Re: HyperDoc replacement, was: Re: [Axiom-developer] CATS - Computer Algebra Test Suite

[Arthur Ralfs]

Martin,

I took a look at the tex4ht package.  Maybe I should tell you
what my "agenda" is.

I am motivated to make mathematical content and functionality
as easy as possible to access on the web.

To me the only possibility for doing this is by using a web
browser as the interface.

Philosophically any software I use must be Free in the sense
of Richard M. Stallman.

I want a potential user to be able to navigate to a URL in their
browser and be exposed to full mathematical content and
functionality.  To be as easy as possible the user should not
have to download or install any other packages or plugins.
This means using only native browser capability.  I think I
have demonstrated that this is possible although of course
much work remains to be done.

The tex4ht package looks like it's too hard to install.  I don't
run a debian based system.  The only way tex4ht would fit
with my plan is if it was bundled and configured with axiom.

I am not against other approaches.  The more the merrier.
But for  now I am only interested in the above.

Arthur


Martin Rubey wrote:
> root <address@hidden> writes:
>
>   
>> I found only 1 mention of tex4ht on MathAction.  It appears in the summer of
>> code section.  Do you have a reference?
>>     
>
> http://www.cse.ohio-state.edu/~gurari/TeX4ht/mn.html
>
>
>
> I have written the beginnings of a replacement for hyperdoc a while ago. I 
> even
> advertised it here, but nobody showed any interest, so I didn't pursue it
> further.
>
> You can find it on 
>
> http://wiki.axiom-developer.org/SandBoxHyperDocReplacement
>
> Before you can use it, you have to install tex4ht, which is standard with, for
> example,
>
> apt-get install tex4ht
>
> As far as I remember, it also depends on a few fixes at the time only found in
> wh-sandbox, but I imagine they have found their way to axiom already.
>
> I'd love to see it resurrected.  It provides dynamic lookup, dynamic 
> generation
> of html -- or mathml, jsmath, etc., that's trivial to configure.
>
> It assumes that people do not have super fast machines, so the results of the
> translation to html are cached.  This could certainly be done better.  I
> thought one would generate the necessary html database during build time,
> adding a constructor would update the database dynamically.  Since I am not
> into databases, I did not waste any time to design a sensible structure, but
> that certainly could be done.
>
> Again, the only point of my little program was to show that it should be 
> fairly
> easy to replace the hyperdoc rendering by a browser rendering, while keeping
> all the other functionality.
>
> Martin
>
>
>