RE: [Axiom-developer] sbcl and Axiom

[Page, Bill]

On Wednesday, July 26, 2006 11:15 PM  Gaby wrote:

> ...
> Tim Daly wrote:
> | 
> | Well, as long as we're talking about changing the input
> | language so that it is type-free, one thing to consider
> | would be to mimic the Maple language. That would give us
> | a large test suite of programs and examples. It also might
> | be fairly enlightening about the issues, both for us and
> | for Maple.

I have been and continue to be a long term frequent user and
beta tester of new versions of Maple. I don't see much point
in mimicing the "Maple language" as an Axiom interface - at
least not from the point of view of the significant features
of Axiom. In fact my dissatifaction with Maple as a formal
mathematical environment and with the direction taken by
the MapleSoft developers a few years ago lead directly to
my current interest (dare I say fascination) in Axiom and
Aldor.

> 
> I consider Maple language to be cripled, grown barnacle
> over the year.  With hindsight, we must be very careful.

I agree strongly with Gaby about this. I think we have to be
very careful. I know that Steven Watt and his research group
are quite actively involved with Maple developers and have
implemented several different interfaces of the kind that
Tim is suggesting. See for example:

A Framework for Using Aldor Libraries with Maple.
Cosmin Oancea, Stephen M. Watt.
www.csd.uwo.ca/~watt/pub/reprints/2004-eaca-mapal.pdf

Since Maple remains a properitary system in more or less direct
competition with Axiom in spite of the fact that Axiom is now
open source, I think this connection between Aldor and Maple
is potentially in conflict with Aldor becoming open source
and available for full integration with the Axiom open source
distribution.

Two thing might dissuade me from this negative opinion: One
is if Maple should move much more strongly towards making
Maple open source then in has in the last few years
(admittedly there has been *some* progress). The second is
if MapleSoft would agree to become an official corporate
sponsor of the Axiom project. (Don't scoff!) Seriously I
don't see any reason why MapleSoft and even Mathematica might
not want to do this - afterall what do they have to lose by
doing this? And how many promises of the good will of
mathematical researchers would it like buy for them?

Anyway, although a Maple or Mathematica-style input language
might have some initial appeal to users attempting to migrate
from one of these other environments to Axiom, I think that
in the long run this path for new users to Axiom is not
likely to be a very effective one for Axiom because it too
strongly de-emphasises the fundamental design distinctions
between Axiom and these other systems. In fact it could be
very misleading for new users to be introduced to Axiom this
way.

> I thought that B-natural was considerd to be the typeless
> language of choice for Axiom?
>

I think I would prefeer to describe B-natural as a singly-
typed language rather than a typeless language because of
the way it must interface with the very complex Axiom type
system. Form my point of view, as a candidate for a new user
input language for Axiom, B-natural remains the most "natural"
choice. :)

Regards,
bill Page.