💡📝H. G. Muller wrote on Fri, Oct 6, 2017 04:10 PM UTC:
Ah, good catch. I always use C and Z, because I can never remember which of L or J is the Camel. Fixed now.
As to the darkening of initial moves: the problem was that the i modifier was listed with the first leg, while it was tested at the destination, after the second leg. The diagram should now remember the i setting from the first leg even for multi-leg moves. I had never used initial multi-leg moves; I would be inclined to write the move as imN2. This does overlap with the normal move, however, generating it twice, with as a result that the move specified last will prevail. So the total move had to be written as imN2N to get the desired effect.
As to o for cyclinder and oo for toroidal: done!
I was aware of the e.p. problem; the imW6 move does not automatically grant e.p. rights. Furthermore, the code currently only uses 2 e.p. squares even when rights are granted, at 3/8 and 5/8 of the multiple push. Here up to 5 squares would have to be tested. So I should completely redesign e.p. testing. Also, the imW6 is not really describing the Pawn, as from 3rd rank it would still have W5 etc. This is really not an 'initial move', but a location-dependent move rule. For which Betza notation isn't really suitable. Unless something is invented to mean: up to half the board. (W* ?).
BTW, the q and z modifiers can now be combined with p and g, to get crooked and circular hoppers.
[Edit]I have redesigned the e.p. testing now. It tests all squares the rights-creating piece has moved through. This will work both for forward orthogonal and for diagonal moves. In addition, an asterisk as range now indicates 'distance to the 'River' (i.e. the line separating the board halves) or 1, whichever is larger. So that fW* can be used to indicate a Pawn that can always move up to River (and fF* a Berolina Pawn with such a range). Such a move will always create e.p. rights.
Ah, good catch. I always use C and Z, because I can never remember which of L or J is the Camel. Fixed now.
As to the darkening of initial moves: the problem was that the i modifier was listed with the first leg, while it was tested at the destination, after the second leg. The diagram should now remember the i setting from the first leg even for multi-leg moves. I had never used initial multi-leg moves; I would be inclined to write the move as imN2. This does overlap with the normal move, however, generating it twice, with as a result that the move specified last will prevail. So the total move had to be written as imN2N to get the desired effect.
As to o for cyclinder and oo for toroidal: done!
I was aware of the e.p. problem; the imW6 move does not automatically grant e.p. rights. Furthermore, the code currently only uses 2 e.p. squares even when rights are granted, at 3/8 and 5/8 of the multiple push. Here up to 5 squares would have to be tested. So I should completely redesign e.p. testing. Also, the imW6 is not really describing the Pawn, as from 3rd rank it would still have W5 etc. This is really not an 'initial move', but a location-dependent move rule. For which Betza notation isn't really suitable. Unless something is invented to mean: up to half the board. (W* ?).
BTW, the q and z modifiers can now be combined with p and g, to get crooked and circular hoppers.
[Edit]I have redesigned the e.p. testing now. It tests all squares the rights-creating piece has moved through. This will work both for forward orthogonal and for diagonal moves. In addition, an asterisk as range now indicates 'distance to the 'River' (i.e. the line separating the board halves) or 1, whichever is larger. So that fW* can be used to indicate a Pawn that can always move up to River (and fF* a Berolina Pawn with such a range). Such a move will always create e.p. rights.