00100		As the Daily reported, I was most  vocal  in  dissatisfaction
00200	with  the SWOPSI proposals for regulating research at Stanford.  Here
00300	are some reasons:
00400	
00500		1. It was proposed that a professor's  chances  of  promotion
00600	not  be  jeopardized  by  does no research provided he teaches a full
00700	load.  This might be OK in some departments, but it wouldn't  do  for
00800	the  Computer  Science  Department,  a  graduate  department  where a
00900	professor not doing research could not do his  share  of  supervising
01000	student research.
01100	
01200		2.  The elimination of salary off-sets whereby a professor is
01300	paid by teaching and research sources in proportion to  his  activity
01400	in  these  areas  would  eliminate  professors  like me who prefer to
01500	arrange our ratio of teaching  and  research  to  match  our  current
01600	interests.  It would also eliminate contact between students and many
01700	research leaders who would then prefer to  work  in  purely  research
01800	institutions.   There are plenty of repectable colleges in the United
01900	States where the faculty teaches full time, but I presume  that  most
02000	students  who  come  to Stanford know that Stanford is an institution
02100	that does both.
02200	
02300		3. The proposal to limit the amount of research support for a
02400	department from a single source is unworkable.  For example, there is
02500	only one major support for research in artificial intelligence in the
02600	U.S.,  namely  the  Information  Processing  Techniques Office of the
02700	Advanced  Research  Projects  Agency  of  DoD.   Unfortunately,  this
02800	research  requires  a  dedicated large computer (ours cost about $1.5
02900	million)  and  therefore  a  large  project.   On  the  other   hand,
03000	theoretical research in computer science is much less expensive.
03100	
03200		4.  The  proposed  tax  on  externally sponsored contracts to
03300	support graduate  student  research  that  does  not  fit  under  the
03400	contracts  is  ok if the external agencies can be persuaded to go for
03500	it.  Actually, we have considerable freedom in allowing people to use
03600	our  facility,  and I would there is a substantial budget of computer
03700	time at the Computation Center for unsponsored research.   Therefore,
03800	I  would  doubt  that  many  Stanford students lack suitable computer
03900	facilities for their desired research and would be interested to hear
04000	about  counter-examples.   On the other hand, while we have been able
04100	to let music students use our computer, we  have  not  been  able  to
04200	provide salaries for students to do research in computer music.
04300	
04400		This  leads  into  some  of  the broader issues raised at the
04500	meeting of whether
04600	support of projects by mission oriented agencies is compatible with
04700	having a good university.  The SWOPSI report seems to doubt it.  My
04800	own opinion is that universities should do research that will help
04900	solve society's problems, but that there should also be substantial
05000	funds available for projects that extend knowledge for which no
05100	application is foreseen.  This is approximately the case at present.
05200	
05300		The SWOPSI report includes the Computer Science Department
05400	in a list of departmens whose research is distorted by outside support.
05500	No reason for this conclusion was given.
05600	
05700		At the meeting, Ms. Farlow said that she doubted that the
05800	Artificial Intelligence Project was appropriate for a university.
05900	When she was asked why, she mentioned only the lack of involvement
06000	of undergraduates in the project.  Various speakers then pointed out
06100	that the university has many activities such as the professional
06200	schools than don't involve undergraduates much.  (I should mention
06300	that undergraduates qualified and wishing to work in artificial
06400	intelligence are welcome if they can find us.)  In fact, Stanford
06500	University has three goals: undergraduate education, graduate education,
06600	and research, and some of its activities contribute to only one
06700	or two of these goals.  The fact that most activities contribute
06800	to more than one is the justification for multi-purpose institutions.
06900	
07000				John McCarthy
07100				Professor of computer science