COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Professor Joel Moses:
C00005 ENDMK
C⊗;
Professor Joel Moses:
Department of Computer Science  and Electrical Engine
M.I.T.
Cambridge, MA 01239

Dear Professor Moses:

	Sorry for the delay.  I thought I'd sent the letter, but
the system must have crashed or something, because I found the file,
and it had nothing in it.

	Anyway I have a good opinion of Halstead and am glad to
recommend him for tenure.  I know him, because he gave a paper
on his Multilisp at the same Lisp conference at which Dick Gabriel
and I gave a paper on our Qlisp (then called Qlambda), and the
approaches were similar.  I agree with the approach to parallel
processing in Lisp he is taking.  It can be characterized as
conservative from the point of view of the Lisp system programmer.
It puts considerable demands on the hardware designer in order
to make sure the machine is programmable.

	Halstead's project is more ambitious than ours, since
he is designing and building his parallel processor, whereas
we are using a commercially available machine.  There is only
one other project using this approach --- Takayasu Ito's
Pailisp project at Tohoku University in Sendai, Japan.
I would say that Halstead's project is ahead of the other two
in implementation.

	Naturally I don't know about his teaching, but his lectures
are clear and so are his papers.

	I have no hesitation in recommending him for tenure.

Sincerely,

John McCarthy