system. There is great potential for your technology to support distributed AI applications 
and CSCW (Computer Supported Cooperative Work). 

Also look at applications of your technology for numerical applications. At a 
workshop in Syracuse in 1990, Geoffrey Fox complained that existing languages for 
parallel numerical applications only support homogeneous parallelism. Perhaps this is 
another application area for a suitably extended KL1, heterogenous numerical applications. 

In some form or other the FGCS project must continue, or the achievements 
will have far less impact than they should. 

The new project, run perhaps by a smaller ICOT, should support, maintain 
and continue to develop KL1, PIMOS and the application support tools. Its 
role should be to help others to use this software, by producing excellent documentation 
and assisting outside groups (in Japan and elsewhere) to develop applications. There is less 
need, now, for ICOT to develop complete applications, except perhaps the `killer' 
application. In addition, ICOT should be adequately funded to continue 
fundamental research into LP and its use on parallel machines. 

Finally, thank you for an exciting 10 years of excellent research into concurrent LP and its 
use. May your good work continue. 


					- 52 -