FGCS Assessment
Gilles Kahn
INRIA Sophia Antipolis, FRANCE
Several groups have designed languages at ICOT. Were certain principles of lan-
guage design systematically used, did you design or use general tools for this task?
Certainly the technology that you have developed can be useful there.
To conclude this paragraph, I believe that having unity of purpose is extremely
useful; it gives everyone a sense of a global objective, a way of measuring progress
But basic research has its own logic as well. If you pass near an important scientific
problem and do not treat it because it is not squarely in the direction of your project,
this will be unfortunate. Because we are in 1992, I cannot avoid the banale remark
that Columbus did not look for America.
Conclusions
In basic research, ten years is NOT a very long time. As I mentioned earlier,
research on parallel processing has been with us at least since the mid-fifties. Many
many subjects of Computer Science have taken longer than that to mature. In the case
of an extremely original project like FGCS, everyone could -and did- predict that if
you were serious about basic research, then it would take more than ten years to reach
your objectives. It may be necessary to arrange the pursuit of your goals differently,
but it is in the nature of good basic research to create constantly new and unforeseen
problems.
In terms of technical achievements, the FGCS has produced many technical papers,
it has accumulated considerable competence on software for MIMD machines, on how
to harness the power of a large number of processors. It has defined and followed what
I believe are fundamentally good strategic
directions of research: software for parallel processing, theorem proving, Object-
oriented Operating systems, computing problems connected with the law or the
understanding of the human genome.
In terms of social achievements, the success is truely remarkable. The project has
developed basic research in computing on an unprecedented scale in Japan, supporting
a number of activities outside ICOT as well. The Journal that it has fostered New
Generation Computing is a good scientific journal. ICOT has developed friendly and
fruitful international contacts with many countries and scientific institutions. This
cannot be overemphasized. For many of us, the project was absolutely crucial in
opening and maintaining contact with Japan. Finally, the project has maintained
faith in Artificial Intelligence as a fundamental research topic with a scientific basis.
I cannot imagine, at this stage, that ICOT will stop brutally. I have understood
that you plan to release software to the world, In a novel policy of your government.
On the basis of the little experience I have in this area, I would say that few researchers
in the world will want to use software if they know that no one is there to maintain,
develop and improve it, as well as basing its own research on it. Software that has not
- 71 -