[GAP Forum] SPECHT - now that GAP3 is long in the tooth.

[Joachim Neubueser]

Dear Dr. Aitchison,

On Fri, Aug 05, 2005 at 11:59:26AM +0100, you wrote to the GAP Forum:
> 
> I've been asked to install SPECHT, and hence GAP3.
> Since GAP3 is no longer supported and we have been using GAP4 for years,
> I'm reluctant to make GAP3 available again. I'm particularly worried
> about people getting wrong answers from it, although I guess that in
> many cases the result can be verified independently once found.
> 
> My instinct is to point the user at a way of solving her problem
> in GAP4 rather than enabling her to use GAP3.
> 
> Comments ?
> 
> -- 
> Dr. Andrew C. Aitchison		Computer Officer, DPMMS, Cambridge
> A.C.Aitchison at dpmms.cam.ac.uk	http://www.dpmms.cam.ac.uk/~werdna

I am  afraid that, while  generally GAP4 should  be used, we  can only
recommend  still to install  GAP3 and  its packages,  if e.g.   a user
needs the functionality of SPECHT.  SPECHT makes use of the large GAP3
package CHEVIE  and both of them, as  well as a few  further ones have
unfortunately not yet been tranferred to GAP4 by their authors.

As you know, GAP is a project depending on the voluntary contributions
of many people and this is in particular true for the 'packages' which
remain fully under the  control and responsibility of their individual
authors.   Although  transfer  of   GAP3  code  to  GAP4  is  strongly
encouraged,  the 'GAP  team'  (whatever  that means),  has  no way  of
putting any pressure on these authors.

As to reliability,  from all we know, GAP3 was in  a pretty good shape
when further development  of it was abundoned because  of the shift to
GAP4.   And  this  shift  was  not  done  because  of  doubts  in  the
reliability but in the hope that  the new system GAP4 would offer more
possibilities of extension, a hope that  in our view has come true. So
we hope that installing GAP3 now  should not be a greater risk than at
the time of  the shift to GAP4.   But (as with GAP4, after  all) it is
given away without any warranty.

Hope that this information will help your decision.

With kind regards    Joachim Neubueser

---------------------------------------------------------------------

Prof. em. J. Neubueser
Lehrstuhl D fuer Mathematik
RWTH Aachen
Germany