💾 Archived View for gemi.dev › gemini-mailing-list › 000215.gmi captured on 2023-11-04 at 12:33:36. Gemini links have been rewritten to link to archived content

View Raw

More Information

➡️ Next capture (2023-12-28)

-=-=-=-=-=-=-

[SPECS text/gemini] Heading lines proposal

cage <cage-dev (a) twistfold.it>

Hi!

When studying the documentation about  the text format (and reading the
mailing list messages) i started  to think that the same argumentation
that lead to adding  a space after the list item  marker (the '*') are
valid to the heading marker as well.

For example if i starts a line mentioning an IRC channel like that:

#channel [...]

this would be presented  to the user as an header of  level 1 which is
not.

The same would be true if i mention a tag for a social network.

As for the  list item i think would  be a good idea to add  a space to
the header mark: so from "#" to  "#<space>" and so on for header level
2 and 3.

Not really sure if this is correct, though so i am looking forward for
your comments.

Bye!
C.

Link to individual message.

solderpunk <solderpunk (a) SDF.ORG>

On Sun, Jun 14, 2020 at 11:07:31AM +0200, cage wrote:

> For example if i starts a line mentioning an IRC channel like that:
> 
> #channel [...]
> 
> this would be presented  to the user as an header of  level 1 which is
> not.
> 
> The same would be true if i mention a tag for a social network.

I guess this is just as good an argument as the *emphasised* words
clashing with list items thing.  AFAIK it hasn't been spotted in the
wild yet, but that's probably just a function of Geminispace being young
and small.

> As for the  list item i think would  be a good idea to add  a space to
> the header mark: so from "#" to  "#<space>" and so on for header level
> 2 and 3.

This would be covered by Petite Abeille's proposal to just put a
mandatory space after *all* the special line indicators.

Is it silly that I'm sad this would require (on account of
### sub-sub-headings) changing "It is possible to unambiguously
determine a line's type purely by inspecting its first three characters"
to "first four characters"?  It is silly.  But three just feels like a
much more *natural* choice of entirely arbitrary small number...

Cheers,
Solderpunk

Link to individual message.

cage <cage-dev (a) twistfold.it>

On Sun, Jun 14, 2020 at 10:16:20AM +0000, solderpunk wrote:

Hi solderpunk!

> On Sun, Jun 14, 2020 at 11:07:31AM +0200, cage wrote:

> > For example if i starts a line mentioning an IRC channel like that:
> >
> > #channel [...]
> >
> > this would be presented  to the user as an header of  level 1 which is
> > not.
> >
> > The same would be true if i mention a tag for a social network.
>
> I guess this is just as good an argument as the *emphasised* words
> clashing with list items thing.  AFAIK it hasn't been spotted in the
> wild yet, but that's probably just a function of Geminispace being young
> and small.

yes, and probably even less likely to occurs than the emphasis.

> > As for the  list item i think would  be a good idea to add  a space to
> > the header mark: so from "#" to  "#<space>" and so on for header level
> > 2 and 3.
>
> This would be covered by Petite Abeille's proposal to just put a
> mandatory space after *all* the special line indicators.
>
> Is it silly that I'm sad this would require (on account of
> ### sub-sub-headings) changing "It is possible to unambiguously
> determine a line's type purely by inspecting its first three characters"
> to "first four characters"?  It is silly.  But three just feels like a
> much more *natural* choice of entirely arbitrary small number...

:) Well i think i can sort of understand, if i would need to choose to
change  the  sentence  above  to  "It  is  possible  to  unambiguously
determine  a  line's  type  purely  by  inspecting  its  first  *four*
characters", and  remove the  sentence for  inexplicable (even  to me)
reason i would choose the second! ;-D

Anyway, i am starting writing a parser, it is refreshing (for my mind)
the fact that the syntax is so simple and still effective! :)

Bye!
C.

Link to individual message.

Alex Schroeder <alex (a) gnu.org>

Hi all
I want to return to a heading lines proposal raised by cage back in
June.
https://lists.orbitalfox.eu/archives/gemini/2020/001662.html

I would like headers to require a space after the #, at least for the
first level.

