jsoftware.com
>Can someone kindly tell what are the
best OCX controls for J and give examples how to use them?<

As I already wrote, I use the controls from Far Point: the spreadsheet,
calendar, and sundry input controls such as date. I chose those after
looking around at what is out there in the market. At the time J couldn't
(or I didn't know how) (now I do) use collections, thus ruling out many
spreadsheet controls. Those I chose are perfectly adequate for my needs.

I also use the Olectra chart control for graphs.

One last one is Xceedzip which I use to compress data with the same
algorithm used in .zip files. Most of the data I store in J files is
compressed. I like saving disk space.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 09:56:11 1999
From: John Scholes 
Subject: RE: APL performance - was Re: Jforum: Even more questions about amend
Date: Wed, 21 Jul 1999 14:47:44 +0100

Not sure if this email came from Mike or David?
Anyway, thanks for the information.
'Fraid that my talk on Dfns is still fairly fluid and will remain so until
around half an hour into the talk itself. Perhaps I can send you a
'postview' afterwards.
John.

On Wednesday, July 21, 1999 10:07 AM, DNess@home.com [SMTP:DNess@home.com]
wrote:
> John
>
> The interchange quoted below is neither here nor there
> ***except*** he's trying to do a fairly standard trick -
>    v <- ? 1000000 rho 1000
>    a <- 1000 rho 0
>    a[v] + <- 1           @ count frequencies
> which I thought was pretty nifty in APL.
>
> His "k" coding is exactly analogous
>    a[v] +: 1
>
> On my "k-lite" trial version (but the licence
> costs $5000! so don't worry) this zips through in
> 0.17 sec on my desktop.
>
> Eugene McDonald's suggestion for doing it in J,
> which is rather counter-intuitive in my eyes,  not
> copied here,  takes ~ 1.3 sec.
>
> In Dyalog APL/W 8.2.2 it seems to take about 12 sec.
> Intermediate timings:
>    1000    ?
>   10000   ~60 ms
>  100000  ~880 ms
>  200000 ~1.9   s
>  300000 ~3.1   s
>  500000 ~5.4   s
>
> I got the same-ish result whether with 20Mb or 30Mb
> of workspace in a 64Mb machine at 200 MHz under Win 95,
> with quite a lot of resources devoted to Office
> Information stuff. But then J & K run in the same
> environment.
>
> Do you think at (y)our age this is right?
>
> PS - any chance of a preview of your talk on dynamic
> functions at APL99? - I'm not going unfortunately.
>
> All the best,
>
> Mike
> 21 7 99
>
> [correspondence of 20 7 99]
> >To:             d=forum@jsoftware.com  ou=smtp o=maff p=maff400
a=attmail c=gb
> >From:           d=owner-jsoftware@lists.interlog.com  ou=smtp o=maff
p=maff400 a=attmail c=gb
> >Subject:        Jforum: Even more questions about amend
> >... preamble...
> >Essentially the problem is to do a frequency count of the occurrence
> >of values [0,...,m] in a n-long vector.
> >
> >I am curious about how one should use `J' to deal with the kind of
> >amend that `K' accomplishes with a construct like:
> >    a[v] +: 1
> >where `a' starts as a vector of 0's (a:1000#0) and `v' is a random
> >vector of [0...999] (v:1000000 _draw 1000).
> >... explores possibilities in J...
> [from David Ness]
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 10:21:29 1999
From: Andrew Nikitin 
Subject: RE: Jforum: more on amendment
Date: Wed, 21 Jul 1999 09:18:09 -0500

> -----Original Message-----
> From:	Seymour Glass [SMTP:glasss@mindspring.com]
> Sent:	Tuesday, July 20, 1999 20:14
>
> Are you counting a gerund as a noun (then it's easy)?  If not,
> I'd like to see the answer.
No, I dont count gerund as noun. And I dont have an answer. I wonder if
it is posiible at all.
It is easy to build wall from one symbol:

B=:( (<(<0 _1),<<0 _1) , (<(<<0 _1),<0 _1) ), (<(<0 _1),<0 _1)
M=: 5 8 $ 'abcde'
BM=:'*' B} M

I've also found one with variable LEFT argument to amend:

LEFT=:($B{M) ([ $ (*/@}.@[) # ]) '-|+'
LEFT B} M
Can it be simplified to CONSTANT left argument?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 10:55:14 1999
From: Andrew Nikitin 
Subject: RE: Jforum: Need help accessing the Windows Registry
Date: Wed, 21 Jul 1999 09:51:14 -0500

> -----Original Message-----
> From:	Seymour Glass [SMTP:glasss@mindspring.com]
> Sent:	Wednesday, July 21, 1999 07:39
>
> Also, does anybody know what the 'shell environment' is
> in J under Windows, for the purposes of 2!:5 ?  Is that
> the Registry, and if so do you know the syntax for
> naming a variable?
It is not registry, it is so called 'environment' variables set by SET
command in your autoexec.bat file.
The most useful of them is PATH variable:
   2!:5 'PATH'
C:\WINNT\system32;C:\WINNT;c:\sqlany50\win;c:\sqlany50\win32;c:\sybtools
\win32;c:\usr\bin;c:\usr\sculptor\bin;C:\WINNT\System32\WBEM

Another useful example is
    2!:5 'windir'
C:\WINNT

You can obtain a list of variables available on your system by typing
SET in dos command prompt (I assume that you use Windows).

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 11:40:07 1999
From: "Seymour Glass" 
Subject: RE: Jforum: more on amendment
Date: Wed, 21 Jul 1999 11:39:03 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
> Sent: Wednesday, July 21, 1999 10:18 AM
> To: 'forum@jsoftware.com'
> Subject: RE: Jforum: more on amendment
>
>
> > -----Original Message-----
> > From:	Seymour Glass [SMTP:glasss@mindspring.com]
> > Sent:	Tuesday, July 20, 1999 20:14
> >
> > Are you counting a gerund as a noun (then it's easy)?  If not,
> > I'd like to see the answer.
> No, I dont count gerund as noun. And I dont have an answer. I wonder if
> it is posiible at all.
> It is easy to build wall from one symbol:
>
> B=:( (<(<0 _1),<<0 _1) , (<(<<0 _1),<0 _1) ), (<(<0 _1),<0 _1)
> M=: 5 8 $ 'abcde'
> BM=:'*' B} M

This is an example of the non-atomic amend that should
be used with caution.  Your example works because you are
trying to modify the top-left cell, but if I change it
to bring the top of the box down one row, look what happens:

  B=:( (<(<1 _1),<<0 _1) , (<(<<0 _1),<0 _1) ), (<(<1 _1),<0 _1)
  M=: 5 8 $ 'abcde'
  '*' B} M
*bcdeabc
********
*cdeabc*
*abcdea*
********

See? the top-left cell is modified too.  So this whole approach
is not recommended for general use.

>
> I've also found one with variable LEFT argument to amend:
>
> LEFT=:($B{M) ([ $ (*/@}.@[) # ]) '-|+'
> LEFT B} M
> Can it be simplified to CONSTANT left argument?

I can do it if you let me use a gerund (which really IS
a noun, syntactically):

   top =. , @: (0 _1&(<@,"0/ i.@{:@$))
   sides =. , @: (0 _1&(<@,"0/~ i.@{.@$))
   corners =. (, <@,"0/~ 0 _1)"_
   topdt =. +:@{:@$@] # 1&{@[
   sidesdt =. +:@{.@$@] # 2&{@[
   cornersdt =. 4: # 0&{@[
   dt =. topdt , sidesdt , cornersdt
   mods =. (top , sides , corners)@]
   frame =. dt ` mods ` ]
   '+-|' frame} 4 6$'a'
+----+
|aaaa|
|aaaa|
+----+

I didn't say it was pretty!  By boxing the selector
for each element, each atom selects a block of the same
shape (viz one atom), and so the idea would work for
a box that didn't touch the top-left cell.  I have taken
advantage of the fact that the modifications are made
starting with item 0, which is not guaranteed by the
Dictionary but seems to be the current implementation.

