💡📝H. G. Muller wrote on Sat, Sep 30, 2017 06:26 PM UTC:
Thanks for warning me. It should work again, now. When I added the interface with the engine two variables were not declared in the diagram script, but in the page that defined the diagram on which I tested. So it did no longer work with the existing pages. Now I declare those variables in the diagram script.
As to your earlier questions: 'q' and 'z' are not yet implemented in the Betza compiler. I did not kow any variants that featured Rose or Crooked Bishop. Do you have need for them? 'pp' For multi-hopper (which I would need for Tenjiku Shogi) is also not implemented yet.
As to 'j': In the original Betza notation this is ill-defined for oblique moves, like its counterpart 'n'. You would not know to which square(s) it appies. It also seems redundant; e.g. a Dababba that must jump ('jD') can also be written as a hopper with range 2 ('pR2'). In XBoard's XBetza implementation I overloaded the 'j', giving it a different meaning on atoms that do not move 'over' any squares, like W and F: on sliders using such steps the 'j' would mean they skip the first square ('ski-slider'). This is just for conveniece, as it can also be written with the aid of the 'again' modifier: a Ski-Rook would be 'ygaW', meaning that after a first Wazir step it toggles range to infiniteboth on an empty square ('y') as on an occupied one ('g'). But 'jR' looks less obscure.
On the 'O' atom XBetza uses for castling, a 'j' would mean to skip one file when looking for the castling parter starting from the corner. So 'sjO2' would mean castling not with the corner piece, but with the pieces one step inward (i.e. the Knights in FIDE). Omega Chess needed this. But the diagram uses a kludge for this: it always allows castling with a piece that moves like a Rook. So in the diagram I currently ignore the 'j'.
Thanks for warning me. It should work again, now. When I added the interface with the engine two variables were not declared in the diagram script, but in the page that defined the diagram on which I tested. So it did no longer work with the existing pages. Now I declare those variables in the diagram script.
As to your earlier questions: 'q' and 'z' are not yet implemented in the Betza compiler. I did not kow any variants that featured Rose or Crooked Bishop. Do you have need for them? 'pp' For multi-hopper (which I would need for Tenjiku Shogi) is also not implemented yet.
As to 'j': In the original Betza notation this is ill-defined for oblique moves, like its counterpart 'n'. You would not know to which square(s) it appies. It also seems redundant; e.g. a Dababba that must jump ('jD') can also be written as a hopper with range 2 ('pR2'). In XBoard's XBetza implementation I overloaded the 'j', giving it a different meaning on atoms that do not move 'over' any squares, like W and F: on sliders using such steps the 'j' would mean they skip the first square ('ski-slider'). This is just for conveniece, as it can also be written with the aid of the 'again' modifier: a Ski-Rook would be 'ygaW', meaning that after a first Wazir step it toggles range to infiniteboth on an empty square ('y') as on an occupied one ('g'). But 'jR' looks less obscure.
On the 'O' atom XBetza uses for castling, a 'j' would mean to skip one file when looking for the castling parter starting from the corner. So 'sjO2' would mean castling not with the corner piece, but with the pieces one step inward (i.e. the Knights in FIDE). Omega Chess needed this. But the diagram uses a kludge for this: it always allows castling with a piece that moves like a Rook. So in the diagram I currently ignore the 'j'.