COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub" source file
C00006 ENDMK
C⊗;
.require "let.pub" source file;
∂AILDr. Herbert Stoyan↓DDR 806 Dresden↓Togliattistr. 40↓East Germany∞

Dear Dr. Stoyan:

Steve Russell is in New Zealand for two years and Dan Edwards probably won't
answer your letter, but I may telephone him for you.

I'm sorry but I don't have time for a joint paper.

Here are answers to your questions.

1. I didn't do computer work before 1956 but was interested in
noncomputer applications to artificial intelligence.

2. Seminars about Fortran were given at MIT starting, probably, in the
fall of 1956.

3. There were no proceedings of the Dartmouth conference published.

4. Allen Newell reported on IPL at the Dartmouth conference.

5. Nathanial Rochester of IBM, who was present at Dartmouth, decided
to make a geometric          based on the idea that Minsky presented
at the Dartmouth conference.  Gelernter, who was also hired by IBM, was
asked to work on the project.  I visited IBM about once a month during
the planning stages of the project and wrote some memos with the
idea of making an algorithm list structure language.

6. 

7. I don't understand the question.  CONS should allow composition.  I
didn't understand that until Gelernter found that out.

8.  I continued consulting on FLPL until the summer of 1958 when my
ideas for recursion were developed and could not be realized within
FORTRAN, making a new start necessary.

9. Among other things, I worked on a             program at MIT.

10. The IF function was call-by-value.  This limited it's usefullness.
I immediately wanted to make it call-by-name although those terms were not
in use.

11. The MIT-AI project started in the fall of 1958.  I started at MIT in
1957.

12.  I don't remember precisely who was in the first batch of students.
Maling was a professional programmer.  Black was a Harvard student.  The
others were MIT graduate students, but I think Henneman didn't come along until
much later.

13. Here I consider the reference counter idea as an alternative to g.c.

14. I had read the first part of Church's calculus of Lamda conversions and
when I needed function algorithms to define the       function 
maplist, I then used Lamda's.

15. It was the clearest language.

16. Not to my knowledge.

17. Yes

18. I'll check.