On 3 Apr 2017, at 23:53, Asmus Freytag <asm...@ix.netcom.com> wrote: > Alternatively, a system that uses no Variation selectors and only relies on > Opentype ligatures might work even better. > > This would require one Empty and one Filled board cell, to ligate with > whatever piece is supposed to sit on top of it. The use of Empty / Filled > board cell would result in the correct metrics and by encoding an empty cell > that is different from a "white square", there's no need to overload the use > of the latter.
That would not be more legible — quite the opposite, in fact — because it would add variable length to lines for every character in the 8 × 8 matrix in any environment where the ligation failed. That’s not chess-legible. Michael Everson
