COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Subject: Report on computing resources
C00007 ENDMK
C⊗;
Subject: Report on computing resources
From: John McCarthy

	I am sorry to have to say this so late in the game, but I cannot put
my name on anything like the present draft report, because I disagree with the
following points:

	1. The conclusion that computing in the future will be based
on personal computers.  Maybe it will, but at present I see only some
group wishful thinking.

	2. When we met in March, I thought we agreed that having many terminals
was more important than having super-terminals.  There is no trace of this
conclusion in the present report.

	3. Not being a TENEX fan myself, I see the complaint about the KI-10
as being a translated complaint about TENEX.  Second I am skeptical of the
statement that D.E.C. has decided to make TENEX their main vehicle, and
I can't see how it could be in their interest to do so, since TENEX is
not better than their own system in many respects, and ARPA makes up only
a fraction of their market.  It seems to me that there is and has been much
wishful thinking about D.E.C. becoming wedded to TENEX.

	4. The draft report doesn't discuss the concrete situation of even
one ARPA site.  Therefore, it is not clear how its generalities would
interact with what the sites plan to do or ARPA's plans for support of
research.

	5. Besides all this, I think that a number of issues have not been
dealt with that are important to ARPA computation over the next years -
in what areas is compatibility essential - displays, printing above the
teletype level, how does the ARPA community fit in with the larger
time-sharing community, and many others which I have not even had time
to formulate.  Therefore, I would suggest meeting some more and going back
to the drawing board on the report.


ADDENDUM From Les Earnest

I too find the report too far from my beliefs to sign.

1.  The section on "personal computing" is still off the mark.  It
recommends the development of a cheap PDP-10 that runs Tenex, which
makes about as much sense as developing an electronic calculator that
displays its output by positioning abacus beads. 

2.  The assertion that these computers ought to have local file
storage is apparently not supported by any studies of
cost/performance tradeoffs and is probably wrong at least for the
forseeable future. 

3.  The recommended network file/archive system won't work with
current ARPAnet bandwidths, yet there is no discussion in the report
of the need for substantially greater capacity.

4.  A prerequisite for efficient exchange of files among network users
is a file format description language.  So far, we don't even have a
good example of one, let alone a standard, yet the report says
nothing about this need. 

Overall, the report presupposes that all prior developments in the
ARPA community are on exactly the right track and does not examine
shortcomings or alternatives in sufficient depth.