perm filename HEWITT.RE1[LET,JMC] blob
sn#396696 filedate 1978-11-22 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00002 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00002 00002 .require "let.pub" source
C00006 ENDMK
C⊗;
.require "let.pub" source
.skip 5
∂AIL Professor Gerald L. Wilson↓Department of Electrical Engineering
and Computer Science↓M.I.T.↓Cambridge, MA 02139∞
Dear Professor Wilson:
Marvin Minsky tells me that I have a reputation for laconic
letters of reference, and here's another.
I think the basic ideas of Hewitt's PLANNER, especially as
implemented in MICROPLANNER, are a major contribution to computer science.
I think they have not been superseded by later decorations - whether by
Hewitt or by others. Hewitt's weakness, shared by the other improvers of
MICROPLANNER, seems to be that he doesn't know when to stop improving and
start simplifying.
Hewitt's second major contribution is the idea of actors and
its associated idea of regarding computation as message passing. This
idea seems very promising to me, but I have not been able to penetrate
Hewitt's or other people's presentations of it. Because much
interaction of processes has the quality of message passing, it seems
certain to me that actors will play an important role once a good
enough formalism has been devised. Of course, Hewitt is not the only
person trying, but he certainly has a head start and has the most
accomplishments to his credit.
As an idea generator in computer science, Hewitt has a high
standing. There he is on a level with Burstall and Landin.
Elaborating his ideas and making them practical have become
minor industries. As a problem
solver he has a lower standing, but still substantial accomplishments.
I don't remember who his graduate students were except for Irene Greif,
who has an excellent reputation. There is every reason to believe
that he will continue as a first class idea generator and that
others will continue finding it worth the effort required to extract
them from his somewhat vague presentations.
I cannot be more concrete, because I have spent my own time
pursuing different ideas.
.next page
I hope this is still of use.
.sgn