Headings:

# Foo
## Bar
### Baz
##Bar if you really want to
###Baz if you really want to

The reason being that I'd love to use hashtags in my posts and I'd love
to add them at the end, like this:

#Backup #Administration

I guess I can live with the requirement to add some filler text in
front of the first tag:

Tags: #Backup #Administration

But I think requiring a space after level 1 headings would make more
sense. What do you think?

Cheers
Alex

Link to individual message.

Ben <benulo (a) systemli.org>

It should be rare for hashtags to appear at the beginning of a line 
(even in your example they didn't), but in such a case there could 
indeed be a conflict. In such cases perhaps something like \# could be 
used to start a line.

The space idea also seems good to me and maybe more efficient. 
Personally I always space my headers as you suggested.

Ben

-- 
gemini://kwiecien.us/

Link to individual message.

Sebastian Higgins <bctnry (a) outlook.com>

IMO a better (way better) option will be just use `=` for headings but 
that conflicts with current spec & all the existing software. Requiring 
(at least one) space after # is equally good as long as it's applied to 
all headings & not just level 1 headings; it is probably not a good idea 
to introduce new corner cases when doing changes like this in spec.
________________________________
From: Gemini <gemini-bounces@lists.orbitalfox.eu> on behalf of Alex 
Schroeder <alex@gnu.org>
Sent: Wednesday, July 22, 2020 7:55
To: Gemini <gemini at lists.orbitalfox.eu>
Subject: [SPECS text/gemini] Heading lines proposal

Hi all
I want to return to a heading lines proposal raised by cage back in
June.
https://lists.orbitalfox.eu/archives/gemini/2020/001662.html

I would like headers to require a space after the #, at least for the
first level.

Headings:

# Foo
## Bar
### Baz
##Bar if you really want to
###Baz if you really want to

The reason being that I'd love to use hashtags in my posts and I'd love
to add them at the end, like this:

#Backup #Administration

I guess I can live with the requirement to add some filler text in
front of the first tag:

Tags: #Backup #Administration

But I think requiring a space after level 1 headings would make more
sense. What do you think?

Cheers
Alex

Link to individual message.

Gary Johnson <lambdatronic (a) disroot.org>

I'm in favor of requiring at least one space after all headline levels for consistency.

Sebastian Higgins <bctnry at outlook.com> writes:

> IMO a better (way better) option will be just use `=` for headings but 
that conflicts with current spec & all the existing software. Requiring 
(at least one) space after # is equally good as long as it's applied to 
all headings & not just level 1 headings; it is probably not a good idea 
to introduce new corner cases when doing changes like this in spec.
> ________________________________
> From: Gemini <gemini-bounces at lists.orbitalfox.eu> on behalf of Alex 
Schroeder <alex at gnu.org>
> Sent: Wednesday, July 22, 2020 7:55
> To: Gemini <gemini at lists.orbitalfox.eu>
> Subject: [SPECS text/gemini] Heading lines proposal
>
> Hi all
> I want to return to a heading lines proposal raised by cage back in
> June.
> https://lists.orbitalfox.eu/archives/gemini/2020/001662.html
>
> I would like headers to require a space after the #, at least for the
> first level.
>
> Headings:
>
> # Foo
> ## Bar
> ### Baz
> ##Bar if you really want to
> ###Baz if you really want to
>
> The reason being that I'd love to use hashtags in my posts and I'd love
> to add them at the end, like this:
>
> #Backup #Administration
>
> I guess I can live with the requirement to add some filler text in
> front of the first tag:
>
> Tags: #Backup #Administration
>
> But I think requiring a space after level 1 headings would make more
> sense. What do you think?
>
> Cheers
> Alex


-- 
GPG Key ID: 7BC158ED
Use `gpg --search-keys lambdatronic' to find me
Protect yourself from surveillance: https://emailselfdefense.fsf.org
=======================================================================
()  ascii ribbon campaign - against html e-mail
/\  www.asciiribbon.org   - against proprietary attachments

Please avoid sending me MS-Office attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html

Link to individual message.

defdefred <defdefred (a) protonmail.com>

On Wednesday, July 22, 2020 9:55 AM, Alex Schroeder <alex at gnu.org> wrote:
> #Backup #Administration

Seems that hashtag don't like to be alone...
Maybe 1 x #word could be a heading and several x #word could be hashtags...

freD.

