Aucbvax.5016 fa.editor-p utzoo!decvax!ucbvax!editor-people Sat Nov 7 18:18:44 1981 EDITOR-PEOPLE digest >From Admin.JQJ@SU-SCORE Sat Nov 7 17:46:26 1981 EDITOR-PEOPLE Digest Saturday, 7 Nov 1981 Volume 1 : Issue 1 Today's Topics: strong typing, Xanadu/Hypertext Note: I'm experimenting with a digest format today, but will probably go back to REMAILing messages tomorrow. Comments to EDITOR-PEOPLE-REQUEST. ---------------------------------------------------------------------- Date: 7 Nov 1981 05:33:02-PST From: ihuxl!jej at Berkeley Subject: strong typing Briefly, since this is editor-people: 1. Isn't it the case that 'packages' avoid redeclaration of external variables? (Besides, there is argument about externals being hazardous to one's health.) 2. Doesn't one at some point have to define what polymorphic functions mean for various combinations (or, heaven forbid, all combinations) of types? Your example of MEMBER simply pushes the proliferation of definitions from MEMBER to the definition of 'matches'. A typed language would not necessarily force one to declare the type of the elements, at least not at the level at which MEMBER is written--such languages as ECL or Russell, which treat 'type' as a type, would allow the type to be passed as a parameter. (Admittedly that may be cumbersome, but people more familiar with the languages might be able to tell us whether the construct type(x) is available, which would make things easier.) I don't think that I'm dogmatic about typing--perhaps in an interactive environment it becomes obnoxious--but one grows paranoid when one sees the messes that result in languages which allow confusion of types, such as C, or the dogged attachment to low-level access via C[AD]+R that one finds in LISP programs that use lists as structures (or records, for Pascal folk). James Jones (ihuxl!jej) -------------------------------- Date: 7 Nov 1981 06:57:54-PST From: cbosgd!mark at Berkeley Subject: strong typing Strong typing can be a pain when first writing a program because the compiler will fuss and fret over little nits that it should have been able to figure out. But it will also make you write cleaner code that will port more easily to other places, and it finds more bugs at compile time than a loosely typed language. I feel that Ada's typing is about right, certainly Pascal's is too loose. I'd make a statement about LISP but no doubt dozens of people would write to me telling me that whatever I said was wrong for THEIR dialect of LISP. Perhaps RMS has never heard of unions. While slightly ugly to use, they very cleanly solve the problem of having a variable (or list or whatever) with elements which can be one of many types. Even Pascal has variant records, but since you can't return records from functions in Pascal, you would have to pass them by reference as parameters. Ugly but not impossible. But the deficiencies of Pascal are no reason to conclude that strong typing is bad - I personally wish I had a much fussier compiler than I currently work with. -------------------------------- Date: 7 Nov 1981 0946-PST From: Craig Taylor Subject: More Xanadu(C) If the book Brian mentioned is "Literary Machines", which discusses Xanadu, hypertext, etc., the first edition is available from Ted by sending a valid check for $15 to the previously mentioned address in PA. It's written as a hypertext and is amusing to read but not very technical. By the way, Xanadu is being claimed as a trademark by Ted, so maybe we'll all be paying him royalties soon. /Craig ------- ------- ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.