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 11:48:50 1999
From: "John D. Baker" 
Subject: Re: APL performance - was Re: Jforum: Even more questions about amend
Date: Wed, 21 Jul 1999 11:47:01 -0400
	charset="iso-8859-1"

For years I've used the little J verb

freqdist =: ~.@] ,: #/.~

To build frequency distribtions of integer data types.  For example

x =:  ? 1000000 $ 1000

y =:  freqdist x

takes roughly 0.3 seconds on my NT machine.  Note that freqdist is
doing more than the APL and K idioms discussed in this thread
To just count the frequencies for integer data all we need is:

cf =: #/.~

This executes so quickly on million element integer vectors I have to
run it a 100 times to get an average time of roughly 0.09
seconds for the problem discussed.   This idiom has been optimized
by Roger (in J 4.02  and greater) and is unquestionably the fastest way
to do this task in J.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 14:16:37 1999
Date: Wed, 21 Jul 1999 14:24:25 -0400
From: Eric Iverson 
Organization: Iverson Software Inc.
Subject: Jforum: wd 90% + 1%
References: <199907192316_MC2-7D8A-9CC3@compuserve.com> <009601bed2ee$01022600$0801a8c0@spin.kharkov.ua>

I agree with Oleg Kobchenko's suggestion that a high priority for a
future release of wd is a form_size event that is nicely integrated with
child size/position commands and the autosize mechanism.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 14:16:37 1999
Date: Wed, 21 Jul 1999 14:12:58 -0400
From: David Ness 
Subject: Re: APL performance - was Re: Jforum: Even more questions about amend
References: <003001bed390$83343260$01ffde6f@papabear>

"John D. Baker" wrote:
>
> For years I've used the little J verb
>
> freqdist =: ~.@] ,: #/.~
>
> To build frequency distribtions of integer data types.  For example
>
> x =:  ? 1000000 $ 1000
>
> y =:  freqdist x
>
> takes roughly 0.3 seconds on my NT machine.

Thanks. EEMCD pointed me to the section of the Phrase book with some
similarly helpful examples. Running your code on my NT 200mhz 64mb
PentII takes closer to 1 sec than your 0.3 sec, but that's probably just
that you have a faster machine or I have a slower version of J.

>  Note that freqdist is
> doing more than the APL and K idioms discussed in this thread
>

For my purposes, actually it is doing a little `less', not `more'. The
K idiom (and maybe the APL idiom as well) returns a vector (a) in order
and (b) with 0s for boxes that didn't get hit. Since that is actually
what I need, at the moment I have to do an `amend' to get what K yields
directly with the `a[v] +:1' construct.

> To just count the frequencies for integer data all we need is:
>
> cf =: #/.~
>
> This executes so quickly on million element integer vectors I have to
> run it a 100 times to get an average time of roughly 0.09
> seconds for the problem discussed.   This idiom has been optimized
> by Roger (in J 4.02  and greater) and is unquestionably the fastest way
> to do this task in J.
>

Takes 0.4 sec on my machine, not counting the amend, but that's
a lot faster than when I started out on this one yesterday...

Thanks (and to EEMCD, too) for the pointers.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 14:35:10 1999
From: Andrew Nikitin 
Subject: RE: APL performance - was Re: Jforum: Even more questions about a
	mend
Date: Wed, 21 Jul 1999 13:32:29 -0500

> -----Original Message-----
> From:	David Ness [SMTP:DNess@home.com]
> Sent:	Wednesday, July 21, 1999 13:13
> > freqdist =: ~.@] ,: #/.~
> For my purposes, actually it is doing a little `less', not `more'. The
> K idiom (and maybe the APL idiom as well) returns a vector (a) in
> order
> and (b) with 0s for boxes that didn't get hit.
	...
> Since that is actually
> what I need, at the moment I have to do an `amend' to get what K
> yields
> directly with the `a[v] +:1' construct.
>
You can easily modify this verb to get the vector with (a) and (b)
properties.
   '' (#/.~@])`(~.@])`(>:@(>./)@] # 0:)} 1 3 3 2 1 2  3
0 2 2 3

One of the strange things about amend is ugly fake left argument that
often appears in the expressions of this kind. This, in turn, leads to
the repetitive @], that leads to necessity of extra brackets.

Does you solution also uses fake left argument?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 15:11:57 1999
Date: Wed, 21 Jul 1999 15:08:40 -0400
From: David Ness 
Subject: Re: APL performance - was Re: Jforum: Even more questions about amend
References: 

Very interesting. Thanks.

My initial try is that this takes about 10 times longer than the
original, but, as you say, it doesn't require any other amend. Later
this afternoon I should get a chance to compare it with the amend
version, to see how they compare.

My first try at the amend part used a real left argument, not a fake one. It is a construct
something like `(#/.~y.) (~.y.)}x.#0' and it
runs just about the same speed as yours, perhaps 1-2% slower.

I'll do some more careful tests later.

Andrew Nikitin wrote:
>
> You can easily modify this verb to get the vector with (a) and (b)
> properties.
>    '' (#/.~@])`(~.@])`(>:@(>./)@] # 0:)} 1 3 3 2 1 2  3
> 0 2 2 3
>
> One of the strange things about amend is ugly fake left argument that
> often appears in the expressions of this kind. This, in turn, leads to
> the repetitive @], that leads to necessity of extra brackets.
>
> Does you solution also uses fake left argument?
>
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 16:30:55 1999
Date: Wed, 21 Jul 1999 16:30:15 -0400
From: Daniel Torres 
Subject: Jforum: J runtime sessions not fully closing.
Content-Type: multipart/mixed;
 boundary="------------0C40B4DB01C53B9141FC7785"

This is a multi-part message in MIME format.
--------------0C40B4DB01C53B9141FC7785

I run a script in Win98, from a shortcut with a command:  "c:\j.exe /rt
c:\script.ijr".

The script's structure is:

SCRIPT=: 3 : 0
try.
...            NB. program here
catch.
wdstatus terminationmessage
end.
)

SCRIPT ''

The script runs perfectly.  Any error, including "Ctrl-Break", causes the
script to terminate and display the status window with a termination message.
While the program is running, pressing "Ctrl-Alt-Del" brings up the the window
displaying active sessions.  "J" does not show up.

Closing this latter window, and then pressing "Ctrl-Break", terminates my
script and displays the termination message in the expected status window.
Pressing "Ctrl-Alt-Del" again, brings up the the window displaying active
sessions.  "J" still does not show up.

Closing this latter window, as well as the termination message from my script,
and then pressing "Ctrl-Alt-Del", brings up the the window displaying active
sessions.  "J" now shows up.

If I select the "J" and click on "End Task" I get a message window titled "J",
with the question "Do you want to close" and two buttons "Yes" and "No".  If I
keep this message open for a few seconds I get a message from Windows titled
"J" and saying "This program is not responding.....". I can "Cancel" or "End
Task".  "End task" in this second message window or "Yes" in the preceeding one
closes the processes a "J" no longer appears in the listing.

Running my script n times and then pressing "Ctrl-Alt-Del", displays a list of
n entries for "J".  These can all be closed via "End Task" but closing some
results in the blue Windows exception or program busy message window.

Is my use of "try. catch. end." incorrect?  Is there something I might do to
prevent this problem.

Thanks.

Daniel
--------------0C40B4DB01C53B9141FC7785
 name="torres.vcf"
Content-Description: Card for Daniel Torres
Content-Disposition: attachment;
 filename="torres.vcf"

begin:vcard
n:Torres;Daniel
tel;fax:1-978.383-5817
tel;home:1-305.441-0369
tel;work:1-305.461-6829
x-mozilla-html:TRUE
adr:;;600 Biltmore Way APT 412;Coral Gables;FL;33134-7529;USA
version:2.1
email;internet:torres@SAmerica.com
fn:Daniel Torres
end:vcard

--------------0C40B4DB01C53B9141FC7785--

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 21 18:33:00 1999
Date: Wed, 21 Jul 1999 18:35:32 -0400
From: Eric Iverson 
Organization: Iverson Software Inc.
Subject: Re: Jforum: J runtime sessions not fully closing.
References: <37962DD7.210AACFE@SAmerica.com>

Daniel Torres wrote:
> ... "J" does not show up. [various runtime mysteries] ...

Windows 95/98 ctrl/alt/del task manager does not list tasks that don't
have a visible window. Is this the case for J in your examples?

There may be a bug in runtime task termination vs try/catch. Does adding
an explicit 2!:55 in both the try and catch section help?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Thu Jul 22 04:19:17 1999
From: "Marwan Khawaja" 
Date: Thu, 22 Jul 1999 10:17:39 +0200
Content-type: text/plain; charset=US-ASCII
Subject: Re: Jforum: J Plot
References: <4.2.0.58.19990717192951.0094bde0@pop.kolumbus.fi>

The definitions in the jplot utility demo shows many options that
are not documented (e.g. cut).  I was trying to plot an age pyramid
by making two horizental bar charts and flipping/rotating one chart --
with not much luck perhaps because I am a newcomer to J.
Any suggestions?
Thanks
Marwan

> Indeed the JPlot utility has undocumented features. keystyle is documented
> values 0 0 to 1 2. However there seems to be implementation for 2  0, 3 0 ,
> etc. giving also linepatterns to the key.
>
> What about the idea to put ... int the documentation to tell that there is
> more, but no one has time to write it? Then we enthusiaists can start our
> own eperiments :)
>
> Anssi
>
> --------------------------------------------------------------------------------
> J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Thu Jul 22 11:05:18 1999
From: Charles Fisk 
Subject: RE: Jforum: Need help accessing the Windows Registry
Date: Wed, 21 Jul 1999 10:15:05 -0500

This is interesting, does  2!5 allow you to rename a file? can you rename
with wild cards?

