Richard Fateman ?????:
> Here is a tex- based solution. Set up a big set of
> macro definitions in the TeX language that actually
> understand something about math (precedence, etc.)
>
> Then send to TeX something that is the moral equivalent
> of the internal form of mathematics in Maxima.
>
> (or the MathML form...)
>
> Then let the TeX macros, which would know the line length,
> font size, and the situation (in-line or display), figure
> out whether to do
> (a+b+c)/(d+e+f) or
>
> a+b+c
> ------
> d+e+f
>
> The problem here is that writing TeX macros is rather
> uncomfortable for most people.
> Another is the size of the macros, which would have to
> somehow be input to the TeX processing of expressions.
>
> This approach was actually used by Dan Zwillinger to
> simultaneously typeset formulas for Gradshteyn & Rhyzik
> and make it possible to input the forms to computer
> algebra programs.
>
> RJF
>
> PS, I agree there are more important issues, but
> this one is so tempting to discuss..
>
Well, sound as a good approach but I expect yet
another trouble. It is memory memagement in TeX.
It is pretty static - no allocation of extra RAM
at run time, no GC etc. Is it possible to
feed in TeX any sufficiently _complex_ a nd _large_
Lisp or MathML expression?
--
Vadim V. Zhytnikov
<vvzhy@mail.ru>
<vvzhy@netorn.ru>