Link to individual message.

luna <luna (a) owlsne.st>

I'm certainly in favor of this; It could help make parsing headers a bit
easier, make formatting more consistent and prevent problems with these
sort of situations.

~luna



-- Email domain proudly hosted at https://migadu.com

Link to individual message.

luna <luna (a) owlsne.st>

I'm certainly in favor of this; It could help make parsing headers a bit
easier, make formatting more consistent and prevent problems with these
sort of situations.

~luna



-- Email domain proudly hosted at https://migadu.com

Link to individual message.

Alex Schroeder <alex (a) gnu.org>

On Wed Jul 22, 2020 at 10:06 PM CEST, defdefred wrote:
> Seems that hashtag don't like to be alone...
> Maybe 1 x #word could be a heading and several x #word could be
> hashtags...

I have my doubts. On my blog I have 4788 pages with tags but 2325 of
them have just one tag.

Cheers
Alex

Link to individual message.

Paper <paper (a) tilde.institute>

I think geminitext should be readable even when it is not being
rendered. Something like this isn't readable:

##something

but this is:

## something

So I like this proposal. I think everyone is using a space
even though it's not mandatory yet.

Paper

Link to individual message.

defdefred <defdefred (a) protonmail.com>

On Thursday 23 July 2020 13:56, Alex Schroeder <alex at gnu.org> wrote:
> I have my doubts. On my blog I have 4788 pages with tags but 2325 of
> them have just one tag.
Leasiness?
:-)

Link to individual message.

Solderpunk <solderpunk (a) posteo.net>

Okay, I am happy with this idea (in fact, for a minute or two I was
thinking "didn't we already do this?!" - no, that was with list items,
to avoid ambiguity with *this* kind of emphasis).

At this point it seems like we may as well go "whole hog" and insist on
at least one space after *all* the line type markers - that would mean
quote lines starting with "> " and link lines starting with "=> ".  Does
anybody object to *that*?  I strongly suspect that almost everybody is
actually writing their content that way anyway, but perhaps not...

Cheers,
Solderpunk

Link to individual message.

Petite Abeille <petite.abeille (a) gmail.com>



> On Jul 23, 2020, at 21:33, Solderpunk <solderpunk at posteo.net> wrote:
> 
> At this point it seems like we may as well go "whole hog" and insist on
> at least one space after *all* the line type markers - that would mean
> quote lines starting with "> " and link lines starting with "=> ". 

Amen.

Link to individual message.

luna <luna (a) owlsne.st>

Sounds like a good plan to me :D

~luna



-- Email domain proudly hosted at https://migadu.com

Link to individual message.

acdw <acdw (a) acdw.net>

On 2020-07-23 (Thursday) at 19:33, Solderpunk <solderpunk at posteo.net> wrote:

> Okay, I am happy with this idea (in fact, for a minute or two I was
> thinking "didn't we already do this?!" - no, that was with list items,
> to avoid ambiguity with *this* kind of emphasis).
> 
> At this point it seems like we may as well go "whole hog" and insist on
> at least one space after *all* the line type markers - that would mean
> quote lines starting with "> " and link lines starting with "=> ".  Does
> anybody object to *that*?  I strongly suspect that almost everybody is
> actually writing their content that way anyway, but perhaps not...


The only possible problem is that, with '###' lines, clients will have to 
read *four* characters from the beginning of a line to determine what it 
is -- but I don't think that's asking too much.

I am 100% on board with this proposal.

-- 
~ acdw
acdw.net | breadpunk.club/~breadw

Link to individual message.

cage <cage-dev (a) twistfold.it>

On Thu, Jul 23, 2020 at 09:33:43PM +0200, Solderpunk wrote:

Hi!

> At this point it seems like we may as well go "whole hog" and insist on
> at least one space after *all* the line type markers - that would mean
> quote lines starting with "> " and link lines starting with "=> ".

In my opinion this is the best thing to do!

(even if i have to change the parser a bit ;-D ;-D)

Bye!
C.

Link to individual message.

colecmac@protonmail.com <colecmac (a) protonmail.com>

> At this point it seems like we may as well go "whole hog" and insist on
> at least one space after all the line type markers - that would mean
> quote lines starting with "> " and link lines starting with "=> ". Does
> anybody object to that?


