H. G. Muller wrote on Wed, Mar 7, 2018 09:27 AM UTC:
Castling where the King moves a single step, with the resulting danger of ambiguity, can also happen in Chess960. This is not a problem for notation there, as there is only one type of castling, so O-O or O-O-O suffices. But it is a problem for computer interfaces. The conventional solution is to enter castling by dragging the King onto your own Rook. (Or click Rook after selecting King, in the case of click-click moves.) In WinBoard I elaborated on that idea for Seirawan Chess, where castling can be accompanied by gating, either at the King or at the Rook square. To indicate the latter, you can drag your Rook onto the King. It does need a disambiguator in notation, however. (O-O/Eh van O-O/Ee.)
With flexible castling this method could be used to indicate castling where the King moves a single step (sO1 in XBetza notation). All other castlings can be entered by moving the King; there is no ambiguity for those. (At least with an orthodox King.) This allows all castlings to be entered by dragging a single piece, or separate clicks on from- and to-square.
When pieces that can move more than one step by itself, it is a bigger problem to allow indicate their castling. The 'advanced' variant of Omega Chess poses this problem: there the Queen can castle with a Rook ('Guarding'). To handle that I generalized the 'friendly-capture' method to multi-leg moves, using locust capture of a friendly piece as a signal that you want to castle with it. WinBoard supported such two-leg locust captures (where the final destination square is not implied by the victim, or vice versa) anyway, to handle the Chu-Shogi Lion.
Castling where the King moves a single step, with the resulting danger of ambiguity, can also happen in Chess960. This is not a problem for notation there, as there is only one type of castling, so O-O or O-O-O suffices. But it is a problem for computer interfaces. The conventional solution is to enter castling by dragging the King onto your own Rook. (Or click Rook after selecting King, in the case of click-click moves.) In WinBoard I elaborated on that idea for Seirawan Chess, where castling can be accompanied by gating, either at the King or at the Rook square. To indicate the latter, you can drag your Rook onto the King. It does need a disambiguator in notation, however. (O-O/Eh van O-O/Ee.)
With flexible castling this method could be used to indicate castling where the King moves a single step (sO1 in XBetza notation). All other castlings can be entered by moving the King; there is no ambiguity for those. (At least with an orthodox King.) This allows all castlings to be entered by dragging a single piece, or separate clicks on from- and to-square.
When pieces that can move more than one step by itself, it is a bigger problem to allow indicate their castling. The 'advanced' variant of Omega Chess poses this problem: there the Queen can castle with a Rook ('Guarding'). To handle that I generalized the 'friendly-capture' method to multi-leg moves, using locust capture of a friendly piece as a signal that you want to castle with it. WinBoard supported such two-leg locust captures (where the final destination square is not implied by the victim, or vice versa) anyway, to handle the Chu-Shogi Lion.