Re: makeinfo - plans to generate XHTML 1.0 Transitional?

[Gerald Pfeifer]

Hi Karl,

On Mon, 4 Nov 2002, Karl Berry wrote:
>> If Texinfo does generate HTML in the "<br />" style of XHTML, then it
>> no longer validates as HTML.
> Argh!!  I guess I misunderstood Janis and Gerald.  The browsers seem
> happy enough with <br />.  It's very sad that it doesn't validate.

you didn't misunderstand what Janis and me wrote; on the other hand, I
cannot reproduce the problem of the Stephen reported.

<br /> (and <img... /> for that matter) both well validate as HTML 4.01.
In fact they validate both as HTML 4.01 Transitional and even HTML 4.01
Strict, as you can verify yourself by means of the example I have put up
at http://www.dbai.tuwien.ac.at/~pfeifer/tmp/valid.html.

> Personally, I think all these incompatible standards and validations are
> absurd.  That's why I don't wish to output a doctype declaration (for
> HTML output) -- there is nothing reasonable to declare.

I don't think it's that bad: Apart from the <meta> and <link> tags, as
far as I know it is very easy to write pages that qualify both as HTML
4.01 and XHTML, and in fact I am doing that for my personal pages as well
as the GCC pages.

> I just want to output something that displays ok for the actual users
> (including ones who (gasp) aren't using the latest browsers), not
> placate standards committees.

In that case the output of the current pre-release of texinfo should be
fine, according to Janis' and my tests.

> Did the texinfo html output ever validate, even before the <br /> stuff?
> That was never a goal, so if it did, it was an accident :).

We now have texinfo HTML output validating fine as HTML 4.01 on some of
our gcc.gnu.org pages.

Stephen wrote:
> An option to output XHTML would change this declaration, and everybody
> would know what they were getting.  I am happy to see that makeinfo
> already does this with the --xml or --docbook options.

Well, I suppose we violently agree here, but Karl mentioned (in an earlier
discussion) that this, unfortuntely, will prove more difficult to implement
than it seems to, so we somehow should fine a compromise for the pending
release.

Since I failed to reproduce the problem you reported to Karl, perhaps you
could put a small demo page online somewhere, similar to mine, which shows
the problem?

Gerald
-- 
Gerald "Jerry" address@hidden http://www.dbai.tuwien.ac.at/~pfeifer/