These note notes have been transcribed from a December 1969 handwritten record
in the first Wita-Awit research notebook.   
Item 1 in http://webdocs.cs.ualberta.ca/~tony/T.A.Marsland.2004-58.pdf

2 Dec						Status of WITA 2

    On Sunday 17th Aug. WITA 2 began successful operation.  The difference 
between WITA 1 and WITA 2 is very fundamental and is centered on the move 
listing procedure.  From any given board position all the available moves for 
both sides is known.  Such a list will also contain "illegal" moves of the type 
which place the King in check.
    Whenever a move is tried, only the pieces directly affected (the moving 
piece, the captured piece --if any-- and those pieces "controlling" the FROM 
square and the TO square) have their move list recalculated.  Thus only 
perturbations of the board are considered.  In recomputing the moves of the 
"controlling" pieces only the sub-list in the direction of the move is computed 
and then edited into the total move list for that piece.  This feature was not 
debugged until November.
    The new state vector includes both mobility of a piece, as used in WITA 1, 
and control of a square.  Special importance is given to control of the squares 
surrounding the King.  However the most important change from WITA 1 lies in the 
way in which the attacking and defensive scores are computed.  Instead of 
referring only to the moving piece, a cumulative score for all the pieces attacked 
and defended is maintained.  By this means it is possible to ensure that the 
scoring function is only dependent upon the current board position and not upon 
the route taken to that position.  There is still an inconsistency, however, since 
discovered attacks or blocking defences are given extra scores.  Since these are 
relatively uncommon conditions they are unlikely to interfere severely with the 
ultimate goal of generating a simple "hash" table to eliminate the analysis of 
duplicate position in the search tree. 
    A series of tests to determine WITA 2's effectiveness and to "tune" its 
scoring function will begin in December.  Matches against Hans Berliner's 
program, Greenblatt's and Roth's (Texas A&M) will be arranged.
    The latest "improvement" to the scoring function, which was implemented during 
November, consists of making "partial" or incremental moves.  Then, instead of 
retracting every move, account is kept of the current position of the moving 
piece and it is incremented to its next position.  A formal retraction to the 
original position of the piece is made after every capture or when the move list 
for that piece is exhausted.

                   signed             T.A. Marsland 2 Dec. 1969