Dear GAP users,

today is my last day as employee at Lehrstuhl D f"ur Mathematik. Allow
me a few words first about GAP's and my common past and then about our
separate futures.

I have been working on GAP since 1985, literally since day 1. I am to a
fair extent responsible for GAP's current state, and most of my work in
those 11 years has been for GAP. I enjoyed this time a lot.

I enjoyed meeting and working with many great persons. Please forgive me
that I do not list them here, to keep this message acceptably short.

I also enjoyed developing, distributing, and maintaing a program system.
I enjoyed discussing with the other developers about where GAP should be
headed and how it should get there. I enjoyed designing the system and
implementing the designs. I enjoyed learning mathematical algorithms and
implementing them. I enjoyed turning the program into a distributable
system and getting it out of the door. I enjoyed your feedback (but I
did not enjoy fixing the bugs ;-). I enjoyed the discussions, the
private ones and the ones in the GAP forum.

All those years I could do this work and earn a living. For this I have
to thank foremost Prof. Joachim Neub"user, who offered me a position at
Lehrstuhl D and convinced the Deutsche Forschungsgemeinschaft to support
me.

Unfortunately continuing this work is not compatible with a usual
academic career. Put differently, the older I get the more difficult
it becomes to find money for me.

I will now take the things I learned in those years and use them to
start my own software business.

I wish it was possible for me to continue to develop mathematical
software. But there are reasons why this would not be a good idea.

First. I believe that mathematical software is best developed by
mathematicians for mathematicians. Such software should be treated
like other mathematical works, i.e., it should be distributed freely
and mathematicians should not have to pay to use it.

Second. It is difficult to earn money with mathematical software such as
GAP, because the number of potential users is relatively small and they
cannot spend large amounts of money on software.

Third. Those areas where there is a sufficiently large market are
already claimed by large companies such as Macsyma Inc., Waterloo Maple
Software, Wolfram Research, and NAG Inc. and I do not believe that I can
compete with them.

Instead I will develop networking software. This is a market whose users
have commercial interests, so they should and can pay for their software.
It is a large market, much much larger than the market for mathematical
software. It is a rapidly changing market, so there are enough niches
that are not already claimed by large software companies.

It is difficult for me to leave now, because I believe that GAP is in
a very exciting phase right now and that it has a very bright future
(I wish I was so certain about my own ;-).

All those years I was on a quest for a type system that would allow to
model algebra well. Time and again I have argued that object oriented
languages such as C++ are not powerful enough to model the complex
relationships in algebra. Scratchpad/Axiom is much better, but is has
its own share of problems. I believe the new type system in GAP 4 comes
closer than anything else (but of course I am biased ;-).

The move of GAP to St. Andrews means (among other things) that GAP will
become a system that will be developed and supported to a larger extent
by its users. This model of development has worked very well for other
systems (foremost for Linux), and I think it will work well for GAP.

Since I am still interested in GAP's future, I will not leave completely.
I will continue to read the GAP forum and (if my work leaves me a few
free minutes) participate. If the developers have problems with code
that I have written, I will certainly try to help them.

But I cannot guarantee how often I will come around to read my e-mail.
So, in your own interest, you should report all GAP related problems to
'GAP-Trouble@Math.RWTH-Aachen.DE' and not to me.

I hope you enjoyed the time with me as much as I enjoyed it,

Martin.

-- .- .-. - .. -.  .-.. --- ...- . ...  .- -. -. .. -.- .-
Martin Sch"onert,   Martin.Schoenert@Math.RWTH-Aachen.DE,   +49 241 804551
Lehrstuhl D f"ur Mathematik, Templergraben 64, RWTH, 52056 Aachen, Germany