Martin Schoenert has said good-bye to us, I guess it is my turn to
return a most grateful good-bye to him, but also to say a few words
about the present situation of GAP, as Martin is leaving it.
Martin has very adequately described his engagement with and for GAP,
and indeed for several years now I have often called him the
'architect' of the system. As I have emphazised in the preface to the
manual, his role in the general design of GAP cannot be overestimated,
but also, and much less noticeable to the majority of the users, his
constant struggle for a system that is truely open because in all
details it has a clean and clear structure that allows the user to
extend and modify it for particular problems.
Such perfectionism has sometimes overthrown too optimistic estimates
for the date of the next release, update etc. and indeed the first
version of GAP 3 did come out much later than was hoped and similarly
now GAP 4 is long overdue measured by some dates that I was bold
enough to mention. However these times were used for long
considerations of possibilities that in the end paid off by their
result. In particular during the last couple of years intensive
discussions of the desirable features for GAP 4 took place between
Martin, Thomas Breuer and Frank Celler. What had to be taken into
account were the problems of
- the possibility to introduce an increasing number of mathematical
- the possibility to handle a fast increasing number of algorithmic
methods for the same or related goals,
- and yet to keep as much upward compatibility as possible ( a point
much more important than in the transition from GAP 2 to GAP 3 in view
of the much bigger number of users).
During the GAP workshop from September 1 - 7 we got valuable, and
generally very positive feedback from about 20 participants from
outside Aachen about the new concepts of GAP 4, but of course with
Martin leaving the official GAP team ( I am glad that indeed we can
count on his continued advice, as he said in his letter) the question
about the state of GAP 4 is most legitimate.
Kernel and compiler of GAP 4 are operational, the new structure will
allow additions to the kernel much more easily. About half of the GAP
library has been adapted to GAP 4 and work by about six people at
Aachen and two in St Andrews on the adaption of the rest is making
good progress. Quite a few new routines are basically ready to be
joined to GAP 4 soon after a first release which will have about the
same mathematical functionality as the present version. We hope that
the following timetable is realistic:
- Until early December '96 work of the above mentioned team with a
continuously growing library under CVS will go on;
- Then there will be weekly developmental versions and the team of
developers will be increased, involving e.g. several participants of
- About early March '97 a beta test version will be available to every
- About May/June '97 GAP 4.1 will be released and with it St Andrews
will take over control of the further development of GAP 4 which will
then lead to further versions adding more new features to the library.
- Moreover it is planned that a last update GAP 3.4.4 of GAP 3 will be
issued (already by St Andrews) in January '97.
Surely there are many sizeable tasks ahead, adapting the rest of the
library and the share packages, and rewriting the Manual for GAP 4,
but I think that Martin is leaving GAP on a safe way to version 4 and
moreover he leaves a number of cooperators who have worked for a long
time closely with him and will continue what he did.
I am not sure, if I can speak for the users, but for the team in
Aachen, I want to express my most sincere and warm thanks to him and
wish him all the best for his future plans.