COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Dear Dave:
C00005 ENDMK
C⊗;
Dear Dave:

	We think ARPA is making a serious mistake in cutting so severely
the basic AI research, formal reasoning and mathematical theory of
computation parts of the proposal.  Basic research is the foundation on
which applied research is based.  We hope that our new proposal, which you
haven't seen yet, makes the connection between basic research and defense
applications better than we have in the past.  We tried to present the
basic research in a way that will make it easy for you to expain its
relevance to management.  Please let me know whether the new version is
helpful.  Nevertheless, we despair that ARPA, having already decided to
reduce the effort, will be fully convinced.

	There is, however, hope of getting other support, especially from
NSF for this work, although it will take some time, and it is difficult to
get money for research associates from NSF.  Do you think the following
would be acceptable to ARPA?

	1. We will proceed at the level of effort planned in our original
proposal.

	2. We will charge ARPA for only a fraction of the project's costs.

	3. The ARPA share would be 220K for basic AI and formal reasoning
and 80K for mathematical theory of computation totalling 300K compared to
566K in our original budget and 200K in your revised budget.

	4. If other support is long delayed, we will sharply reduce the
level of effort, but we should have some word from NSF by Fall.

	If ARPA cannot provide the additional 100K (to bring its share to
300K), then we will have to reduce the level of effort in July.

					John McCarthy