Let B be the board state, M be a move of a piece, and R be the current ruleset. These variables are related by one of the below functions:
B(M), implying an internally managed board and ruleset
B(B, M), implying internally managed rules and a user provided board & move
B(R, B, M), implying user provided rules, board state, and move, with Boardscript's implementation serving only as a simple input output function.
Which of these should be boardscript's behavior model?
B(M) is a monolithic continuously running program, requiring minimal input from clients and managing everything itself. This limits flexibility heavily, but it's in question whether that even matters. Boardscript is the game being played.
B(B, M) requires loading only the irules of the game being played, allowing it to serve as a way of transforming boards that already exist according to a move operation. Boardscript is an engine that facilitates the playing of games.
B(R, B, M) requires doing more but keeping track of naught. Boardscript applies M transformation to B according to R, then terminates. Boardscript is a stateless function.
Upon thinking, it seems the only difference between these 3 is WHEN the board or rules are loaded. As such, it may be worth simply launching the program in 3 different modes, "Game Mode," "Engine Mode," and "Function Mode" respectively for the above functions of B. Polymorphism perhaps?