COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	IBM GRANT MONEY
C00008 ENDMK
C⊗;
IBM GRANT MONEY
Jeff,

As we discussed briefly before, I have two suggestions for spending
some of the IBM grant money.

1.  A distinguished lecturer series.
The format  would be  to  invite a  leading  computer scientist  to  visit
Stanford for a couple days, during which time s/he would deliver a general
colloquium to the department, meet with relevant faculty and students, and
give a second, more detailed talk for them.  I suggest each CS area choose
one speaker, thus making a  total of about 5 or  6 for the year.  AI,  for
instance, might invite someone of the stature of Newell or Simon.  Most of
the speakers  will have  their own  source  of travel  funds, so  I  would
recommend  awarding  them  each  a  $1000  honorarium,  and  local  living
expenses, but have them  pay for travel themselves.   Thus the total  cost
would be $5500-6500.

2.  An "immigration course" for new PhD students.
Incoming students have  a great variety  of backgrounds, and  it might  be
worth taking a couple  weeks to try to  equalize their skills somewhat  in
programming, mathematics, problem solving, etc.  In addition, there are  a
plethora of machines, OSs,  editors, document prep  systems, etc. in  this
environment, and all too often a  student learns a bare subset and  clings
to that even though there might be some more appropriate tool if only s/he
were aware of it.  Finally,  there is a need  for students and faculty  to
"pair up" early  in the student's  first quarter here;  by having a  short
introduction where each faculty  member and each student  got to meet each
other,  this  pairing  might  be  more  suitable  to  both  parties.   The
"immigration course" would occupy a  full-time week before classes  start,
and some  of the  first week  of classes.   In it,  graduate students  and
faculty would give talks related to specific goals:  familiarity with  the
systems in this environment; background knowledge which will be assumed by
the introductory AI course, the introductory NA course, etc.  Each faculty
member will  also  deliver a  brief  talk  on his  current  research,  and
interests and  opportunities for  graduate students.   At the  end of  the
second week, students will be asked to fill in a sheet listing the faculty
members who are  their first, second,  third, and fourth  choices to  work
with.  The  faculty  will  meet  and discuss  these  sheets,  and  usually
all the students -- and faculty -- will get their first or second choices.

The IBM money would be used over this coming summer, to prepare  materials
for such a course.  It is VERY intensive, and requires a tremendous amount
of coordination  and  preparation.  One  full-time  student, plus  a  fair
amount of faculty time, are needed.  The latter can be handled by whomever
is going to handle it  in the fall (I volunteer  to do this for the  first
year, at least).   Thus IBM would  primarily be paying  for one  full-time
summer student (or two half-timers)  to assemble problem sets, codify  the
material which is  required for  the introductory courses  (e.g., talk  to
every area  except  Hardware to  find  out exactly  what  knowledge  about
hardware they assume when they teach their introductory courses; then talk
to hardware  people  to  find  out  how  to  present  just  THAT  material
succintly.)  I would also use the IBM money to have that same student  (or
two) act  as  TA's during  the  Immigration Course,  and  to pay  for  the
computer time, copying charges, and other  material costs used as part  of
that course.  The total cost would then run about $7000.


Doug