> -----Original Message-----
> From:	Andrew Nikitin [SMTP:anikitin@fastenal.com]
> Sent:	Wednesday, July 21, 1999 9:51 AM
> To:	'forum@jsoftware.com'
> Subject:	RE: Jforum: Need help accessing the Windows Registry
>
> > -----Original Message-----
> > From:	Seymour Glass [SMTP:glasss@mindspring.com]
> > Sent:	Wednesday, July 21, 1999 07:39
> >
> > Also, does anybody know what the 'shell environment' is
> > in J under Windows, for the purposes of 2!:5 ?  Is that
> > the Registry, and if so do you know the syntax for
> > naming a variable?
> It is not registry, it is so called 'environment' variables set by SET
> command in your autoexec.bat file.
> The most useful of them is PATH variable:
>    2!:5 'PATH'
> C:\WINNT\system32;C:\WINNT;c:\sqlany50\win;c:\sqlany50\win32;c:\sybtools
> \win32;c:\usr\bin;c:\usr\sculptor\bin;C:\WINNT\System32\WBEM
>
> Another useful example is
>     2!:5 'windir'
> C:\WINNT
>
> You can obtain a list of variables available on your system by typing
> SET in dos command prompt (I assume that you use Windows).
>
>
>
> --------------------------------------------------------------------------
> ------
> J Forum: for information about this list, see
> http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Thu Jul 22 11:15:04 1999
Date: Thu, 22 Jul 1999 11:12:19 -0400
From: David Ness 
Subject: Re: Jforum: Need help accessing the Windows Registry
References: 

I must miss the point. What would `Getenv' have to do with renaming?

Charles Fisk wrote:
>
> This is interesting, does  2!5 allow you to rename a file? can you rename
> with wild cards?
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 23 13:44:42 1999
From: Andromeda 
Subject: Jforum: ?? is ..
Date: Fri, 23 Jul 1999 19:41:05 +0200

 there any help about that MAPI-problem and this strange behaviour of the
 of the DLL-error 11 if sending attachments?
 Has anybody tried it and detected also similar behaviour ..?

 //- Milan

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 23 14:58:54 1999
Date: Fri, 23 Jul 1999 14:58:47 -0400
From: Daniel Torres 
Subject: Re: Jforum: J runtime sessions not fully closing.
References: <37962DD7.210AACFE@SAmerica.com> <37964B34.5314@interlog.com>
Content-Type: multipart/mixed;
 boundary="------------F0ADF3AC87ED1B42C04D456A"

This is a multi-part message in MIME format.
--------------F0ADF3AC87ED1B42C04D456A

The script in question only produce status windows, which may explain why the
task manager does not list the script as a task while it runs.  Thanks.

Adding 2!:55 eliminated the problem of the session remaining open after the
script had terminated.  Unfortunately, it did away with a behavior I was
counting on:  My very simple script only uses status windows -- no forms.  I
was hoping I could have the final status window stay open until the user closed
it.  Without the 2!:55, it works exactly the way I want, excepting that the
session stays open in the task manager.  With it, there isn't even time to read
the final status window, because the terminated session closes it.  Bummer.
But, thanks for solving the mystery.

Is there a way to accomplish the desired behavior using only wdstatus?

Thanks

Daniel

Eric Iverson wrote:
>
> Daniel Torres wrote:
> > ... "J" does not show up. [various runtime mysteries] ...
>
> Windows 95/98 ctrl/alt/del task manager does not list tasks that don't
> have a visible window. Is this the case for J in your examples?
>
> There may be a bug in runtime task termination vs try/catch. Does adding
> an explicit 2!:55 in both the try and catch section help?
>
> --------------------------------------------------------------------------------
> J Forum: for information about this list, see http://www.jsoftware.com/forum.htm
--------------F0ADF3AC87ED1B42C04D456A
 name="torres.vcf"
Content-Description: Card for Daniel Torres
Content-Disposition: attachment;
 filename="torres.vcf"

begin:vcard
n:Torres;Daniel
tel;fax:1-978.383-5817
tel;home:1-305.441-0369
tel;work:1-305.461-6829
x-mozilla-html:TRUE
adr:;;600 Biltmore Way APT 412;Coral Gables;FL;33134-7529;USA
version:2.1
email;internet:torres@SAmerica.com
fn:Daniel Torres
end:vcard

--------------F0ADF3AC87ED1B42C04D456A--

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 10:33:05 1999
From: Charles Fisk 
Subject: RE: Jforum: Need help accessing the Windows Registry
Date: Mon, 26 Jul 1999 09:26:53 -0500

I guess I was trying to find out if there is a way for J to execute "DOS"
style commnads, such as copy, rename, etc.

> -----Original Message-----
> From:	David Ness [SMTP:DNess@home.com]
> Sent:	Thursday, July 22, 1999 10:12 AM
> To:	forum@jsoftware.com
> Subject:	Re: Jforum: Need help accessing the Windows Registry
>
> I must miss the point. What would `Getenv' have to do with renaming?
>
> Charles Fisk wrote:
> >
> > This is interesting, does  2!5 allow you to rename a file? can you
> rename
> > with wild cards?
> >
>
> --------------------------------------------------------------------------
> ------
> J Forum: for information about this list, see
> http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 10:50:14 1999
From: Andrew Nikitin 
Subject: RE: Jforum: Need help accessing the Windows Registry
Date: Mon, 26 Jul 1999 09:46:24 -0500

AFAIK there is no direct way to do this. When I need this I create
temporary .bat file with desired commands and run it with wd 'winexec'.

> -----Original Message-----
> From:	Charles Fisk [SMTP:charles.fisk@usaa.com]
> Sent:	Monday, July 26, 1999 09:27
>
> I guess I was trying to find out if there is a way for J to execute
> "DOS"
> style commnads, such as copy, rename, etc.
>
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 10:55:15 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Need help accessing the Windows Registry
Date: Mon, 26 Jul 1999 10:53:27 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

Yes; look in the wd commands for 'winexec'.

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Charles Fisk
> Sent: Monday, July 26, 1999 10:27 AM
> To: 'forum@jsoftware.com'
> Subject: RE: Jforum: Need help accessing the Windows Registry
>
>
> I guess I was trying to find out if there is a way for J to execute "DOS"
> style commnads, such as copy, rename, etc.
>
> > -----Original Message-----
> > From:	David Ness [SMTP:DNess@home.com]
> > Sent:	Thursday, July 22, 1999 10:12 AM
> > To:	forum@jsoftware.com
> > Subject:	Re: Jforum: Need help accessing the Windows Registry
> >
> > I must miss the point. What would `Getenv' have to do with renaming?
> >
> > Charles Fisk wrote:
> > >
> > > This is interesting, does  2!5 allow you to rename a file? can you
> > rename
> > > with wild cards?
> > >
> >
> >
> --------------------------------------------------------------------------
> > ------
> > J Forum: for information about this list, see
> > http://www.jsoftware.com/forum.htm
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From ljdickey Mon Jul 26 11:09:26 1999
Subject: RE: Jforum: Need help accessing the Windows Registry

     0 !: 1 'mv foo goo'

From owner-jsoftware@lists.interlog.com  Mon Jul 26 12:08:21 1999
From: "Oleg Kobchenko" 
References: 
Subject: Re: Jforum: Need help accessing the Windows Registry
Date: Mon, 26 Jul 1999 19:05:37 +0300
	charset="koi8-r"

Look at this piece of session log:

   wd'winexec "command.com /c dir"'
0
   load'dir'
   dir'isij.*'
isij.ttf    22344 19-May-94 17:12:40

   wd'winexec "command.com /c copy isij.ttf *.qq"'
0
   dir'isij.*'
isij.qq     22344 19-May-94 17:12:40
isij.ttf    22344 19-May-94 17:12:40

But this is very basic functionality. It does not
allow redirection/capture of I/O or at least have
a change to pass and get a string from/to stdin/out.

The 2!:n foreign conjunction somehow is not yet
implemented for Windows. However the help system
contains very promising example, ununequivocally
indicating the OS:

2!:0 'dir *.exe'

The same concerns 1!:1/2, which is only partially
works for stdin/out in Windows.

