[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Fri, 15 Jul 2005 17:49:33 +0200
Mozilla Thunderbird 1.0.2 (X11/20050404)
John W. Eaton wrote:
This was what I was trying to say in my original post, but since english
isn't my first language, this message properly got lost in translation.
On 15-Jul-2005, Søren Hauberg wrote:
| But it wouldn't be integrated with Octave, right? I wouldn't be able to
| do a "help -i my_function"? So if I create a new package the
| documentation that comes with the package wouldn't be integrated with
Currently this is probably true because Octave only looks in one file
when doing a help -i search. But we can always fix Octave if it does
not do the correct thing.
A way to support this would be to generate an info file at startup (or
the first time in the session help -i is called) that links to the
installed info files. Packages would then need a way of telling Octave
that they provide an info file.
Another way to support this would be if the user asks for "help -i
some_function", where "some_function" lives in the directory DIR, let
Octave check if an info file called DIR/doc.info (or something similar)
exist. If it does, then Octave should use that file, if not Octave
should the default info file.
| I used the .texi files in doc/interpreter. I then used a combination of
| info2html and latex2html. The "see also"'s doesn't work at the moment,
| but it shouldn't be to hard.
I tried playing with makeinfo 4.7 (the latest is 4.8, I'll have a look
at it later on). The problem is that I can't seem to make makeinfo
generate math (either MathML of images). texi2html can call latex2html
and have it generate images for math contents. So makeinfo isn't quite
as good as texi2html. Since one should be able to generate MathML from
TeX, I guess it shouldn't be to big a task to include MathML within the
Octave manual (wrapped in @ifhtml).
Please try using a current version of makeinfo to generate the HTML
and then write a .css file that gives the appearance you want. I
think this has the potential for much better and more reliable results
than any other kind of conversion because makeinfo should be able to
correctly parse all the Texinfo markup.