Re: Introducing CAS To Engineers (was Mathamatica vs MATLAB)

Richard J. Fateman wrote: 


JohnCreighton_@xxxxxxxxxxx wrote:

I am resurrecting this post from the Mathamatica vs MATLAB thread and I
would like to say that it is my view that the best way to introduce
computer algebra to engineers is to better integrate the symbolic
packages with MATLAB. ...


Cleve Moler, who responded, is the original author of Matlab and the
founder of the company.  All he says is that putting symbolic
processing more centrally into Matlab would require substantial
changes, and that Matlab has a primarily numerical focus. The
choices he made (in spite of design errors you might find), have
produced a valuable and widely-used software package.


I do not see any disagreement, except that Cleve (and many others)
view symbolic manipulation as secondary to numerical computation
as an engineering tool.

There is no doubt that links from a computer algebra system to
a numerical system can be constructed. Macsyma/Maxima had links to
Matlab about 25 years ago.


That's it. LINKS, not *integration*.
Matlab in its current incarnation evolves to a fully-fledged object-
oriented package with dynamical type system. Permitting all the Matlab
objects to be any "symbolics" (deeply embedded hierarchical entities
with rich internal structure) might simply degrade the performance
of the system as a whole. Now, <<cui bono>>?

Do you know a small novel by Jules Verne: "The Master of the World"?
Verne ressuscitated Robur the Conqueror, and made him construct a
super-car, a "bolid which could attain almost 200 mph", which at
will could be changed into a submarine, or a flying craft. Now, the
civilization went another way...

What's wrong with specialized tools?
Forgetting that people's *visions* are very different and sometimes
incompatible leads to abomination. Here and there in France there
were attempts to use Maple as the FIRST programming language in
high schools. Why?: because it is a nice language, not too
complicated, nicely interfaced. You may conclude yourself what
disgraceful mistake it was...

Jerzy Karczmarczuk
.

Relevant Pages

  • Re: Introducing CAS To Engineers (was Mathamatica vs MATLAB)
    ... Richard Fateman disagrees with my suggestion that swelling Matlab ... with symbolics could degrade its performance. ... > to mimic what humans doing mathematics might do. ... where math plays *some* role... ...
    (sci.math.symbolic)
  • Re: Introducing CAS To Engineers (was Mathamatica vs MATLAB)
    ... would like to say that it is my view that the best way to introduce computer algebra to engineers is to better integrate the symbolic packages with MATLAB. ... Demonstrations of useful engineering calculations while help further. ... It is my view that the way the symbolic package was implemented is not the best way to implement the symbolic package without making deep changes to the MATLAB interpreter. ...
    (sci.math.symbolic)
  • Re: Mathematica Vs. Matlab. It is not a superset
    ... It might seem to some people that Mathematica ... Matlab, or "even faster" routines. ... could a CAS be a superset of Matlab? ... Engineers at my school are keen advocates of Matlab and various of its ...
    (sci.math.symbolic)
  • Introducing CAS To Engineers (was Mathamatica vs MATLAB)
    ... I am resurrecting this post from the Mathamatica vs MATLAB thread and I ... It is my view that the way the symbolic package was ... It might seem to some people that Mathematica ... >Engineers at my school are keen advocates of Matlab and various of its ...
    (sci.math.symbolic)
  • Re: Small MCRInstaller.exe is needed
    ... alone binary to provide it to the other engineers. ... actually run a Matlab program, ... Matlab programs no longer compile into C code that calls upon ...
    (comp.soft-sys.matlab)