Re: [GSoC] Proposal for JavaScript for info-style navigation.

[Mathieu Lirzin]

Per Bothner <address@hidden> writes:

> On 03/30/2017 03:40 PM, Mathieu Lirzin wrote:
>
>> Per Bothner <address@hidden> writes:
>>> While I re-emphasize this is my personal roadmap of how I think texinfo 
>>> should
>>> evolve (not everyone agrees ...), I think there is general agreement of the
>>> value of a nice JavaScript-based UI for texinfo/html documents.
>>
>> Yes, that was my understanding.  I guess the replacement of the info
>> format with hinfo and the use of epub for packaging it are the more
>> controversial propositions?
>
> I'm guessing, but I can't speak for anyone else.
>
>> If I remember past discussion on this ML there was a disagreement
>> regarding wether HTML export should use HTML4 or HTML5 standard (the
>> later would help manipulating the DOM with Javascript) but since the
>> goal is to use a HTML5 compatible format for the xhtml output.  I guess
>> this is not a controversy anymore.
>
> I'm not sure xhtml output is the best way to go.  I suspect the best is
> "polyglot" output, with ".html" extension, valid as html, and well-formed as 
> xml.
> ("Well-formed" and "valid" are technical XML terms.)
>
> One point that I didn't realize in the earlier discussion
> (https://lists.gnu.org/archive/html/bug-texinfo/2016-02/msg00056.html):
> The form <a name="LABEL"> is not "deprecated" in HTML5 - it is not
> valid HTML5 at all.  While using id="LABEL" is valid in both html4 and html5.
>
> Which IMO makes using id attributes a no-brainer.

Thanks for the precision.  The last message of this discussion
(https://lists.gnu.org/archive/html/bug-texinfo/2016-02/msg00080.html)
seems to suggest that a solution for providing "id" in the output is
already available.  However I don't think the patch Gavin provided has
been applied yet.

Anyway this is a good indication that the Javascript UI might be able to
work without too much trouble with the HTML output of 'makeinfo'.

>> What are the known controversial parts of your personnal roadmap?
>
> Not sure.  Probably dropping info format.  But let's focus on html format.
>
>>> Do you know Perl?  Are you comfortable making changes to the makeinfo 
>>> source,
>>> which is mostly Perl?  It not necessary, and it may actually be an unwise
>>> distraction from focusing on the JavaScript, just using the existing 
>>> kawa.epub.
>>> However, the way kawa.epub is built by first translating to DocBook isn't
>>> really satisfactory, so a some point we want to use makeinfo's html output
>>> directly, which means cleaning up and enhancing the the latter,
>>
>> I have some basic Perl knowledge, however modifying the makeinfo source
>> would require a deep study of the AST structure and the output
>> generation mechanism before being able to actually work on it (I would
>> estimate the effort to 2~3 weeks).  If the Javascript implementation
>> goes well, I can definitely spend some time on the html/xhtml5 output.
>
> We do have another student who expressed interest in the "improve makeinfo"
> aspect.  It's likely texinfo will only get funding for a single GSoC
> student - but if there are extras, and we have two strong proposals, we
> might talk the GNU GSoC people into giving texinfo one more, as we "gave
> back" a slot last year.

That would be great.

Thanks.

-- 
Mathieu Lirzin
GPG: F2A3 8D7E EB2B 6640 5761  070D 0ADE E100 9460 4D37