COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\|\\M1BASL30\M2NGR40L\M3NGR25\M4NGR20\F2\CSTANFORD UNIVERSITY
C00008 ENDMK
C⊗;
\|\\M1BASL30;\M2NGR40L;\M3NGR25;\M4NGR20;\F2\CSTANFORD UNIVERSITY
\F3\CSTANFORD, CALIFORNIA 94305
\F4ARTIFICIAL INTELLIGENCE LABORATORY\←L\-R\/'7;\+R\→.\→S   Telephone:
\←S\→.415-497-3330

\F1\CMay 29, 1979


Mr. Harry Shershow
Personal Computing
1050 Commomwealth Avenue
Boston, Mass. 02215

Dear Mr Shershow:

\JLet me first answer your question regarding Strachey's part in the early
work on checkers.

Stachey was indeed interested in checkers at a relatively early date,
although perhaps not as early as 1947 when I started to work on my program
while at the University of Illinois.  Babbage had earlier proposed having
his Analytic Engine play draughts and chess so we were both scooped.

My first checker program for the ILLIAC was never actually run as the
ILLIAC was still only a paper design when I left the University for IBM in
1949.  It was not until early 1952 that I had my program running on the
experimental model of the 701.  Incidentally, this first program was
written directly in machine code before we even had a symbolic assemblier.

I learned of Strachey's work from the paper that he presented in Toronto
in Sept. 1952.  Since he was by then at the publication state, I must
concede defeat.  It was not until 1954, with the advent of the IBM 704
that my program was able to play an interesting game.  My contribution was
to add learning to the program, and I believe that I can claim a first for
this.

As soon as I could, I arranged to meet Strachey in Cambridge England, It must
have been in December of 1953 because Boden gave me a copy of his book during
this same visit.  At any rate, Stachey had not progressed much further than what
was reported in Boden's book and his program was not then in a playable
condition.  We did have a very interesting meeting and Strachey was full of
ideas as to how I could improve my program.  In reviewing the credit lines in my
1959 paper, it now seems that I did not give him enough credit, although I did
state that "The basic program used in these experiments is quite similar to the
program described by Strachey in 1952".  Strachey and I became good friends and
my wife and I were invited to his home on a number of occasions.  He was a very
brilliant man and it was a real loss when he died so young.

Now as to my Vice Versa program- There is really very little to say except
that this game seems to be ideally suited for a computer with a display,
such as the Video Brain, because of the practical difficulty of making
plays correctly when using an ordinary board.  The game lends itself well
to an Alpha-Beta tree search and it has a reasonable branching factor so
the program can look quite far ahead.  The evaluation aspect is a bit
tricky as much less is known about this game than is known about checkers
or chess.  It was also a challenge to get it all within the RAM and ROM
limits imposed by a micro computer but I had already faced this problem
with my checker program.  The Video Brain version is really quite some fun
to play.\.

\←L\→S\←R\-L\/'2;\+L\→L
Sincerely yours,



Arthur L. Samuel
\←S\→L
ALS:pdp10