Proposal to add standardized variation sequences for chess notation

[Michael Everson]

NOTE: A number of messages I sent in the last two days were scrubbed by the Unicode list software because they contained images. I will re-send these with links now. 

From: William_J_G Overington <wjgo_10009 at btinternet.com <mailto:wjgo_10009 at btinternet.com>>
Date: 2 April 2017 at 12:05:03 IST

> I included the regular Unicode chess pieces themselves, and for each chess piece also versions on a white square and on a black square in the Private Use Area of my Quest text font.

OK, I’m looking at this. William’s description uses rather different terms than mine does, so I’ll try to translate. 

First, he’s describing a font he made in 2004 or 2005, not an implementation of my proposal. 

> Free download of the Quest text font from http://www.users.globalnet.co.uk/~ngo/fonts.htm <http://www.users.globalnet.co.uk/~ngo/fonts.htm>
> 
> Thus, for, say, White King, there are three glyphs.

Correct, just as my proposal would have it. And the metrics for e.g. white-king-for-use-in-text and white-king-for-use-in-chessboard are different. 

> The Quest text font has descenders, so that while the glyph for White King itself is sat on the baseline, the glyph for the White King on a white square has the chess piece positioned lower vertically. The background shading for White King on a black background goes down to the WinDescent level.

By this he means more or less that the glyphs intended to produce a board have em-square metrics, while the base glyphs do not have square metrics and would be more suitable for in-text usage.

[Picture of William’s three characters showing metrics]

http://evertype.com/standards/unicode-list/overington-glyphs.png

> Line spacing could be an issue, but it need not be as long as the OpenType-supporting application where the font is used has the facility to set type with no additional spacing. I use Serif PagePlus X7 and the facility is there, so diagrams look fine.
> 
> I hope that Michael's proposal goes forward and is accepted.

So does Michael. 

> Regarding the borders. I note that use of a variation selector is not suggested.

Nor should it be. 

> As it happens, Quest text also has eight glyphs for producing a border, all eight being in the Private Use Area. They are rather ornate. They are at U+E5B0 through to U+E5B7.

They are there. I had to figure out how the should be used. They are put together in a very different way than the borders of any other font I have seen are. I am not sure, but I think he’s intended to use them thus:

[Pic of the Looking-Glass board in William’s font]

http://evertype.com/standards/unicode-list/overington-board.png

William’s design is decidedly non-traditional, and not (to my eye) particularly easy to read, but it doesn’t matter. The picture here shows his glyphs configured in exactly the same way as specified in my proposal. IT WORKS. (There are some hairline gaps in the border and the top left corner piece is a little less well aligned than one would if one were preparing to ship the font.) 

The underlying text just the same text that I used to set the Looking-Glass board in my proposal, variation selectors and all. There are no variation selectors used (or needed) for the border, though its glyphs are certainly unconventional horizontal lines. ;-) 

> The empty squares in the chess diagram each use a variation selector.

I dont’ see how. There were no OpenType instructions in the font or variation selector characters in the font.

> I opine that it would be helpful if a variation selector were to be used for each of the eight border items. 

No, because this would lead to potentially infinite variety of borders within any font, and it would be better to restrict this. I wouldn’t even want a VS to distinguish a single-line border from a double-line border, and there’s an enormous variety of ornamental borders one could put on the glyphs for 

> Using a variation selector would mean that a diagram could be produced without relying on the basic designs of the eight character sorts used to produce the border and also would allow a stylish border design to be included in a font.

A chess font is best when optimized to the design the designer wants, but honestly, the model proposed is simple and robust and does not need more tinkering or more complexity. It is able to support William’s design as well as the more traditional ones in the proposal while remaining parseable plain text. That should be enough. That is what takes the mess that current non-Unicode chess fonts are in and normalizes them for use. 

> Best regards,
> 
> William Overington

Thank you for sharing your font, William. I’ll send you the ttf of this one so you can tinker with glyph placement as you wish, if the proposal is accepted and the standardized variation sequences accepted.  

Michael Everson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://unicode.org/pipermail/unicode/attachments/20170405/d63c27c6/attachment.html>