[Axiom-mail] RE: [Axiom-developer] conference

[Page, Bill]

On Tuesday, April 26, 2005 10:37 AM Martin Rubey wrote:

>Camm Maguire writes:
> > Hi Tim!  Great conference!
>
> So, what happened?

I attended both the pre-conference (sprint) day and the main
conference. The two days seemed to go by very quickly which
for me usually means that it was interesting, worthwhile, but
that there was too much to do in too little time. While it is
still fresh in my mind I will try to list below some of the
things that stood out for me and I hope other participants
will add their comments as well. 

I will let Tim report on the actual statistics, but I think
the conference was quite well attended, considering that it
was the first ever meeting about the new open source version
of Axiom. We had about 7 or 8 people the first day and somewhere
between 30-50 people the second day.

The City College campus was a great venue. The apple and cherry
blossoms highlighted the wonderfully maintained historical
buildings which seemed to me to blend well with those of more
modern architecture and the distant backdrop of New York city
skyscrapers ... (there that is the extent of my capacity for
poetic imagery :). Plus I would like to say thank you to
Gilbert Baumslag and Bernice Ravitz for their generous
hospitality.

The discussion on the sprint day was lead by Tim Daly and
revolved around future directions and the larger, longer term
context in which open source Axiom will develop. For those
people who have been following the discussion here on the
Axiom-developer list for the last two years, much of what we
discussed would have seemed familiar. We talked about many
things, but here as some that come immediately to mind:

- literate programming and documentation, pamphlet format
  files (noweb) and implementation on the web

- mathematical versus program correctness and relationship
  to proof systems like ACL2,

- user interfaces, both desktop and web-based; and their
  integration (the doyen project)

I am sure we can add other things to this list. It did seem to
me however that (as usual and normal in an open source project)
we did not really make any firm decisions. We have many more
ideas for projects and research than there are available
resources.

During the latter part of the day we spent some time on more
technical issues such as the role of GCL in support of platform
independent graphics (for both Linux/Unix and Windows). Tim
demonstrated a possible Java-based solution but I think the
idea of adding a Java dependency to Axiom motivated all of us
to think of ways that this could be achieved more directly with
GCL and tcl/tk.

I also helped set up a Windows development environment for
Axiom on one of the workstations in the open source development
lab. Tim and I worked (without a complete resolution) on a
problem of building the new Axiom book volumes from the tla
archives in the Windows environment. There seems to be an
issue with noweb even though it works properly for building
Axiom from source on Windows.

The main conference day was to be devoted to applications of
Axiom. Perhaps there was not enough discussion of actually
applications, however William Sit did give some examples of the
using Axiom "for both fun and research" and during part of my
talk (which was mainly about the MathAction website and how it
supports both Axiom development and Axiom users) I did point
out the work on "guessing integer sequences" that Martin Rubey
has posted on www.axiom-developer.org

Tim Daly opened the conference with an overview of Axiom and
the concept of the "30 Horizon": 30 years back from a historical
perspective and 30 years forward. He gave reasons why we might
believe that Axiom will still be actively used and developed
30 years from now and what we might be able to do now to help
ensure this.

Richard Fateman discussed the "state-of-the-art" in what one
might call the "artificial intelligence view" of computer algebra
systems, he pointed out that although current systems like Axiom,
Maxima and commercial systems like Mathematica and Maple are
useful for symbolic computation, they are still a long way from
being "artificial mathematicians". I think also that with his
analogy of a "mathematical golem", Richard might have also been
trying to suggest that the goal of creating artificial
mathematicians might not be such a good idea. :)

Camm Maquire gave a talk on the history and potential of open
source software. To me this was a bit of a surprise given Camm's
specific and detailed technical contributions to both GCL and
Axiom, but at a conference focused on **applications** I think
these are issues that need to be addressed. How can we convince
academic and commercial budget managers that open source can
compete effectively with commercial products when the economics
and motivations are so different?

In all of the talks there seemed to be a common thread of the
need for collaboration - between Axiom developers; among Axiom
users and between mathematicians and the Axiom system. I think
there was a general agreement that in the open source Axiom
project so far we have achieved some significant steps in the
right direction but that there remains a lot more to do.

Finally, I want to add a special personal thanks to Richard
Fateman for solving a last minute technical problem with
network access to the axiom-developer.org website by the
loan of his laptop computer and wireless Internet connection.
Without his help I don't think I could have given a particularly
convincing verbal demonstration of the potential for the
MathAction website to support the kinds of collaboration that
Axiom needs.

Bill Page.