Proposal to add standardized variation sequences for chess notation

William_J_G Overington wjgo_10009 at
Wed Apr 5 07:22:46 CDT 2017

Asmus Freytag wrote:

> .... - relying solely on ligatures has the benefit of not involving the UTC at all, therefore it could be implemented today without delay).

I am wondering whether that is correct.

Where one implements a ligature using a ZWJ without the Unicode Technical Committee having agreed then that is fine where the meaning of the text is unchanged: for example, if one chooses to include, say, a pp ligature in a font.

Yet to implement a ligature using a ZWJ where the meaning is changed, then I am wondering whether that needs the agreement of the Unicode Technical Committee.

There have been some recent encodings where ZWJ has been used with two or more emoji characters to produce a new emoji character where the meaning of the result is different from the combined meanings of the ingredients, the meaning of that new character not always or maybe never being congruently obvious unless one already knows the meaning.

If a ZWJ encoding for producing chess diagrams were to be introduced, then if it is not UTC that decides the detail, then who does decide? Would a non-UTC decision be interoperable, would it be supported?

There are details that would need to be decided, such as how to produce an unoccupied white square and how to produce a white knight upon a white square. In my opinion a white knight upon a white square would need a ligature as the glyph might be different from a white knight in running text.

It would be helpful if at the next UTC meeting the UTC could issue a statement clarifying with precision the situation over using ZWJ in this manner, maybe in relation to emoji as well.

Going back to Michael's proposal, I like the proposal for the board and I hope that UTC accept it for inclusion in The Unicode Standard.

I opine that the encoding should allow that the glyph for a white knight upon a white square is different from the glyph for the white knight that is used in running text. The advance widths of the two glyphs might be different each from the other, and the vertical position of the contours within the glyph may be different each glyph from the other.

With regard to the border of the board I opine that it would improve the proposal if a variation selector also applied to the eight characters used for the border.

This would mean that the glyphs for a chess board and its border could all be separate from the glyphs of other items in the font.

This would mean that where there is an open source font available and licenced for making derivate fonts provided that the name of the font is changed then chess diagram glyphs and chess diagram border glyphs could be added to a font and satisfactory results obtained.  

On a different aspect of this thread, I have a metal type chess fount, bought in the 1960s. The fount is suitable for handsetting and printing chess diagrams. It was great fun, and changing a diagram after printing it by moving a knight was quite interesting as that process involved four pieces of type: removing a knight on one colour of square, putting an empty square of that colour in that place, removing an empty square of the other colour and then putting a knight of the same original knight colour on a square of that other square colour in its place. The knight did not take any other piece in that move. 

The fount was cast by the typefounder from matrices supplied by the Monotype corporation.

The white squares were included in the fount, one does not have to rely on using normal spacing material for a white square.

There were also four long thin typemetal pieces for the border.

So here is a puzzle that results from that experience and yet also relates to the encoding of chess diagrams as in this thread.

Suppose that one has a diagram for a valid position in a game of chess.

Next one wants a diagram for the next valid position in the game.

For the second diagram, first make a copy of the first diagram and then change some of the glyphs.

How many glyphs need to be changed depends on the first position and the move that is made.

Can you find examples, where 2, 3, 4, 5 and 6 different glyphs are involved: that is, the total number of both glyphs that are moved out from the diagram and glyphs that are moved in to the diagram?

I have just made up that puzzle and I think that a result for one of the numerical values may not be possible, though I am not sure of that,  but that all of the others are possible sometimes.

With a metal chess fount, removed characters are carefully cleaned and then carefully placed back into the typecase.

William Overington

Wednesday 5 April 2017

More information about the Unicode mailing list