----- Original Message -----
From: Andrew Nikitin 
Sent: Monday, July 26, 1999 5:46 PM
Subject: RE: Jforum: Need help accessing the Windows Registry

> AFAIK there is no direct way to do this. When I need this I create
> temporary .bat file with desired commands and run it with wd 'winexec'.
>
> > -----Original Message-----
> > From: Charles Fisk [SMTP:charles.fisk@usaa.com]
> > Sent: Monday, July 26, 1999 09:27
> >
> > I guess I was trying to find out if there is a way for J to execute
> > "DOS"
> > style commnads, such as copy, rename, etc.
> >
> >
>
> --------------------------------------------------------------------------------
> J Forum: for information about this list, see http://www.jsoftware.com/forum.htm
>

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 14:18:52 1999
From: gosi@centrum.is
Subject: Re: Jforum: J runtime sessions not fully closing.
Date: Mon, 26 Jul 99 18:15:58 +0000

Daniel Torres wrote:

> The script in question only produce status windows, which may explain why the
> task manager does not list the script as a task while it runs.  Thanks.

>  With it, there isn't even time to read
> the final status window, because the terminated session closes it.  Bummer.
> But, thanks for solving the mystery.
>
> Is there a way to accomplish the desired behavior using only wdstatus?

You can put a delay in if the program is too fast. I use it often together with
status windows.

6!:3 y	Delay.  Delay execution for y seconds.  For example, 6!:3 (2.5)

/Gosi

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 14:47:50 1999
Date: Mon, 26 Jul 1999 14:49:52 -0400
From: Tom Chwastyk 
Organization: US Naval Research Laboratory
Subject: Jforum: Timer vs. Delay [WAS: J runtime sessions not fully closing.]
References: <379ca5de2008001@blik.skima.is>

I find delay awkward because it seems to suck up 100% of the CPU time during delay,
making windows response sluggish. Starting a timer (possibly with a count down in
the status window?) imposes negligible system load. See wd 'timer ' and
sys_timer. Still no need for a form.

That brings up my other question, though. If there is only one sys_timer event
handler, is there an accepted etiquette for coordinating multiple events that all
want to use timer independently and transparently? Or have I missed something about
locales here?

gosi@centrum.is wrote:

> Daniel Torres wrote:
> [snip]
> >  With it, there isn't even time to read
> > the final status window, because the terminated session closes it.  Bummer.
> [snip]
>
>
> You can put a delay in if the program is too fast. I use it often together with
> status windows.
>
> 6!:3 y  Delay.  Delay execution for y seconds.  For example, 6!:3 (2.5)
>
> /Gosi

--
Tom Chwastyk (sounds like Fosdick;
Polish: CH='H, W=F (here), A=AH,
styk=stick, soften strict 'HFAH-stick)

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Mon Jul 26 15:05:10 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Timer vs. Delay [WAS: J runtime sessions not fully closing.]
Date: Mon, 26 Jul 1999 15:04:32 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <379CADD0.227D7AAB@cms.nrl.navy.mil>

There is only one sys_timer, and it runs in the base locale.

If you want to have multiple events, you have to code that
on your own.  In fact, on Windows95 at least, you need a
bit of code even if you have only one event, because Windows
sometimes gives you a sys_timer earlier than you asked for
(I assume this is a Windows, not a J, feature).

I don't know of an accepted etiquette.  If you want something
to start with, I will send you the code that I use to
manage multiple timer threads.

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Tom Chwastyk
> Sent: Monday, July 26, 1999 2:50 PM
> To: forum@jsoftware.com
> Subject: Jforum: Timer vs. Delay [WAS: J runtime sessions not fully
> closing.]
>
...
> That brings up my other question, though. If there is only one
> sys_timer event
> handler, is there an accepted etiquette for coordinating multiple
> events that all
> want to use timer independently and transparently? Or have I
> missed something about
> locales here?
...

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 10:23:11 1999
From: "Seymour Glass" 
Subject: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 10:18:46 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <379ca5de2008001@blik.skima.is>

The problem is,

  Define the verb 'allones' which operates on a boolean
  noun of arbitrary shape, and returns 1 if all the atoms
  are 1 (or there are no atoms), 0 otherwise.

Obviously this is trivial: (*./ @: ,) would work.  But
there is another requirement:

  The verb should use the minimum space possible.

Solutions are to be judged with the following testbench:

   a =: 50 50 50 $ 1
   allones =: NB. 
   7!:2 'allones a'
6848

which is my best result so far.

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 10:47:46 1999
From: Andrew Nikitin 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 09:44:41 -0500

> -----Original Message-----
> From:	Seymour Glass [SMTP:glasss@mindspring.com]
> Sent:	Tuesday, July 27, 1999 09:19
>
>   Define the verb 'allones' which operates on a boolean
>   noun of arbitrary shape, and returns 1 if all the atoms
>   are 1 (or there are no atoms), 0 otherwise.
>
>   The verb should use the minimum space possible.
spoiler follows

> Solutions are to be judged with the following testbench:
>
>    a =: 50 50 50 $ 1
>    allones =: NB. 
>    7!:2 'allones a'
> 6848
As far as all possible inputs (a noun) consist of only ones than
allones=:1:
will be solution.
   allones a
1			NB. correct!
   7!:2 'allones a'
1024

that is less than 6848.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 10:59:25 1999
From: Andrew Nikitin 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 09:56:24 -0500

> -----Original Message-----
> From:	Seymour Glass [SMTP:glasss@mindspring.com]
> Sent:	Tuesday, July 27, 1999 09:19
>
>   Define the verb 'allones' which operates on a boolean
>   noun of arbitrary shape, and returns 1 if all the atoms
>   are 1 (or there are no atoms), 0 otherwise.
>   The verb should use the minimum space possible.
>
> Solutions are to be judged with the following testbench:
>
>    a =: 50 50 50 $ 1
>    allones =: NB. 
>    7!:2 'allones a'
> 6848
Another spoiler:

   a =: 50 50 50 $ 1
   t=:7!:2@('allones a'"_)
   t 1 [. allones=:($:@(*/))`]@.(0: = #@$)
6656
   allones a
1

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 11:15:20 1999
From: "Seymour Glass" 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 11:14:44 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

allones=:($:@(*/))`]@.(0: = #@$)

is a nice solution.  My solution was

allones=:*./^:_

which is the same idea as yours, but since ^:
keeps a copy of v. I extended it to

allones=: *./^:_ @: (*./) @: (*./)

which takes just a smidgen more space.

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Andrew Nikitin
> Sent: Tuesday, July 27, 1999 10:56 AM
> To: 'forum@jsoftware.com'
> Subject: RE: Jforum: J puzzle (all-ones)
>
>
> > -----Original Message-----
> > From:	Seymour Glass [SMTP:glasss@mindspring.com]
> > Sent:	Tuesday, July 27, 1999 09:19
> >
> >   Define the verb 'allones' which operates on a boolean
> >   noun of arbitrary shape, and returns 1 if all the atoms
> >   are 1 (or there are no atoms), 0 otherwise.
> >   The verb should use the minimum space possible.
> >
> > Solutions are to be judged with the following testbench:
> >
> >    a =: 50 50 50 $ 1
> >    allones =: NB. 
> >    7!:2 'allones a'
> > 6848
> Another spoiler:
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>    a =: 50 50 50 $ 1
>    t=:7!:2@('allones a'"_)
>    t 1 [. allones=:($:@(*/))`]@.(0: = #@$)
> 6656
>    allones a
> 1
>
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 11:16:58 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Questions
Date: Tue, 27 Jul 1999 11:11:16 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <379DCA6D.CB6A429B@worldnet.att.net>

> 3) Two of the systems will be put on a network and used by multiple
> users. Each user can have multiple sets of data, e.g., a set of files
> for each calendar quarter. I would like to do something like the
> following:
>
> The program is prog.ijr. A file, sys.ijf, resides in the same
> directory as prog.ijr. When the program is invoked, it determines who
> the user is and shows a list of file sets for the user to choose from.
> The user can also edit this list.
>
> My specific questions are:
>
>    - how does prog.ijr locate sys.ijf?
>    - how does prog.ijr identify a user?
>

I had this very problem last week.  I decided to use the
Windows Registry to hold machine-specific information.
This might make sense if each user runs only from his
desktop computer.

If you want the functions for accessing the Registry
I will send them.

