Enter Your Reply

Why is it impossible to give feedback (add rating or comments) on the
CVpage for FFEN / ffen2htm? I wanted to add the following comment:

It seems two separate issues got entangled here: extending Forsyth-Edwards
notation to unambigously denorte positions from Chess-variant games, and
the creation of a graphical tool for generating diagrams from such
positions.

W.r.t. the first issue I see no logical reason why FEN notation should be
limited to boards consisting of a grid of squares. In paricular, the
Xiangqi board is fully equivalent to a board of squares. But also boards
of hexagons (with overal hexagonal or diamond shape) can also be
raster-scanned for description as a FEN with, possibly, unequally sized
ranks.

Indicating if the board is checkered, and what is the color of its
lower-left squares are not part of the game-state at all, and thus do not
belong in a (F)FEN. They are merely input to a diagram-generator tool,
like font information, diagram size, etc.

I don't think it is a realistic desire that an FFEN uniquely specifies
the variant it represents a position from. This problem is well
illustrated in Janus vs Capablanca Chess. These games seemingly share many
positions (after the Chancellors in Capablanca got traded), but as long as
there are Pawns, the positions are still different, as in Capablanca the
Pawns can promote to Chancellor, which does not exist in Janus. In
addition, the rules for Queen-side castling are different. It would be
very inconvenient to require that such details are unambiguously described
by the FEN, as this would quickly make them infinitely complicated.

Note that FENs are mostly used in a context where the game they represent
is known: e.g. in a tag of a PGN file, which also containst a variant TAG.
So if a variant is known to have a diamond-shaped 8x8 board consisting of
hexagons, there is no reason why it could not be represented by a FEN as
would be used for normal Chess. Assuming the variant to be known would
also solve the problem that you run out of latin characters long before
the list of fairy-pieces is exhausted. That problem would not exist if the
same letter represents different pieces depending on the variant. E.g. a C
could denote Chancellor in Capablanca Chess, but Cannon in Xiangqi.

This is the way I implemented the FEN reader in WinBoard_F. Which, btw,
could also be used as a tool to generate diagrams as GIF files for many
variants (only with square boards of conventional topology, though). For
instance, the Xiangqi array would look like this:

http://home.hccnet.nl/h.g.muller/XQpetite.gif