Proposal to add standardized variation sequences for chess notation

[Michael Everson]

On 5 Apr 2017, at 15:52, Garth Wallace <gwalla at gmail.com> wrote:

> […] I'm just saying that if having symbols without VS not match either of the VSes is a sticking point, it's not hard to work around.

Oh, I see. �� Well, yes, I agree with you in part. But here’s the thing. 

It is *permissible* for proportional-inline-chesspieces to be identical to emsquare-chessboard-chesspiece if a designer *wants* to do it that way. But it is *just* as permissible for proportional-inline-chesspieces to be truly proportional and unsuitable for chessboard typesetting (and that’s how it has been since Unicode 1.1).

Look, here is a choice:

U+2654 - WHITE CHESS KING whose width might or might not be 
U+2654 FE00 - WHITE CHESS KING whose glyph is a white/light em-square for chessboards
U+2654 FE01 - WHITE CHESS KING whose glyph is a black/dark em-square for chessboards

I think this is enough. Or it could be:

U+2654 - WHITE CHESS KING whose width might or might not be 
U+2654 FE00 - WHITE CHESS KING whose glyph is the same as the unmodified U+2654, whatever it is
U+2654 FE01 - WHITE CHESS KING whose glyph is a white/light em-square for chessboards
U+2654 FE02 - WHITE CHESS KING whose glyph is a black/dark em-square for chessboards

There’s some precedent for this, where some symbols have one VS for “text glyph” and a different VS for “emoji glyph” and of course the unmodified symbol can be used and will display as the font has it.

I don’t think the second is necessary. It’s not necessary for this, for example:

U+0030 - DIGIT ZERO
U+0030 FE00 - short diagonal stroke form
U+0030 FE0E - text style
U+0030 FE0F - emoji style

OK, “text style” is identical to unmodified U+0030, but the only reason that attribute exists is in distinction to “emoji style”. Compare also:

U+1000 - MYANMAR LETTER KA
U+1000 FE00 - dotted form

>>> Currently, chess fonts can be (roughly) divided into "diagram fonts" and "notation fonts”.
>> 
>> That’s not true. There are some which do all three.
> 
> There are, sure. I said roughly: many don't do both & rely on font-switching.

But even more of them can’t rely on font-switching because the encoding of the piece on light and dark chessboard varies from supplier to supplier. All current chess fonts are ASCII hacks. 

>>> None of the features required for a diagram font are unacceptable in figurine notation:
>> 
>> The white ones may be too wide for use in text.
> 
> Not visually ideal, but legible.

Yes but if we were to unify unmodified chesspieces with the pieces on white squares it could invalidate the metrics of text like http://evertype.com/standards/unicode-list/34-variantim.png

As I say, it’s *permissible* to have the unmodified chesspiece glyph be the same as the white-square chesspiece glyph, but it’s not obligatory, and we must preserve font designer choice here. 

Michael Everson