00100	                     Computer Science Department
00200	                         Stanford University
00300	                     Stanford, California 94305
00400	
00500	                          January 20, 1971
00600	
00700	Mr. John Parker
00800	Department of Mathematics
00900	Palo Alto High School
01000	Palo Alto, California
01100	
01200	Dear Mr. Parker:
01300	
01400		At  your  request,  I  am  putting  into  writing some of the
01500	comments that I made at  the  January  19  meeting  of  the  COMPUTER
01600	EDUCATIONAL  TECHNOLOGY  COMMITTEE.   As  you  know,  I  am  somewhat
01700	uncertain of the correctness of all these remarks, since I  have  not
01800	been  able  to take a regular part in the activities of the committee
01900	nor to investigate the issues raised.   Nevertheless,  I  made  these
02000	remarks,  because  it  seemed  to  me that some important issues were
02100	being postponed that might well be dealt with sooner.
02200	
02300		1. Large scale computer aided instruction is  not  economical
02400	at present.  Its development will require major long term investments
02500	by government  agencies,  computer  firms,  and  organizations  doing
02600	lesson  programming.   When  and  if  cost-effective  systems  become
02700	available, they will be packages  like  textbooks  and  will  not  be
02800	locally   developed   although  they  may  provide  scope  for  local
02900	adaptation and modification.
03300	
03400		2.  The  plan to acquire the CDC 1700 was probably a mistake,
03500	since  the  machine   is   not   cost-effective   for   Palo   Alto's
03600	data-processing  needs,  and  Palo Alto cannot really afford to carry
03700	out the planned experimental program  in  various  kinds  of  on-line
03800	work.  Whether the plan should be changed at this time in view of the
03900	long term lease and the investment in software is  questionable,  but
04000	it  should  be  investigated.  The committee, which started under the
04100	chairmanship of the author of this plan, has pussy-footed around this
04200	question.
04300	
04400		3.  The  greatest reduction in costs of data-processing would
04500	probably be made if an educational data-processing service  could  be
04600	purchased, for example, from the County or State. A penalty in having
04700	to accept the standard procedures of this service  would  undoubtedly
04800	have   to   be  paid.   An  objective  decision  about  whether  this
04900	possibility  exists  in  a  cost-effective  way   and   whether   the
05000	standardisation  required  is  harmful cannot reasonably be requested
05100	from the EDS group.
05200	
05300		4. It is my impression that the projected hardware  costs  of
05400	adding  more  terminals  to  the  CDC  1700  are excessive in today's
05500	market.  On the other hand, the software costs  of  making  effective
05600	use of such terminals are greatly underestimated.
05700	
05800		5.  It  will  be  quite  unfortunate  if  the  study does not
05900	emphatically point out the need for computer facilities for  teaching
06000	programming in the high schools.  The present hand-to-mouth situation
06100	should not be allowed to continue.
06200	
06300		I hope these considerations are of interest to you,  but  let
06400	me again emphasize their tentative character.
06800	
06900						Sincerely yours,
07000	
07100	
07200	
07300						John McCarthy
07400						Professor of Computer Science