H. G. Muller wrote on Thu, Jan 28, 2021 10:51 PM UTC:
The problem with Fairy-Max was mainly that at the time I configured it for Wildebeest, there was no convention in the XBoard protocol for encoding such moves. And I didn't bother to think about it, because it seemed a quite useless move in the first place; normally you would want your King to go as far into the corner as possible.
But the problem can be solved, by adopting conventions that are not that strange if you think about it:
A move of the King to the Rook square can indicate a castling where the King indeed moves there. So it basically uses the normal encoding, and the only thing that is somewhat unusual about it is that the King seems to go to an occupied square.
A move of a Rook to the King square can be used for one-step (O1) castling. This is also not so strange, as indeed this would be where the Rook ends. (Both in classical and fast castling!)
The problem with Fairy-Max was mainly that at the time I configured it for Wildebeest, there was no convention in the XBoard protocol for encoding such moves. And I didn't bother to think about it, because it seemed a quite useless move in the first place; normally you would want your King to go as far into the corner as possible.
But the problem can be solved, by adopting conventions that are not that strange if you think about it: