Re: Proposal to add standardized variation sequences for chess notation

From: Michael Everson <>
Date: Thu, 6 Apr 2017 13:57:50 +0100

On 6 Apr 2017, at 11:00, Christoph Päper <> wrote:
> Michael Everson <>:
>> Standardized variation sequences are the best way to achieve this simply and without needless duplication. :-)
> I still agree with this assertion.

So do I.. ;-)

>> Yes but you still want it to be reasonably legible when the OpenType ligatures fail.
> This is were I don't follow.

Why wouldn’t you want it to be reasonably legible when the OpenType ligatures can’t be displayed?

is far better than this:
▕□︀□︀□︀♙︁♛︀□︀□︀□︀▏<< Is it the pawn or the queen that’s on the black square?

> It *looks* far better in a multi-line plain text environment, but that's a glyphic/typographic/stylistic argument.

It’s an argument for legibility.

> The semantics conveyed are redundantly encoded this way, so I wouldn't say it was far better. This alternating pattern is far more redundant than, say, pairs of opening and closing characters (brackets, quotation marks).

It’s not redundant to the reader. The reader of the second one has to remember that the dark square is the lower left, and then count in order to know the colour of any given square. The reader of the first one doesn’t have to do this, because we have both ▨︁ and □︀, two encoded characters, and we use them for convenience.

> Aside, good fallback isn't something the UTC seems to be concerned with lately,

Inconsistency on the part of the UTC is not my concern. I have to

> see emoji subregion flags that are all represented by Waving Black Flag in legacy implementations (possibly followed by TOFU).

Yes, well, that’s an example of a decision that didn’t have good oversight or feedback, perhaps. I do know that falling back to a black flag rather than to the Union flag for Wales, England, and Scotland doesn’t seem very sensible. Leaving out the de-facto flag of Northern Ireland wasn’t very wise either, though nobody asked the UK or Irish representatives of SC2 their opinion about it.

>> See? To parse this one you have to remember which of the white squares are the alternating black ones.
> No, you only have to remember that A1, i.e. the lower left square initially occupied by a white rook, is black.

You have to remember that, and then you have to count every other square in whatever direction to know what colour a given square is. That’s not very user-friendly. And it’s easy to be user friendly. Just use both ▨︁ and □︀.

> For legal moves, the color pattern hardly matters, unless - regarding pawns - it was common practice to render the board turned, i.e. with the white player not at the bottom, but at the top (or left or right) side, and without alphabetic column and numeric row labels.

For legal moves, no. But this is text. The table is meant to be read. Since it is, good fallback is better than bad fallback.

>> The colour of the matrix is NOT redundant for a human reader.
> That's what this proposal is all about. It's a good and sound proposal, except for the empty square.

Do you mean “except for the light and dark squares without a piece on them” or “except for the light square without a piece on it”? The convention is to have two alternating shades on the squares and there’s no advantage to the human reader to quash this distinction.

What is your specific counter-proposal?

Michael Everson
Received on Thu Apr 06 2017 - 07:58:15 CDT

This archive was generated by hypermail 2.2.0 : Thu Apr 06 2017 - 07:58:15 CDT