MathGroup Archive 1992

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: What should Mma be?

Jingbin Cao (jcao at writes:

> Improving math algorithms and presentation are not mutually exclusive
> goals. As a matter of fact, they require different skills and talent.

I agree.  But they are paid out of the same total budget, so support  
for one comes at the expense of support for the other.

> ... Before the minimalists can justify their assertion, they should
>  tell the rest of us which presentation program can do the job
> (presentation), and do so with minimum fuss. I am very doubtful they
> can.

After trying more than a dozen systems over the last 20 years, I have  
come to do almost all my memos, letters, papers and reports in LaTeX  
--- it can produce professional-level formatting of math and other  
technical material, it is easy to use (but not easy to learn), the  
files I created in 1985 still work (and can easily be sent to  
colleagues using different systems), editing is far faster than with  
other presentation-systems. While not everyone uses TeX/LaTeX/AmTeX/etc  
(TeX, for short), the number of books produced with TeX has now passed  
1,000. Including "Mathematica: A system For Doing Mathematics by  

"... minimum fuss"? I would not suggest that everyone finds TeX easy to  
*learn*, but then few of the really good tools are:  How long does it  
take someone to learn FORTRAN well?  Or Mma?  Those colleagues whom I  
have introduced to TeX are productive on the first day; but they are  
still learning after a month. And longer.  I find TeX easy to *use* ---  
it produces fine results with "no fuss".

I've used TeX on MS-DOS machines, Atari ST's, Mac's, and others.  But  
my run-away favorite is LaTeX on the NeXT because *all* the Apps work  
together so marvelously!.  Mma graphics (and other graphics objects)  
are handled as PostScript, can be directly imported into drawing Apps  
for further work (addition of "arrows", shadings, labels, etc.), and  
then incorporated into documents. (And sent to colleagues: just about  
everyone can render PostScript.) And everything happens fast enough,  
and simply enough, that I can concentrate on expressing the ideas, and  
not the computer-mechanics.

There are some plotting packages that are better than Mma's for some  
jobs. For example, SciPlot (by  Michael Wesemann for the NeXT) is  
excellent for 2D plots, and RenderMan (included as part of the NeXT's  
software distribution) is industrial-strength for 2D and 3D.  (I will  
stand aside for Mac and MS-DOS users to report their own favorites ---  
there are some superb products for these machines, too.) Hence, many of  
us prefer to use Mma for "prototyping", and would like Mma to output  
its internal data in a form that can be used by these (or similar)  

So, that's how I do it.

Henning Leidecker

  • Prev by Date: need e-mail address for Ross Moore
  • Next by Date: Telescoping series
  • Previous by thread: Re: What should Mma be?
  • Next by thread: What should Mma be?