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 11:26:01 1999
From: "Oleg Kobchenko" 
References: 
Subject: Re: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 18:23:23 +0300
	charset="iso-8859-1"

   a=. ?2$~3#50
   7!:2'v a'
6144
   b=. 1$~3#50
   7!:2'v b'
6144
   v a
0
   v b
1

------------------------------
   v
]`($:@:(*./))@.(*@:#@:$)

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 11:26:51 1999
From: Andrew Nikitin 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 10:23:53 -0500

> -----Original Message-----
> From:	Seymour Glass [SMTP:glasss@mindspring.com]
> Sent:	Tuesday, July 27, 1999 10:15
>
> allones=:($:@(*/))`]@.(0: = #@$)
>
> is a nice solution.
allones=:$:@:(*/)`]@.(0&=@#@$)
takes even less space (6400 vs 6656). BTW, I don't understand why.

>   My solution was
> allones=:*./^:_
> which is the same idea as yours, but since ^:
> keeps a copy of v. I extended it to
>
> allones=: *./^:_ @: (*./) @: (*./)
>
> which takes just a smidgen more space.
Good idea. I was scared by huge number given by */^:_
   7!:2 'allones a' [. allones=:*/^:_
137792
and abandoned it immideately.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 12:01:47 1999
Date: Tue, 27 Jul 1999 11:04:13 -0400
From: Brian Bambrough 
Subject: Jforum: Questions

I am working on systems for two clients. There are some things that I
need to do that I can't find in the documentation or by experimenting.

1) How can I give the user access to context sensitive help? Ideally,
I would like to use the right mouse button to provide help for
individual controls. Failing that, is it possible to implement a
"What's this?" question mark control?

2) I use wdview to allow the user to see data. Every time it is used,
the wdview window must be manually maximized to be useful. Is it
possible to invoke wdview so that it is automatically maximized?

3) Two of the systems will be put on a network and used by multiple
users. Each user can have multiple sets of data, e.g., a set of files
for each calendar quarter. I would like to do something like the
following:

The program is prog.ijr. A file, sys.ijf, resides in the same
directory as prog.ijr. When the program is invoked, it determines who
the user is and shows a list of file sets for the user to choose from.
The user can also edit this list.

My specific questions are:

   - how does prog.ijr locate sys.ijf?
   - how does prog.ijr identify a user?

4) In the release notes there is a brief description of all the neat
things that can be done with the J editor. I can't find documentation
on this.

If anyone can help me with any of these questions I would greatly
appreciate it.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 13:37:22 1999
From: "Nichols, Peter" 
Subject: Jforum: Size of Text
Date: Tue, 27 Jul 1999 12:26:55 -0500
	charset="iso-8859-1"

I would like to be able to size multi-line edits and rtf's based on the text
I am going to display.

Is there a way to calculate the size of a control needed to display text
that is variable in characters and fonts?

Thanks!
Peter Nichols

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 13:42:58 1999
From: "Stefano Lanzavecchia" 
References: 
Subject: Re: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 19:37:07 +0200
Organization: APL Italiana
	charset="iso-8859-1"

----- Original Message -----
From: Seymour Glass 
Sent: Tuesday, July 27, 1999 4:18 PM
Subject: Jforum: J puzzle (all-ones)

: The problem is,
:
:   Define the verb 'allones' which operates on a boolean
:   noun of arbitrary shape, and returns 1 if all the atoms
:   are 1 (or there are no atoms), 0 otherwise.
:   The verb should use the minimum space possible.
:
: Solutions are to be judged with the following testbench:
:
:    a =: 50 50 50 $ 1
:    allones =: NB. 
:    7!:2 'allones a'
: 6848

allones=. 13 : ((_1}.;15#<'(*./)@'),'y.')

I believe:

allones<-{not}{jot}(0{jot}{belongsto})

in Dyalog APL is as efficient as if not better. The equivalent solution in J
is prevented from the bizzarre definition of e. which does not accept scalar
left argument against an array on the right of any rank.

I tried several other solution slightly more elegant in J using ^:_ or
^:(#@$) but they all seemed to need more space.
I could not make the gerund based solution to work because of a rank error
that I don't understand. Can somebody please tell me what I am doing wrong?

notscalar=.0: ~: $@$

]`(*./)@.notscalar a
| rank error

What's wrong? Why does:
]`(*./)@.1 a
works instead and clearly
notscalar a
is a 1?
Thanks a lot in advance for you help.
--
WildHeart'99
Homepage: http://www.insight.dk/stf/

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 13:56:06 1999
From: Andrew Nikitin 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 12:53:50 -0500

> -----Original Message-----
> From:	Stefano Lanzavecchia [SMTP:stf@apl.it]
> Sent:	Tuesday, July 27, 1999 12:37
>
> I could not make the gerund based solution to work because of a rank
> error
> that I don't understand. Can somebody please tell me what I am doing
> wrong?
>
> notscalar=.0: ~: $@$
> ]`(*./)@.notscalar a
> | rank error
> What's wrong? Why does:
You shoud define notscalar as

notscalar =: 0 ~: #@$   NB. or *@#@$ that takes less space on huge
arguments

so it return scalar, not 1 element vector. At some point J became even
less tolerant to difference between scalar and 1 element vector.

nsg

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 14:06:11 1999
From: "Stefano Lanzavecchia" 
References: 
Subject: Re: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 20:01:50 +0200
Organization: APL Italiana
	charset="iso-8859-1"

----- Original Message -----
From: Andrew Nikitin 
Sent: Tuesday, July 27, 1999 7:53 PM
Subject: RE: Jforum: J puzzle (all-ones)

: You shoud define notscalar as
:
: notscalar =: 0 ~: #@$   NB. or *@#@$ that takes less space on huge
: arguments
:
: so it return scalar, not 1 element vector. At some point J became even
: less tolerant to difference between scalar and 1 element vector.

Thank you. I was expecting something like that. Yet:
]`(*./)@. (,1) a

works. And it must by definition since
]`(*./)@. (1 1)

is valid (it returns a hook).

There is still something fishy.
--
WildHeart'99
Homepage: http://www.insight.dk/stf/

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 14:24:04 1999
From: Andrew Nikitin 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 13:21:34 -0500

> -----Original Message-----
> From:	Stefano Lanzavecchia [SMTP:stf@apl.it]
> Sent:	Tuesday, July 27, 1999 13:02
> To:	forum@jsoftware.com
> Subject:	Re: Jforum: J puzzle (all-ones)
>
> I was expecting something like that. Yet:
> ]`(*./)@. (,1) a
>
> works. And it must by definition since
> ]`(*./)@. (1 1)
>
> is valid (it returns a hook).
>
Dictionary says:
	m@.n is a verb defined by the gerund m with an agenda specified
by n; that is, the verb represented by the train selected from m by the
indices n.  If n is boxed, the train is parenthesized accordingly.  The
case m@.v uses the result of the verb v to perform the selection.

@. conjunction defined differently for noun and verb case despite the
last sentence. Exactly the same case as with } adverb.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 18:28:29 1999
Date: Tue, 27 Jul 1999 18:25:01 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: RE: Jforum: Questions
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id SAA22355

Message text written by INTERNET:forum@jsoftware.com
>I had this very problem last week.  I decided to use the
Windows Registry to hold machine-specific information.
This might make sense if each user runs only from his
desktop computer.<

I try my best to stay away from the registry.

I would use 4!:4 to find out where prog.ijr is (looking for the script that
made one of my verbs),
which would tell me where the data file is.

As for the user, I would have a configuration file right there or a sign-on
procedure, or a command-line parameter.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Tue Jul 27 23:59:55 1999
Date: Tue, 27 Jul 1999 23:54:48 -0400
From: Brian Bambrough 
Subject: Re: Jforum: Questions
References: 

Henry;

> I had this very problem last week.  I decided to use the Windows
> Registry to hold machine-specific information. This might make sense
> if each user runs only from his desktop computer.

> If you want the functions for accessing the Registry I will send
> them.

Thank you for your kind offer. I truly appreciate it. But . . . modify
the registry? Moi? I do not have that much courage.

This afternoon I came up with a similar solution to the one Alain
offered. In my case I used 4!:40. This tells me where J is. I tested
it out with the script below. I put it into a directory holding j.exe
and j.dll. I glued stdlib.ijs, etc., to the front of it and, from the
DOS command line, ran: j /rt findpath.ijs.

Alain;

> As for the user, I would have a configuration file right there or a
> sign-on procedure, or a command-line parameter.

I was hoping to automate this. My thought is that the network knows
who the user is, so all I have to do is figure out a way to get it to
tell me.

