From: Anders Sandberg (asa@nada.kth.se)
Date: Mon Aug 27 2001 - 02:45:25 MDT
On Mon, Aug 27, 2001 at 08:22:51AM +0800, Chen Yixiong, Eric wrote:
> > > and the fact that Mathematica provides most or all of Wolfram's income, I have
> > > to ask: is there an open source version of Mathematica? The concept is
> >
> > No, but there are other numerical algebra packages. None of them are as
> > advanced and rounded as Mathematica or Maple.
>
> I think, why don't we write a better one than theirs, one radically
> different but much better?
Sure, but the problem is nontrivial. Symbolic algebra and especially
analysis is rather hairy when you get to the details. Doing a bit of
derivation, substitution and pattern matching is a typical computer
science course exercise, but when you want to get performance, the
ability to handle integration, differential equations and start linking
with with numerical analysis the detail-devil jumps out with his
pitcfork bifurkation. Just take a look at the papers at
http://dft.if.uerj.br/odetools.html and
http://www.SymbolicNet.org/areas/GB_and_CS/index.html, or the quick
explanation at http://www.SymbolicNet.org/areas/factorization/index.html
BTW, there seems to be some downloadable software at
http://www.SymbolicNet.org/ftpsoftware/ftpsoftware.html
and a version of venerable MACSYMA is apparently going gnu:
http://www.ma.utexas.edu/maxima.html
> Consider this: They do use programmatic mathematical notations to
>model the universe, however, not in the way I had originally thought.
>My version has more of a programmatic than a maths structure.
This is the true engine of progress in computer science (or any other
field): hey, he did something in a way I think I can do better! Lets
make a better program!
> In fact, if you actually look at what I now have in mind, it would
>seem like some Visual Basic and C++ hybrid, very very much like
>programming code. You can even make a compiler and run that, I think.
>This consists of a subtle but important difference. Maybe he had
>thought of that too, but from what I managed to read up to now, no.
As a computer scientist, I would suggest looking beyond what computer
language to use and instead try to formulate the *problem* you are
trying to solve more clearly. Once the problem is defined, you can
select a language or (more likely) write your own.
-- ----------------------------------------------------------------------- Anders Sandberg Towards Ascension! asa@nada.kth.se http://www.nada.kth.se/~asa/ GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:10:08 MST