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

Re: comparison of Maple and Mathematica

• To: mathgroup at yoda.physics.unc.edu
• Subject: Re: comparison of Maple and Mathematica
• From: twj
• Date: Tue, 24 Aug 93 09:29:24 CDT

Christopher Lee writes.

>Hello, our lab is considering purchasing a math
>package for our SGI's or SPARCS.  We are trying
>to decide between Mathematica and Maple.  Most of
>us have only used Mathematica.  However, Maple
>would interface with this other very expensive
>visualization software package,PV-WAVE, that we
>are considering purchasing.  I'm looking for
>advice on the comparitive advantages and
>disadvantages of the two packages to guide our
>decision from people who have looked at both.




There are fundamental differences between Mathematica and 

Maple as systems.  These are often discussed and addressed
in the relevant mailgroups.   I won't address this other than
to say that I believe that Mathematica is more consistently
designed and extensible than Maple, it also provides a far 

wider range of functionality.


To this particular question of connecting to a visualisation
package I say this:

If one is considering a connection from either Mathematica
and Maple to a dedicated visualisation product it is important 

to remember that the Maple connection is usually engineered by 

making a special Maple version for a particular connection 

for a particular machine.   This means that the version
of Maple is not guaranteed to be the latest version or 

available for all platforms.  Also since development
is being carried out for a minority of users it is not
certain that it will be supported for future versions.


Mathematica takes a different approach: to provide MathLink.

MathLink is the Mathematica communication protocol.   

This protocol is public and used by an increasing 

number of applications.   What is important is that 

it can be extended and developed by anyone who has a 

knowledge of programming.  Thus WRI put their development
effort into developing MathLink rather than lots of 

individual and different links.

Since MathLink provides the same interface, developing
a variety of connections which make use of it is quite 

straightforward since there is no special engineering
required.  Even though anyone could develop these, at WRI we
have worked to break the ground by building a number of 

connections to other applications.  Some of these
applications have been visualisation products and a
number of connections, for example to  SpyGlass Transform 

and AVS, are available.  In progress are connections
to Explorer and LabView.

These connections do not rely on WRI supporting the 

particular link but on WRI supporting MathLink which 

is a fundamental part of Mathematica.  To use these
connections you just require a normal version of 

Mathematica.

In addition many of the links are distributed in
source code form and people are free to enhance
and extend them as they wish.  Therefore there is
maximum flexibility.

Another advantage of MathLink is that it is network
based. Since it runs across a network it is possible to
connect Mathematica to some other application across a 

wide variety of different computer systems.  Thus both
a SPARC version of Mathematica and a Macintosh version of
Mathematica could connect to a visualisation application
running on an SGI. 




Tom Wickham-Jones
WRI