MathGroup Archive 1994

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

Search the Archive

Re: your mail

  • To: mathgroup at
  • Subject: [mg243] Re: your mail
  • From: mcdonald at (William MacDonald)
  • Date: Mon, 28 Nov 94 13:10:03 EST

> Dear mathgroup:
> Did some of you at any point investigate how to emulate
> the APL language in Mma?  I located a table of corresponding
> functions in the first version of the Wolfram book, but I am
> happy to receive more pointers on this.  
> The reason for asking this is that I picked up a book that
> employs APL to illustrate some algorithms and datastructures,
> and, sure enough, I never did any APL.  Translation into Mma
> could be a fine exercise.
> Regards, Stig.
> ==================================================================
> Stig F. Mjolsnes, research scientist
> SINTEF DELAB, The Norwegian Inst. of Technology, N-7034 Trondheim
> Internet: Mjolsnes at
> Phone: +47 73 592024
> Fax:   +47 73 592636
> ==================================================================
Mathematica supports many of the APL concepts, and I think that one
could rewrite APL algorithms in Mathematica (if one could understand
them). But APL has some idiosyncracies not built into Mathematica,
like right to left processing etc.  The second volume of Wolfram's
book really omits reference to APL, and omits the equivalencies listed
in the first volume.  My experience is that people who have known and
loved APL quickly become expert in Mathematica. But those APL aspects
of Mathematica are most difficult for others to appreciate.  I see
many programs written by people using C- and FORTRAN like loops,
rather than using element-by-element processing that is built into all
Mathematica functions (Listable).  It seems that Wolfram is now
downplaying the APL aspects of Mathematica. Too bad, because they are
very useful.


William M. MacDonald
Professor of Physics
University of Maryland
Internet: mcdonald at

  • Prev by Date: answers to regress
  • Next by Date: Question
  • Previous by thread: answers to regress
  • Next by thread: Question