I do sort of object to this. I don't think there's much value in switching
the other definitions when a need hasn't been demonstrated. I have seen many
links that don't have a space after the "=>", and I don't see why there needs to
be a space after the quote marker either. It's nicer to read in both cases,
but I'm not in favour of creating changes when they don't seem to be required.

Why not just change the heading lines?

Cheers,
makeworld

Link to individual message.

Mad Ogrit <madogrit0 (a) gmail.com>

Requiring a space for structural formatting seems wise. I don't believe the
proposal implied that navigational formatting (=>) would require one.

But regarding quotes, although the > prefix may be visually familiar from
text threads its use is inconsistent at best with regard to space before or
after. Requiring a space will definitely break multiply-nested multi-line
things (or make them redundant).

Which brings me to... If quotes are only meant for a single depth, but
represent semantic rather than structural formatting (sourced content,
could reasonably have an author or other meaningful attribution), would it
make more sense to treat them like preformatted lines (line toggled) and
retire the prefix entirely?

If they are really only for visual distinction... Well I guess I don't
think they ought to be.

So as long as you're talking about updating the spec...


Context:

Hi! Long time observer (nearly a week!) first time poster. I hadn't seen
deedum when I first began using my enthusiasm for both classic gopher and a
lighter better secure-er web to begin writing dart/flutter libraries for
Gemini.

I'm a very low vision user and this approach to content is very welcome.
It's been a fun exercise and there still seems to be room for a formal
library on pubget and so forth, and I am imagining a kind conversational
model where browsing and contributing are possible within the same
experience. Anywho...

Things I noticed in this process where I've tried to include robust edge
case tests and simulate workflows that may be common to streamline
anybody's ability to create content:

preformatted content in that, while it can be wrapped and flowed, you're
often quoting something by copy-pasting at which point the original layout
can be relevant

levels with hideous line breaking

least useful.

specify more quote-like behavior

I'll share links in a few days when the repo moves from embarrassing to
merely incomplete.


On Thu, Jul 23, 2020, 6:08 PM <colecmac at protonmail.com> wrote:

> > At this point it seems like we may as well go "whole hog" and insist on
> > at least one space after all the line type markers - that would mean
> > quote lines starting with "> " and link lines starting with "=> ". Does
> > anybody object to that?
>
>
> I do sort of object to this. I don't think there's much value in switching
> the other definitions when a need hasn't been demonstrated. I have seen
> many
> links that don't have a space after the "=>", and I don't see why there
> needs to
> be a space after the quote marker either. It's nicer to read in both cases,
> but I'm not in favour of creating changes when they don't seem to be
> required.
>
> Why not just change the heading lines?
>
> Cheers,
> makeworld
>

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great Mad Ogrit once stated:
> Requiring a space for structural formatting seems wise. I don't believe the
> proposal implied that navigational formatting (=>) would require one.
> 
> But regarding quotes, although the > prefix may be visually familiar from
> text threads its use is inconsistent at best with regard to space before or
> after. Requiring a space will definitely break multiply-nested multi-line
> things (or make them redundant).

  The gemtext spec doesn't allow nesting at all, so multply-nested
multi-line things are already a problem (as I found out when trying to
convert HTML to gemtext).

> Which brings me to... If quotes are only meant for a single depth, but
> represent semantic rather than structural formatting (sourced content,
> could reasonably have an author or other meaningful attribution), would it
> make more sense to treat them like preformatted lines (line toggled) and
> retire the prefix entirely?

  No.  My blog (phlog, gemlog) makes extensive use of both <BLOCKQUOTE> and
<PRE> tags.  For instance this post from last year:

	gemini://gemini.conman.org/2019/03/31.1

In the originial HTML (http://boston.conman.org/2019/03/31.1), the format is
something long the lines of:

<BLOCKQUOTE>
  <P>...</P>

  <P>...</P>

</BLOCKQUOTE>

<P>...</P>

<PRE>
...
</PRE>

<P>...</P>

  And while I had to go back quite a bit to find an entry with both a
blockquote and preformatted text, I do use blockquotes and preformatted text
quite often, and they have different semantic meanings.

  That post, by the way, shows how I deal with ordered lists in HTML (when
gemtext only supports a type of unordered list).

