COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\magnification\magstephalf
C00004 ENDMK
C⊗;
\magnification\magstephalf
This is to strongly recommend Dan Pehoushek for admission as a
graduate student in computer science in September 1990.
Pehoushek has worked for the Formal Reasoning Group on the Qlisp
project since the summer of 1987.  His work as a programmer has
been excellent, but this wouldn't justify this recommendation.
The recommendation is based on a major piece of research that is
one of the main results of the project so far.  Everyone had
thought that it would be necessary for Qlisp programmers to
carefully avoid too much parallelism in order to avoid tasks that
require less computing than is required to schedule them.
Pehoushek got the idea for a scheme of dynamic scheduling that
would avoid this automatically.  He implemented it, and convinced
the rest of us that he was right and we were mistaken.  In my
opinion this work was of a magnitude and quality sufficient for a
PhD thesis.  Therefore, he clearly has more than research
potential---demonstrated research ability.

	If we compare him with Stanford graduate students in
computer science who get PhDs, he is in the top half in
demonstrated research ability.
\bye