FINDPATH=: 0 : 0
pc findpath;
xywh 27 7 241 184;cc stcShowPath static;
xywh 220 202 50 11;cc btnRun button;cn "Run";
xywh 220 217 50 11;cc btnExit button;cn "Exit";
pas 6 6;pcenter;
rem form end;
)

findpath_run=: 3 : 0
wd FINDPATH
NB. initialize form here
wd 'pshow;'
)

findpath_close=: 3 : 0
wd'pclose'
)

findpath_btnRun_button=: 3 : 0
path=:1!:40 ''
scr=:4!:3 ''
wd 'set stcShowPath "',path,LF,(,(>scr),"1 LF),'"'
)

findpath_btnExit_button=: 3 : 0
findpath_close''
)

findpath_run ''

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 00:54:32 1999
From: "Roger Hui" 
References:  <003001bed856$e3a0b5e0$57a4b9c2@apl.it>
Subject: Re: Jforum: J puzzle (all-ones)
Date: Tue, 27 Jul 1999 21:38:22 -0400
	charset="iso-8859-1"

Stefano Lanzavecchia writes on Tuesday, July 27:

> I believe:
>
> allones<-{not}{jot}(0{jot}{belongsto})
>
> in Dyalog APL is as efficient as if not better. The equivalent solution in J
> is prevented from the bizzarre definition of e. which does not accept scalar
> left argument against an array on the right of any rank.  ...

The definition of e. is not bizarre, just different from APL.
x e. y tells you which of the cells of x are items in y.
The APL behaviour can be obtained readily as the hook
(e. ,) , but the J behaviour can not be easily defined in APL.
(The J definition of e. is probably defined in lots of APL
utility toolkits.)

By the way, when the right argument is a vector, the items are
scalars, and the left argument can be a scalar, so it is not true
that e. does not accept scalar left arguments against an array
on the right of any rank.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 03:38:15 1999
From: "Stefano Lanzavecchia" 
References:  <003001bed856$e3a0b5e0$57a4b9c2@apl.it> <006001bed89b$b586e680$34b0ba89@f3nbp>
Subject: Re: Jforum: J puzzle (all-ones)
Date: Wed, 28 Jul 1999 09:26:30 +0200
Organization: APL Italiana
	charset="iso-8859-1"

----- Original Message -----
From: Roger Hui 
Sent: Wednesday, July 28, 1999 3:38 AM
Subject: Re: Jforum: J puzzle (all-ones)

: Stefano Lanzavecchia writes on Tuesday, July 27:
:
: > I believe:
: >
: > allones<-{not}{jot}(0{jot}{belongsto})
: >
: > in Dyalog APL is as efficient as if not better. The equivalent solution
in J
: > is prevented from the bizzarre definition of e. which does not accept
scalar
: > left argument against an array on the right of any rank.  ...
:
: The definition of e. is not bizarre, just different from APL.
: x e. y tells you which of the cells of x are items in y.
: The APL behaviour can be obtained readily as the hook
: (e. ,) , but the J behaviour can not be easily defined in APL.
: (The J definition of e. is probably defined in lots of APL
: utility toolkits.)
:
: By the way, when the right argument is a vector, the items are
: scalars, and the left argument can be a scalar, so it is not true
: that e. does not accept scalar left arguments against an array
: on the right of any rank.

Of course both the comments are true, and they could not be otherwise. Maybe
"bizarre" is not the word I should have chosen. Maybe "frustrating" (at
least) would have been more appropriate in my case. I can see that the
definition in J, in general, is more powerful because it cover cases that in
APL are simply dreams, and then simple hook (e. ,) does the trick. But it
would appear (from the memory benchmark suggested by the puzzle) that J
needs to take a copy of the right argument in order to apply the hook, and
that does not SEEM (I am not sure about the internals) to be required in
APL.
Before somebody gets strange ideas, I'd like to point out that I am not out
to seek for weaknesses in J or APL, I am just trying to understand the
differences, and the way I do it usually is by direct comparison, because it
helps me remembering, at least at the beginning.
About the second comment: one of my attempts at solving the proposed puzzle
was:
allones=.]`($:@:(-.@(0&e."1)))@.(0: ~: #@$)

Somehow it needs more memory even then the simpler *./@:, but that's beside
the point. I knew (at least I had figured out) it is possible to use e. in a
scalar vs. vector context. I was just surprised I could not extend it. But
enough of this. Thanks for the clear explanation, and while I am at it,
thanks to mr. Andrew, for the clarification about the difference of the
behaviour of @. when used with a noun and when used with a verb.
--
WildHeart'99
Homepage: http://www.insight.dk/stf/

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From ljdickey Wed Jul 28 05:27:56 1999
Subject: Re: Jforum: J puzzle (all-ones)

To "Stefano Lanzavecchia" 

You raise a very interesting point here:

> But it
> would appear (from the memory benchmark suggested by the puzzle) that J
> needs to take a copy of the right argument in order to apply the hook, and
> that does not SEEM (I am not sure about the internals) to be required in
> APL.

I do not know know about the internals either.  But I think
that in either language, J or APL (if the data are not boxed
or enclosed) it is not necessary for ravel to make a copy of
the data, since it is possible to  perform the the interpreter
to carry out the operation ravel by changing only the header
information for the data, and not changing the data itself.

I believe that some APL implementators have done
ravel this way, by writing the internal code to
do this header action without changing the
data, exactly for the reason to save time and space.
But not I am equally sure that not all APL implementors
have done this.  This kind of code was not introduced
into APL until APL was a relatively mature product.

Whether or not the current J code does this, I do not know.
Certainly it could, and it would then, for ravel, it would
be special case code.

Lee Dickey
--
Prof. Leroy J. Dickey, Faculty of Mathematics, U of Waterloo, Canada  N2L 3G1
         ljdickey@math.UWaterloo.CA	1-519-888-4567, ext 5559
	 http://math.uwaterloo.ca/~ljdickey

From stf@apl.it  Wed Jul 28 06:51:55 1999
From: "Stefano Lanzavecchia" 
References: <199907280927.FAA24371@pythagoras.math.uwaterloo.ca>
Subject: Re: Jforum: J puzzle (all-ones)
Date: Wed, 28 Jul 1999 12:26:36 +0200
Organization: APL Italiana
	charset="iso-8859-1"

----- Original Message -----
From: Lee Dickey 
Sent: Wednesday, July 28, 1999 11:27 AM
Subject: Re: Jforum: J puzzle (all-ones)

: or enclosed) it is not necessary for ravel to make a copy of
: the data, since it is possible to  perform the the interpreter

: Whether or not the current J code does this, I do not know.
: Certainly it could, and it would then, for ravel, it would
: be special case code.

Actually J has the possibility of some extra optimization. To be sure I
should be able to put my hands on the book which describes the internals of
J. Even if it's very out of date I am sure the  general structure of the
interpreter has not changed too much (except for the way the header is now
separated from the data to support memory mapped files). I seem to remember,
or maybe somebody told me, that the parser in J is very simple because some
of the pathological problems (like bracket indexing) were removed from the
language. And such a simple parser can be easily driven by a table with all
sort of special cases for idioms. I don't see why, except possibly lack of
time, Roger Hui could not code the hook (e. ,) in case of a simple left
argument, not to restructure the right argument at all, but just pretend
that it's a simple ravelled vector and quickly scan it for matches. In
theory the simple case where the left argument is a scalar could be
amazingly fast and amazingly cheap in terms of memory use.
I don't know if it is really better, but I would imagine that, given b, a
boolean array of any rank, in APL:
    1{belongsto}b
should be quicker and cheaper than a naive implementation of
    {or}{reduce},b

Just to give an example, I was told that in Sharp-APL for mainframes, +/b
was implemented as a table lookup. I am not saying these kind of
optimazations is not possible in modern APL interpreters, but somehow the
nested array extensions, the defined operators, the ambiguity of symbols
like {reduce} which can mean, in a different context, {replicate}, and the
bracket syntax for indexing, axis specification makes the task of the parser
much harder, and much harder seems to me to code special cases for idioms.
When I read the release notes of the new versions of J, I keep on finding
"optimize the following hook/fork/whatever". This makes me think that it's
possible and relatively easy, in J, to code against these cases. So why not
do it? In fact there are many cases where the hook (v ,) could be optimised
in such a way that it's not actually necessary to ravel the right argument
to calculate the result.
Cheers,
--
WildHeart'99
Homepage: http://www.insight.dk/stf/

From owner-jsoftware@lists.interlog.com  Wed Jul 28 08:59:53 1999
Date: Wed, 28 Jul 1999 08:56:21 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Re: Jforum: Questions
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id IAA15041

Message text written by INTERNET:forum@jsoftware.com
>My thought is that the network knows
who the user is<

I like that. I don't know how to get the information (username) but it
seems a good way of going about it if yo can find the information.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 10:07:15 1999
Date: Wed, 28 Jul 1999 10:03:26 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: Current user
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id KAA05459

I suspect that reading the registry key

\HKEY_LOCAL_MACHINE\Network\Logon\usename

would return who the current user is.

