FGCS Project Evaluation Report
Sten-Ake Tärnlund
by STU and the creation of SICS. In addition, private companies have made significant
efforts e.g., IBM.
3.5 Economics
As pointed out above, Japanese manufacturers have got the opportunity to go
from the PIM prototypes to competitive massive parallel computers with more market
hardware and software. Although such a move involves more than technical aspects,
this economical opportunity is a pleasant spin-off from the FGCS project.
3.6 Public relations
We have seen and will see many press reports on the FGCS project. Some may
be somewhat superficial as the article in the International Herald Tribune, the other
day. Some will be more sophisticated as the two consecutive editorials in the New
York Times, some time ago, interpreting the FGCS project not only as a success but
recommending that the US should learn from this project and launch a new industrial
policy for its computer industry. This recommendation is, of course, equally valid for
Europe. In general, the results of the FGCS project are not yet fully understood and
appreciated, and we would benefit from a better understanding of the achievements of
the FGCS project.
4 A FEW RECOMMENDATIONS
Japan has set the stage for future research based on the success of the FGCS project.
I shall only briefly mention some potentials for an international cooperation based on
the Japanese results and the international results, in particular, from the many re-
sponses to the FGCS project around the world.
There are several interesting subject matters e.g., from parallel inference machines
to advanced AI via programming methodology. Logic programming would be helpful
to strengthen the coherence of such an effort. This research paradigm was considered
the most powerful for the FGCD project. I do not think it has been replaced, new
projects would energize the original vision and be a challenge for the new millenium.
Some principal areas would be:
- parallel inference machines;
- software methodology e.g., specifications, transformations and
verifications;
- software development methodology;
- artificial intelligence methodology;
- advanced applications.
- 132 -