RE: [Axiom-developer] Ask Slashdot

[Page, Bill]

On Tuesday, December 14, 2004 12:33 PM C Y wrote:
> ...
> b)  Mathematica has a very high reputation and is seldom 
> challenged for two reasons - power and interface.  Most
> people will not push Mathematica into regions where it
> has trouble, and the ability to do 2D input in Mathematica
> is a big, big, big usability boost.  That alone puts it
> forward in people's minds as a polished tool.

I have used both Maple and Mathematica extensively and all
I can say is that Mathematica has a better marketing strategy.
My experience with both is the main reason I am motivated to
help make Axiom available again. I think Axiom got a raw deal
when the commercial territory was staked out a few years ago.

> --- Eugene Surowitz <address@hidden> wrote:
> 
> > Actually the real problem here is the advertising leverage 
> > that query and the uncritical nature of the response give
> > to Mathematica. How could such queries be spotted and more
> > appropriate pointer responses to math sites be generated?
>

Basically I think it is a lost cause. These people aren't
listening - they are talking.
 
> Best case, spot the article and be ready with some concrete
> examples where Mathematica should work and doesn't.

A waste of time - mostly. If you spend any time on the
Mathematica and/or Maple mailing lists you can find a lot of
examples where they don't work but should. That is normal.
People write research papers about this. I am more interested
however to know what the developers do when such problems
are identified. I think that it is here that open source
development could have a major advantage.

> Realistically, make a product that the casual user can't
> distinguish from Mathematica as far as usability goes.
> That's probably what it will take.

Again I think it isn't worth the effort. If people think they
want Mathematica (and have the money), great. The more people
learn about computer algebra systems the better. Later they
might learn what else they really want in such a system.
I think one has to take a long term view.

> ...
> I guess I envision Maxima as starting on the road to being
> a better Mathematica, and Axiom as being on the road to
> being something purer than either, but a tad harder for
> mortals to get started on.

I agree completely. I hope that we can do more to make
Axiom easier for the first time user. But I expect that
Axiom will continue to have the most appeal for "advanced"
users and developers.

> 
> [About exaggerated extolling of the virtues of Mathematica]
> I agree, but they are right in two ways currently - as of
> yet no TeXmacs or Emacs environment for any of the free CASs
> comes even close to Mathematica's front end for smoothness and
> friendliness.  I still think that interface can be equalled or
> even exceeded, but to give credit where it's due it won't be
> a simple task.

I have frequently argued against the amount of time and effort
that Maple's developers have devoted to user interface over the
last 5 years. (I think it is currently better than Mathematica
in some ways - worse in others) And in the most recent shift to
Java as the user interface platform it very nearly broke their
product until hardware advances and the subsequent .5 software
release caught up. Anyone can do user interface. But as far as
I am concerned it is the mathematics inside that counts.

Neither Maple nor Mathematica have anywhere near as complete
a mathematical document capability as TeXmacs and I don't
think they should even try to get to that stage. It would be
much better if they were to continue to move to a more open
development environment and encourage others (like TeXmacs)
and Scientific Workplace etc. to provide the interface.

All that I expect of a user interface for research purposes
is that it not be too distracting from the usual mathematical
notations used. From this point of view both Maple and
Mathematica have been adequate since earlier releases (aside
from some obvious typographical deficiencies which still exist
in both).

> 
> The second way Mathematica wins is third party extras.  The 
> biggest one I know of for Mathematica in academia is probably
> Feyncalc (I suspect engineering has their own I don't know
> about.)

There are many many well developed engineering and physics
applications for both Mathematica and Maple. E.g.

http://www.maplesoft.com/applications/index.aspx

Most of these have been developed by users. Most of these
users could have just as easily developed software for Axiom
or Maxima (some already have).

> ... unfortunately there ARE still a few benefits to a
> massive staff and development budget :-(.

I don't think that either Mathematica or Maple could be
described as having a "massive staff and development budget".
Both depend primarily on user developed applications and
research done by third parties. In the early days both of
them benefited greatly by the pioneering efforts of the
Axiom developers.

If you forget about user interface and look at how the
Maple development environment is changing over the last
few releases, it is obvious that it is evolving towards
the places where Axiom has already been. I don't think that
this is by accident.

Regards,
Bill Page.