On Dim, nov 13, 2005 at 09:12:56 -0800, Richard Fateman wrote:
The sourceforge project has some features that were never added to the commercial Macsyma, and even some bug fixes. Though the loss of the Macsyma Inc / Symbolics legacy is unfortunate. Downloading it and setting it up is trivial. (Click on the install icon).
I never used Macsyma, thus I can't make any comparison, but I used much MuPAD 2.5.3 and GNU Maxima; though the library of MuPAD probably is bigger, I preferred Maxima as soon as I discovered it.
[...] There are other systems like Jacal, MuPAD (formerly free, but maybe no longer), Pari, ...
To my knowledge, something indeed has changed int Mupad's license between releases 2.5.3 and 3, but I never knew what. I first thought it was not free any longer, but I installed the new version on the computer of my wife and got some kind of strange license that didn't cose anything to me, though obviously it was something different and more complicated to get than previously (until 2.5.3 when running MuPAD, my name would be written on the screen at the beginning; in some 3.1.? version there was rather an ugly number instead of my name), but in fact I didn't have the time to use it since... I discovered GNU Maxima which suits my needs perfectly (besides I had to use Linux emulation in order to make MuPAD run and I can use GNU Maxima natively on a FreeBSD system). I also use Pari/GP for more "computational" things: - Maxima is the best free CAS I know; - Pari/GP seems to be the best (free or not free) for numerical manipulations; note it is not really a CAS and can't be compared to Maple, Mathematica, Macsyma, Maxima, MuPAD, etc. though it has some efficient support for rational functions; but it is a much better tool for eating numbers rather than RAM... when I have to make a computation last several hours, I wouldn't even think to use something like Maxima (though I did it once): I would either code in into C with GMP or more often use Pari/GP. Note that it has many line features: - support for readline (of course I have something equivalent with all my other tools since I use a line editor, 'ledit', but it is a very important feature anyway); - probably one of the most efficient support for finding linear dependancies or minimal polynomial for a numeric number (even for finding linear dependancies between vectors); - a full integration to C programming whenever speed is an issue; - very efficient numerical routines for general purposes factoring (well I don't mean breaking a RSA code, but when you need to factor some 20 or 30 digits number for studying what it is made with, Pari is much quicker than Maxima - I don't remember with MuPAD, but I think MuPAD was very slow); - In fact I think Pari is intended to handle with seevral thousands digits numbers which is a very good choice (the manual explains that it is not intended to handle with several millions digits, but indeed most computations have to be optimized for hundreds or thousands of digits rather than millions (where using a library and coding at a low level is an issue), this makes Pari/GP very efficient; - ... I could add many things here, since for precise purposes, I really think Pari/GP is much more easy and efficient, but precisely now I think to something I used yesterday: most systems have a "system()" function (and GNU Maxima has it); what made my life very happy yesterday was to discover the "external()" function which takes the output as a literal value (something like `...` in a shell): a = external(concat("myscript ",Str(value))); where myscript is whatever you want (in my case it was a C program writing anything more or less valid according to Pari; in my case the C program makes a request to a database on a network and merely writes a vector in the standard output: [ 1214265167, 1212176271 ,12126188, 90398938, ...] Of course I could have done with any other tool with some heavy input/output feature, but this is the real Unix way to do it with this kind of "shell escape" and again I was happy to do it like that. Regards, -- Thomas Baruchel Home Page: http://baruchel.free.fr/~thomas/