These note notes have been transcribed from an April 1970 handwritten 
record in the first Wita-Awit research notebook.   
Item 1 in http://webdocs.cs.ualberta.ca/~tony/T.A.Marsland.2004-58.pdf

                               WITA 3

    During February 1970 I began experimenting with a method for Repetition 
Pruning.  That is, the recognition of transposition of moves during the 
current search tree, thus leading to subtrees that have already been 
examined.  Richard Greenblatt (1967) used a hash table to help with 
that problem, but it required 32K words of memory and I did not have 
that much "core memory" to work with.

    In April 1970 the program was renamed WITA 3, shortly after 
establishing contact with Monty Newborn.  My notebook shows:

1 April.  Received a call from Prof. Newborn of Columbia, 212-280-4229.  
Interested in entering my chess program in the ACM Show Sept. 1-3.  
Look for sponsor(s).  Consider BTL, COMNET, Y&R and Burroughs.

10 April.  Have now renamed the chess program to WITA 3.  New name 
reflects the new organization of the scoring function.  An attempt 
to reduce the referencing of REMAKE.   Compile time = 97 secs. 2449 cards 
(about 600 do not generate executable code).  Will try timing comparison 
with WITA 2.

4 June.  Very little activity on WITA during April and May.  Minor 
changes to organization.  Played one 25-move match with Bob Stubblefield, 
testing the lookahead procedure.  No great success.  Program still 
inadequate in the opening.

1 July.  WITA 3 back together again.  Played 6 more ASSIAC problems at an 
average speed of 2 mins/move.  Total elapsed time 76 mins! CPU = 10 min, 
I/O = 2.4 min.  Implemented a means of distinguishing between different 
kinds of pruning.  Will try once again to have Bob play the program.  Am 
going up to MIT this weekend will try to play Greenblatt's program.

29 July.  Met with Monty Newborn, Dennis Cooper, Keith Gorlen and Bob 
Stubblefield at Newark Airport last night.  We discussed the ground 
rules for our chess exhibit.
    After much heated discussion it became apparent that we could not 
agree on any suitable way of comparing the computers we were using, 
which include the U 1108, GE 635, CDC 6400, IBM 360/67, 360/91, B-5500 
and B-6500, and therefore my proposal for a "weighted" clock along with 
the times of the handicap system used in boat racing was not approved.
Certainly the complexities in the calculation are enormous.  My other 
suggestion that the relative speeds of the machines be compared via some 
"benchmark" program which was representative of a "fundamental" piece 
of chess program was not agreed upon since we could not describe 
(at that time) such a fundamental piece.
    Finally settled for a simple elapsed time clock rate of 40 in 
2 hours.  Tournament Director can stop any game after 5 hours.  Programs 
are to be run "hands off" and can only be altered with the T.D.'s 
permission.
    In case of Hardware, Program or Communication error the clocks will 
be stopped for up to 20 mins to allow correction of error.  A maximum 
of 1 hour time outs are permitted.
    As a result of these rules a minor modification to the program will 
be required, one which modifies the depth of search during TOURNAMENT 
mode based on the average elapsed time.  

14 Oct. 1970
    In September Joe Frisch wrote that WITA had failed miserably at the 
ACM'70 Tournament.  Although I can't imagine what the problem must have 
been, it looks as though the heavy load on the machine plus the 
mechanical difficulties with the TTY made it necessary to restart the 
program frequently.   From the game scores that I received it seems 
that WITA did not use lookahead.  Either the feature was not set, or it 
was automatically turned off because of the slow rate of play -- 
heavily loaded machine.
    Replied to Joe's letter about the middle of September.
    
1 Feb. 1971.  Received a tape from Ralph Sobek, containing the July 1 
copy of WITA 3.  This would appear to be the most recent version of 
WITA that is available.  

    A couple of weeks ago I wrote to Joe Frisch and asked that he try 
to locate the August 17 version of WITA 3 and to send a copy to me on 
the tape that I enclosed.
    Last week I wrote to Dennis Cooper enclosing a mini-tape, in the 
hope that he could put on a copy of his program.

------------------------------------------------------------------
From the above it is clear that WITA was nearly lost.  Recovery was 
possible only from copies given to other researchers and the source 
left on the Young & Rubicam host computer.   At some later stage WITA 3 
was converted to AlgolW (during 1971-73) and in 1975 that "new" version 
played in the 1975 ACM FJCC in Minneapolis.  
    Prior to that, a copy of WITA was given to the RIBBIT computer chess 
group at the University of Waterloo for testing, before they competed in 
the First World Computer Chess Championship in Stockholm, 1974.

Tony Marsland, Qualicum Beach, June 2011.