[texhax] Building CHM help

[Robin Fairbairns]

> > Use pdftex to generate the printed documentation and another tool such
> > as latex2html to make the .html, then process the .html w/ Microsoft's
> > HTML Help Workshop or some similar tool to make the .chm
> 
> That's exactly what I've tried first. However latex2html appeared to
> be quite old and besides not very simple to use. That's why I decided
> to ask here if there is any straight way to get what I want.
> 
> What is the difference between latex2html and tex2ht? Which would you
> recommend? May be there is some better tex->html converter?

with latex2html, you run your document in the normal way.  latex2html
tries to understand the document and .aux file and to emit
corresponding html.

with tex4ht, you add a package to your document and compile it with
latex; you then run an application that reads the .dvi file and
produces the html.

other candidates are tth and hevea which both attempt to understand
the document, like latex2html does.

there are problems with the "decode" approach of latex2html, tth and
hevea, since one cannot in principle understand all of latex without
writing a complete tex engine.  however, the _process_ is much more
readily understandable (imo); of the three, i know (from experience)
that latex2html can be spectacularly slow (on the machines i used to
use back when i did it, for sure), and i've been told that tth is
really fast.

tex4ht of course proceeds on the basis of a tex engine, so it doesn't
suffer from the theoretical problem.  i'm told it's pretty quick to
run, but i've not used it enough to convince myself either way.

(for things i'm going to put on the web, i have a little script that
restricts what i'm allowed to write in my latex, and does a very
simple translation of the commands it believes in.  but i don't write
much other than the faq -- which has its own translation mechanism --
and department documentation, and i can change my script easily if i
decide it needs to do something new.)