MathGroup Archive 2004

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

Search the Archive

Re: what kind of a programming language is Mathematica?

  • To: mathgroup at
  • Subject: [mg49270] Re: what kind of a programming language is Mathematica?
  • From: AES/newspost <siegman at>
  • Date: Sat, 10 Jul 2004 02:48:42 -0400 (EDT)
  • References: <cclfk4$kfa$>
  • Sender: owner-wri-mathgroup at

In article <cclfk4$kfa$1 at>,
 gaylord at (richard j. gaylord) wrote:

> One of the reasons that many people have problems with writing
> Mathematica programs is that they fail to understand how Mathematica
> evaluates an expression that has been entered and this is directly
> related to how Mathematica works as a TRS. In order to program in
> Mathematica, it is ESSENTIAL to understand the evaluation process.
------------(stuff snipped)---------------------
> a final note regarding elegant vs. non-elegant programming - note:
> RJF: don't feel  compelled to make a  response to this, especially of
> a personal , nasty nature. as it might set off my ICD  :)    -  the
> issue of elegance in programming is very much a  personal matter (de
> gustibus non disputum). i personally love the style of APL and J which
> is so succinct and compact (the famous one-liner) but barely 
> readable. On the other hand a program consisting of a series of
> rewrite rules employed in a shell for applying the rules' is very
> appealing to me (and my students have indicated the same). it makes
> modifying the program trivial - all one has to do is delete, add, or
> change the rules. the longer runtime the program may take to execute
> allows one to think of other things or go do something else.

Thanks for generosity in donating this stuff.

I guess I was brought up on numerical computing at a stage when if you 
wanted to do some calculation, such as solve a set of linear equations 
or invert a matrix or whatever, you learned how to do this from a book 
(like maybe Hamming's famous text; was matrix inversion called the 
Cayley-Hamilton method?); attempted to understood the process yourself; 
then coded it into a procedural language step by step; then ran that 

[Actually, I go back far enough that there was an initial stage when you 
carefully planned and wrote the detailed step-by-step instructions, took 
them upstairs to where the computation staff was based (one of ours had 
literally been at one time a cavalry officer in the Austrian Imperial 
army), and they tediously and methodically ran through them time after 
time by hand, using Marchant mechanical accumulating calculators with 
huge 100-key keyboards.]

As a result, though I've become an intensive Mathematica user, I have a 
very hard time moving beyond a procedural programming way of thinking 
and coming to grips with rule-based programming or, worse, 
pattern-matching calculations.  Only if I can read a program and follow 
all the DO loops and the rest of the explicit program flow do I feel I 
understand what's happening and have confidence that the program is 
doing what I want it to do.

So, I'll look forward to your notes (and hope, some day, to understand 
object-based programming as well).  Just one minor quibble:  Look back 
at your first paragraph and consider that not all of us may know what 
"TRS" stands for.  At least I don't, though I can make a guess.

  • Prev by Date: customizing Integrate with Unprotect
  • Next by Date: Hypergeometric and MeijerG
  • Previous by thread: Re: what kind of a programming language is Mathematica?
  • Next by thread: Re: what kind of a programming language is Mathematica?