> * Quoted content when visually distinct has more in common with
> preformatted content in that, while it can be wrapped and flowed, you're
> often quoting something by copy-pasting at which point the original layout
> can be relevant

  True.  Trying to quote preformatted text is not easy with gemtext.  That's
one of the reasons I still prefer HTML (which Gemini *can* serve, by the
way).

  -spc

Link to individual message.

Ryan Knipe <sario528 (a) gmail.com>

I'm on board. It improves human readability, and ensures consistency.

On Thu, Jul 23, 2020, 2:38 PM Solderpunk <solderpunk at posteo.net> wrote:

> Okay, I am happy with this idea (in fact, for a minute or two I was
> thinking "didn't we already do this?!" - no, that was with list items,
> to avoid ambiguity with *this* kind of emphasis).
>
> At this point it seems like we may as well go "whole hog" and insist on
> at least one space after *all* the line type markers - that would mean
> quote lines starting with "> " and link lines starting with "=> ".  Does
> anybody object to *that*?  I strongly suspect that almost everybody is
> actually writing their content that way anyway, but perhaps not...
>
> Cheers,
> Solderpunk
>

Link to individual message.

Mad Ogrit <madogrit0 (a) gmail.com>

On Thu, Jul 23, 2020, 8:44 PM Sean Conner <sean at conman.org> wrote:

> It was thus said that the Great Mad Ogrit once stated:
> > Requiring a space for structural formatting seems wise. I don't believe
> the
> > proposal implied that navigational formatting (=>) would require one.
> >
> > But regarding quotes, although the > prefix may be visually familiar from
> > text threads its use is inconsistent at best with regard to space before
> or
> > after. Requiring a space will definitely break multiply-nested multi-line
> > things (or make them redundant).
>
>   The gemtext spec doesn't allow nesting at all, so multply-nested
> multi-line things are already a problem (as I found out when trying to
> convert HTML to gemtext).
>

Right, so unless I'm willing to manually clean up copy-pasted material I
end up with a bunch of individual lines treated as quote lines that then
also contain superfluous text like >>>, >> > >, and >|>>|.

>
> > Which brings me to... If quotes are only meant for a single depth, but
> > represent semantic rather than structural formatting (sourced content,
> > could reasonably have an author or other meaningful attribution), would
> it
> > make more sense to treat them like preformatted lines (line toggled) and
> > retire the prefix entirely?
>
>   No.  My blog (phlog, gemlog) makes extensive use of both <BLOCKQUOTE> and
> <PRE> tags.  For instance this post from last year:
>
>

Sorry, I think I was being imprecise. I wasn't proposing to not support
quotes as a separate concept from pre, but rather add an additional line
toggle, eg ___ or something unlikely to appear in nominal plaintext, which
would be handled in the same manner as pre - a toggle for blockquote,
allowing for things like:

So I tried
 ```lang=go
func primes ()...
 ```
But Hofstadter's paper said
___src=gemini://princip.ia/hoff.gmi?335
Remember to add a terminal case like
 ```lang=python
def isPrime()...
 ```
___

(The src= is just a silly example and definitely serves a different purpose
than lang=)

Of course, you made me realize this would significantly complicate the
toggle approach in the case of unmatched toggles overlapping and the
inevitable nesting of alternating toggles.

Come to think of it, I assume that pasting gemtext containing pre inside of
pre toggles would still require some manual editing, for instance to add
white space before the embedded pre. I'm guessing this is why nesting of
any kind is not supported :-)

I haven't fully grokked the archives yet so this may have already been
discussed.

>
>
>   True.  Trying to quote preformatted text is not easy with gemtext.
> That's
> one of the reasons I still prefer HTML (which Gemini *can* serve, by the
> way)
>

My hope with supporting and popularizing gemtext is specifically to
counteract the ways in which HTML has become very easy to make very
inaccessible - even just the semantics of layout and flow. I'm in it to
make and support an accessible client with font sizes I can read /
screen-read no matter what the content is :)

Even markdown, as many things do, has become just this side of easy to make
hard to read and hard to make easy to edit and copy paste universally, so I
do really value the notion of an intrinsic format that is easy to
understand and reliably forward only parse-able.

I originally felt the line toggle approach seemed like an odd burden on the
client given the simplicity and line orientedness of the rest of the
format, but I recognize it has the benefit of not requiring an an author to
touch up content, reducing the barrier to entry of copy paste or generally
including things that are otherwise hard to reformat (eg turn a funky table
into tsv and dump it in pre).