You still should use 4!: something to find out where your files are.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 11:49:30 1999
From: "Chris Burke" 
References: <01074BD4EEC4D1118E3200805F6542A7177FD9@SPRDALLAS2>
Subject: Re: Jforum: Size of Text
Date: Wed, 28 Jul 1999 11:45:58 -0400
	charset="iso-8859-1"

> I would like to be able to size multi-line edits and rtf's based on the
text
> I am going to display.
>
> Is there a way to calculate the size of a control needed to display text
> that is variable in characters and fonts?

This can be quite tricky. It is easier to fix the control size and let the
user scroll through text.

You can calculate the size of text using glextentfont. For the "gl" verbs,
you need a form with an isigraph control (which can be hidden).

Size for multiline output line by line. You start by guessing how much will
text will fit on a line, then move backwards and forwards through the text
until you find out exactly what will fit. Then go on to the next line. This
technique is used in the "write" script.

A simple example of using glextentfont is in the script below.

I dont know about sizing for rtf's

===================================

require 'gl2'

LOGXY=: 4 5 { 0 ". wd'qscreen'
SCALE=: LOGXY % 72 * 20

NB. allow for room around the text in the edit control:
PAD=: 10 8

NB. define fonts
NB. TFONT has size in twips (1/20th of point)
TXT=: 'this is a test for qextent 1 2 3 ABC XYZ'

FONT0=: '"MS Sans Serif" 10'
TFONT0=: '"MS Sans Serif" 200'  NB. 20*10

FONT1=: 'Arial 18 bold'
TFONT1=: 'Arial 360 bold'       NB. 20*18

NB. ==============
TEST=: 0 : 0
pc test closeok;
xywh 6 4 106 5;cc g isigraph;
xywh 6 13 209 9;cc s0 static;
xywh 6 24 40 9;cc e0 edit ws_border es_autohscroll;
xywh 6 38 209 13;cc s1 static;
xywh 6 53 40 9;cc e1 edit ws_border es_autohscroll;
xywh 190 76 32 11;cc close button;cn "Close";
pas 6 6;pcenter;ptop;
rem form end;
)

test_run=: 3 : 0
wd TEST

wd 'setshow g 0'
NB. calculate extent with mapping mode in twips
glmap MM_TWIPS
glextent 0
glextentfont TFONT0
size0=. glqextent TXT

glextentfont TFONT1
size1=. glqextent TXT

wd 'set s0 *',FONT0,' (',(":size0),')'
wd 'setfont s0 ',FONT0
wd 'setfont e0 ',FONT0
wd 'set e0 *',TXT
'x y w h'=. 0 ". wd 'qchildxywhx e0'
wd 'setxywhx e0 ',": x,y, PAD+ >. size0 * SCALE

wd 'set s1 *',FONT1,' (',(":size1),')'
wd 'setfont s1 ',FONT1
wd 'setfont e1 ',FONT1
wd 'set e1 *',TXT
'x y w h'=. 0 ". wd 'qchildxywhx e1'
wd 'setxywhx e1 ',": x,y, PAD + >. size1 * SCALE

wd 'pshow;'
)

test_close=: wd bind 'pclose'
test_close_button=: test_close

wd :: ] 'psel test;pclose'

test_run''

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 12:20:46 1999
From: M.Day@fscii.maff.gov.uk
               Wed, 28 Jul 1999 17:02:00 +0100
               Wed, 28 Jul 1999 16:58:27 +0100
               Wed, 28 Jul 1999 17:01:05 +0100
Date: Wed, 28 Jul 1999 17:01:05 +0100
Content-Identifier: m1210728170054aa
Alternate-Recipient: Allowed
In-Reply-To: 
Subject: Re: Jforum: J puzzle (all-ones)

My best (but standing on the shoulders of giants) is

   $a
50 50 50
   allones=: ]`($:@:(*./@(*./)))@.(*@(#@$)) NB. adapted from others'
   space  =: 7!:2
   space'allones a'
5440
   allones a
1

Mike Day

>The problem is,
>  Define the verb 'allones' which operates on a boolean
>  noun of arbitrary shape, and returns 1 if all the atoms
>  are 1 (or there are no atoms), 0 otherwise.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Wed Jul 28 21:44:45 1999
From: "Seymour Glass" 
Subject: RE: Jforum: J puzzle (all-ones)
Date: Wed, 28 Jul 1999 21:44:06 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

Elaborating on Mike's version:

   allones =: (] ` $: @. (*@#@$))@:((*./)@(*./))
   7!:2 'allones a'
5248

order of parenthesizing is important.

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of
> M.Day@fscii.maff.gov.uk
> Sent: Wednesday, July 28, 1999 12:01 PM
> To: Receipt Notification Requested (Non Receipt Notification Requested)
> Subject: Re: Jforum: J puzzle (all-ones)
>
>
> My best (but standing on the shoulders of giants) is
>
>    $a
> 50 50 50
>    allones=: ]`($:@:(*./@(*./)))@.(*@(#@$)) NB. adapted from others'
>    space  =: 7!:2
>    space'allones a'
> 5440
>    allones a
> 1
>
> Mike Day
>
>
> >The problem is,
> >  Define the verb 'allones' which operates on a boolean
> >  noun of arbitrary shape, and returns 1 if all the atoms
> >  are 1 (or there are no atoms), 0 otherwise.
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
> http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Thu Jul 29 12:44:37 1999
From: "Gordon Sutcliffe" 
Subject: Jforum: Problem understanding plot step option from online manuals
Date: Thu, 29 Jul 1999 17:37:17 +0100
	boundary="----=_NextPart_000_002A_01BED9E9.00CA0700"

This is a multi-part message in MIME format.

------=_NextPart_000_002A_01BED9E9.00CA0700
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

In J Release 4.03b for Windows 95 Evaluation Edition, I have tried a =
plot of type step:

   load 'plot'
   'type step' plot i.6
=B3invalid type: step: signal
=B3       signal'invalid type: ',CurType

Please can anyone supply an online manual reference to help me sort the =
problem.

------=_NextPart_000_002A_01BED9E9.00CA0700
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable










In J Release 4.03b for Windows 95 =
Evaluation=20
Edition, I have tried a plot of type step:


 


   load =
'plot'
   'type=20
step' plot i.6
³invalid type: step:=20
signal
³       signal'invalid =
type:=20
',CurType


 


Please can anyone supply an online =
manual=20
reference to help me sort the problem.


------=_NextPart_000_002A_01BED9E9.00CA0700--

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 30 13:40:38 1999
From: "J Tibollo" 
Subject: Jforum: Farpoint Spread Control
Date: Fri, 30 Jul 1999 13:37:33 -0400
Organization: BACK Information Services
	charset="iso-8859-1"

I would like to build a viewer application in J that uses a spreadsheet
control to display output reports.  I have looked at the J spreadsheet
object and have concluded that it is missing too many features and have come
across the Farpoint Spread Control.

Has anyone had any serious problems interfacing to this Control?  From the
documentation the control supports custom printing, and the importing and
exporting of Excel files (ideal for my application).  I used to use the
Formula One Control until they priced themselves out of consideration.

If anyone has had any experience with interfacing the Farpoint Spread
Control to J please take a few minutes to discuss your experience
(especially if you do not recommend it).

Regards,
Joe Tibollo

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Fri Jul 30 23:37:42 1999
Date: Fri, 30 Jul 1999 23:35:15 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: Farpoint Spread Control
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id XAA08332

Message text written by INTERNET:forum@jsoftware.com
>
If anyone has had any experience with interfacing the Farpoint Spread
Control to J please take a few minutes to discuss your experience
(especially if you do not recommend it).<

I have been using it. It works fine. It is my standard spreadsheet control.
J interfaces very well with it. Run time registration works fine.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 01:05:48 1999
From: k-list@iname.com
Date: Sat, 31 Jul 1999 01:03:39 -0400 (EDT)
Content-Type: Text/Plain
Subject: Jforum: crash

from "bill lam" 

it often crashes when closing a form that contains OCX or
OBDC connection. Did anyone have similar experience?
At first I thought it was codedestroy'' or wd'oleenable that
cause null pointer during closing form. But crash still occurs
(but less frequently?) after these verbs were removed.

