Moderated FORUM on Computer Algebra Systems

The format of this discussion will be a panel of selected representatives of
various points of view: developers of general purpose and specialized systems,
users from various disciplines (mathematics, science, engineering, education,
business, etc.), reviewers of computer algebra systems or anyone else with an
interesting perspective.  Each panel member will make a statement, providing
some commentary on an outstanding issue of interest in the community.  This
will then be followed by a discussion, starting first with any panel members
who care to comment, after which the discussion will be opened up to the
audience.  This will continue until either a set amount of time has elapsed or
the moderator decides that the discussion has wound down.

Each panel member will have an opportunity to bring up at least one issue.  If
time allows, additional issues can be brought up or discussions cut short
earlier can be continued.  The commentary may last a few minutes and the panel
member may even use ONE viewgraph (but no more!) to help present his or her
point.

The focus on this session will be computer algebra systems as this is the main
interaction between the user community with developers and researchers.  Topics
might include (some based on discussion in sci.math.symbolic):

o what kinds of abilities (and facilities) are lacking in the current systems
  that users feel are needed?
o what kinds of abilities (and facilities) are lacking in the current systems
  that developers think are needed?
o What good ideas in some of the current systems should spread widely?
o how easy is it to use the current systems?
o how easy is it to find out if a system can do a particular operation?
o does one have to be too much of a computer scientist to effectively use the
  current systems?
o how easy is it to manipulate and transform expressions?
o do the current systems use too much brute force and not enough finesse?
o is the algorithm a system employs too dependent on the exact form of the
  input [e.g., integrate(x^2/(1 + x^3)^(3/2), x = 0..2) may produce a very
  different looking result than integrate(x^2/sqrt((1 + x^3)^3), x = 0..2)]?
o are the algorithms employed covering well areas of applied interest? 
o how well do the current systems interact with other programs?
o how well do the current systems manage data from external sources?
o what quality of output and graphics is produced and how could it be improved?
o how well do the current systems generate code in numerical languages like C
  and Fortran?
o how well do the current systems cope with large problems?
o how well and how should systems cope with the different viewpoints presented
  by various flavors of mathematicans, scientists, engineers, educators,
  business users, etc.?
o what is the role of reviews?
o commercialization of CASs
o whatever else seems appropriate

Although the topics may be broad, the panel member should strive to inject some
specifics in his or her presentation to help focus the discussion.  The
ultimate goal is to engender some good interactions within the developer/user
community.