Obviously, the above silly example could just be decomposed into a single
quote line, a single pre-block, and a link. But doing that requires the
person composing this text to think about and do that decomposition
possibly from text that they are simply copy pasting an existing pre-block
in existing  gemtext.

Link to individual message.

Gary Johnson <lambdatronic (a) disroot.org>

I also support the "spaces after all line type markers" idea, And I
agree with the other poster (Paper?) that one of gemtext's best
qualities is that it is intrinsically (and intuitively) human-readable
even without rendering. It is, in fact, a totally WYSIWYG format.

Into the future!
  Gary

P.S. Astrobotany is going to need to update its links though...


Ryan Knipe <sario528 at gmail.com> writes:

> I'm on board. It improves human readability, and ensures consistency.
>
> On Thu, Jul 23, 2020, 2:38 PM Solderpunk <solderpunk at posteo.net> wrote:
>
>> Okay, I am happy with this idea (in fact, for a minute or two I was
>> thinking "didn't we already do this?!" - no, that was with list items,
>> to avoid ambiguity with *this* kind of emphasis).
>>
>> At this point it seems like we may as well go "whole hog" and insist on
>> at least one space after *all* the line type markers - that would mean
>> quote lines starting with "> " and link lines starting with "=> ".  Does
>> anybody object to *that*?  I strongly suspect that almost everybody is
>> actually writing their content that way anyway, but perhaps not...
>>
>> Cheers,
>> Solderpunk
>>


-- 
GPG Key ID: 7BC158ED
Use `gpg --search-keys lambdatronic' to find me
Protect yourself from surveillance: https://emailselfdefense.fsf.org
=======================================================================
()  ascii ribbon campaign - against html e-mail
/\  www.asciiribbon.org   - against proprietary attachments

Please avoid sending me MS-Office attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html

Link to individual message.

cage <cage-dev (a) twistfold.it>

On Thu, Jul 23, 2020 at 10:08:02PM +0000, colecmac at protonmail.com wrote:

Hi!

> > At this point it seems like we may as well go "whole hog" and insist on
> > at least one space after all the line type markers - that would mean
> > quote lines starting with "> " and link lines starting with "=> ". Does
> > anybody object to that?
>
>
> I do sort of object to this. I don't think there's much value in switching
> the other definitions when a need hasn't been demonstrated. I have seen many
> links that don't have a space after the "=>", and I don't see why there needs to
> be a space after the quote marker either. It's nicer to read in both cases,
> but I'm not in favour of creating changes when they don't seem to be required.
>
> Why not just change the heading lines?

This is a good question, let me show my opinion about why i agree with
these changes.

In short  to me, the reason  to put a space  after ">" or "=>"  is the
same that changed  "*" to "*<space>".  If, in a  distant future, a new
convention appears for identify a word, or a text starting with a word
using  ">"  (for   example)  the  gemini  parser  will   steps  in  an
inconvenience  like the  ones we  are discussing  in this  thread with
"#". Because  is (in  my opinion)  unlikely that a  person (i  mean an
human being  :-)) writing a text  that will be read  by other persons,
marks  a word  using a  string prefix  that include  a word  separator
(including space)  after and  before or  in the middle  of it  i think
would be a good  idea change the format to require a  space for any of
the symbols Solderpunk mentioned.

This is, of course, probably not  the only reason to make this changes
(and maybe is not very valid :)) but i can not find any better at this
time! ;D

I hope i was able to explain my point even with my bad English!

Bye!
C.

Link to individual message.

Solderpunk <solderpunk (a) posteo.net>

On Fri Jul 24, 2020 at 12:08 AM CEST,  wrote:

> Why not just change the heading lines?

I guess my main motivation was to make the spec more regular/consistent.
If half the line type markers require a space and half of them don't,
writing an correct client requires constant reference back to the spec
to double check these fine details.  Ideally Gemini should "fit in your
head" easily, as somebody else described it, and consistency goes a long
way toward that.

Cheres,
Solderpunk

Link to individual message.

---

Previous Thread: [ANN] Git front-end for Gemini

Next Thread: gemini-mode.el is now in MELPA