---------------------------------------------------
Get free personalized email at http://www.iname.com

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 12:24:07 1999
From: "Gilles Kirouac" 
Date: Sat, 31 Jul 99 12:08:51 -0500
Subject: Jforum: Member (In)

On Tue, 27 Jul 1999 21:38:22 -0400, Roger Hui wrote
about J puzzle (all-ones) :

>x e. y tells you which of the cells of x are items in y.

 Dictionary gives the rank of e. as _ _ _ and
 describes dyad e. : If x has the shape of an item of y,
 then x e. y is 1 if x matches an item of y.
 (Note: we are not told what happens if x does NOT have
 the shape of an item of y.)

 Since rank is _ _ for dyad e. , then the cell is
 the whole argument both on the left and the right.

    5 e. i. 2 3     NB.Example 1
 rank error
    $5              NB.Shape of cell of x

    }. $i. 2 3      NB.Shape of an item of y
 3

 As I understand it, a rank error is reported because
 "x does not have the shape of an item of y".

    (i. 2 3) e. 5   NB.Example 2
 0 0 0
 0 0 1

   $i. 2 3   NB.Shape of x
 2 3
   }. $5     NB.Shape of an item of y

   ($i. 2 3) -: }. $5 NB. Do they have the same shape?
 0

 Here since "x does not have the shape of an item of y",
 I would expect the same behaviour as above, a rank error.
 But a result is produced!
 Why does (i. 2 3) e. 5 produce a result?

 Where am I wrong?

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 13:32:35 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Member (In)
Date: Sat, 31 Jul 1999 13:27:10 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <199907311612.MAA26426@riq.qc.ca>

(The following explanation is not official J terminology,
but it helps me understand things.  It is my interpretation
of the description in the Dictionary)

e. (and its cousin dyad i.) can be thought of as having
two ranks.  They don't really have two ranks, of course,
but it helps to think of them that way.  The 'outer rank'
tells the size of the pieces the verb will operate on, and
gives the rank of a derived verb like x@e.; the 'inner rank'
depends on the rank of the operands, and is the rank that
you use for understanding the operation of the verb.

The outer rank of e. and i. is _ _, so x@e. has
infinite rank.

To find the inner rank, you have to know the operands:

   (i. 2 3) e. i. 3
1 1 1
0 0 0

The inner rank was 0 1 .  The rank of y. is 1, so the left inner rank
is 0, one less than the rank of y.  The right outer rank is the rank of
y.

   (i. 2 3) e. 5
0 0 0
0 0 1

The inner rank was 0 0 .

   (i. 2 3) e. ,: i. 3
1 0

The inner rank was 1 2 .

The notion of 'inner rank' also applies to negative ranks.
In the verb u"_1, the _1 produces an inner rank; the outer
rank of u"_1 is _ .  (The way I interpret it, the Dictionary is
inaccurate: it suggests that u"_1 is equivalent to
u"(rank_of_y. - 1), but it really is equivalent to
u"(rank_of_y. - 1)"_, as you can see by

   +"_1 b. 0
_ _ _

or by making experiments with u@(v"_1)  ).

Henry Rich

> -----Original Message-----
> From: owner-jsoftware@lists.interlog.com
> [mailto:owner-jsoftware@lists.interlog.com]On Behalf Of Gilles Kirouac
> Sent: Saturday, July 31, 1999 1:09 PM
> To: forum@JSoftware.Com
> Subject: Jforum: Member (In)
>
>
> On Tue, 27 Jul 1999 21:38:22 -0400, Roger Hui wrote
> about J puzzle (all-ones) :
>
> >x e. y tells you which of the cells of x are items in y.
>
>  Dictionary gives the rank of e. as _ _ _ and
>  describes dyad e. : If x has the shape of an item of y,
>  then x e. y is 1 if x matches an item of y.
>  (Note: we are not told what happens if x does NOT have
>  the shape of an item of y.)
>
>  Since rank is _ _ for dyad e. , then the cell is
>  the whole argument both on the left and the right.
>
>     5 e. i. 2 3     NB.Example 1
>  rank error
>     $5              NB.Shape of cell of x
>
>     }. $i. 2 3      NB.Shape of an item of y
>  3
>
>  As I understand it, a rank error is reported because
>  "x does not have the shape of an item of y".
>
>
>     (i. 2 3) e. 5   NB.Example 2
>  0 0 0
>  0 0 1
>
>    $i. 2 3   NB.Shape of x
>  2 3
>    }. $5     NB.Shape of an item of y
>
>
>    ($i. 2 3) -: }. $5 NB. Do they have the same shape?
>  0
>
>  Here since "x does not have the shape of an item of y",
>  I would expect the same behaviour as above, a rank error.
>  But a result is produced!
>  Why does (i. 2 3) e. 5 produce a result?
>
>  Where am I wrong?
>
>
> ------------------------------------------------------------------
> --------------
> J Forum: for information about this list, see
http://www.jsoftware.com/forum.htm

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 15:11:11 1999
Date: Sat, 31 Jul 1999 15:04:57 -0400
From: Alain Miville de =?ISO-8859-1?Q?Ch=EAne?= 
Subject: Jforum: crash
	 charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by plus.interlog.com id PAA27019

Message text written by INTERNET:forum@jsoftware.com
>it often crashes when closing a form that contains OCX or <

Make sure you have the most recent release of J4.03. I had that, but it is
fixed.

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 17:40:00 1999
From: "Seymour Glass" 
Subject: Jforum: J events at APL99?
Date: Sat, 31 Jul 1999 17:37:34 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: <199907311505_MC2-7F44-5695@compuserve.com>

Does ISI or anybody else plan to have sessions, BOFs, or
the like at APL99 concerning J's future, or jobs
available for J coders, or anything else J?  I've
seen the program but am wondering if anything else
is planned.

Henry Rich

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 17:55:17 1999
From: "Seymour Glass" 
Subject: RE: Jforum: Member (In)
Date: Sat, 31 Jul 1999 17:53:33 -0400
	charset="iso-8859-1"
Importance: Normal
In-Reply-To: 

My previous note was circular in its definition of e., and
I should use our wonderful formal notation to say what I
mean:

edot =: 4 : 'x. +./ @: (-: "_ _1)"((0 >. <: # $ y.),_) y.'"_

which I think gives the right anwers except that it doesn't
detect rank error.  The dependence of the inner rank on the
shape of the data is apparent.

Henry

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm

From owner-jsoftware@lists.interlog.com  Sat Jul 31 22:24:00 1999
Date: Sat, 31 Jul 1999 22:21:01 -0400
From: Brian Bambrough 
Subject: Jforum: Viewing large arrays. User help
Content-Type: multipart/mixed;
 boundary="------------E9EB5F157E3748E74A522E3B"

This is a multi-part message in MIME format.
--------------E9EB5F157E3748E74A522E3B

A few days ago I posted some questions on the J Forum. Two of them
were to do with wdview and "what's this?" kind of help. Nobody had
anything to offer, so I tried to answer them myself. The resulting
script is included as an attachment.

For some time now I have been working on three commercial systems in
J. I have found the wdview facility to be a marvelous convenience.
Yet it has two deficiencies. One is that it doesn't have a capability
to initialize maximized. I was using it so much that I found it a real
nuisance to have to maximize it by hand every time. The other
limitation is that it can't handle big arrays. The attached script is
not limited in these ways.

As a user I am often puzzled as to what a control is supposed to do.
With many Microsoft products there is a "what's this?" question mark
in the top right corner of the window. Click on it, then click on a
control, and the working of the control is explained. As far as I know
there is no "what's this?" facility in J. A long time ago I wrote a
DOS system in BASIC. It used a character mouse. To find out what any
control did, the user simply right-clicked on the control. Simple and
convenient. But, as far as I know, there is no way to do this in J.

So, how to solve the problem of providing help to the user? The
solution I used was to have a "Button help" check box. When the box is
checked, clicking on a button displays a help message. When it's off,
the button works normally.

Actually, I think this is better than "what's this?". Typically, when
a user comes to a new screen he or she wants to see what all the (non-
obvious) controls do. Instead of having to click back and forth
between the question mark and the controls, the box is checked, all
the controls are examined, then the box is unchecked.

The script is fairly complete except for needing a print facility.
Also I have only run it on my 1024x768 screen - it may look weird at
lower resolutions.

If anyone has suggestions for improvements I would be greatly
interested.

--------------E9EB5F157E3748E74A522E3B
Content-Type: application/x-unknown-content-type-ijs_auto_file;
 name="view.ijs"
Content-Transfer-Encoding: base64
Content-Disposition: inline;
 filename="view.ijs"
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--------------E9EB5F157E3748E74A522E3B--

--------------------------------------------------------------------------------
J Forum: for information about this list, see http://www.jsoftware.com/forum.htm