Uncertain date in PGN

Let's say I wanted to create a PGN of a chess game and I only knew the century or decade it was played in, but not the exact year. How would you go about recording that info in a format-compliant way?

I'm guessing something like:

[Date "15??.??.??"]
[Date "196?.??.??"]

Is this correct?

https://www.chessclub.com/help/PGN-spec

Excerpt, bold text is my emphasis:

8.1.1.3: The Date tag

The Date tag value gives the starting date for the game.  (Note: this is not
necessarily the same as the starting date for the event.)  The date is given
with respect to the local time of the site given in the Event tag.  The Date
tag value field always uses a standard ten character format: "YYYY.MM.DD".  The
first four characters are digits that give the year, the next character is a
period, the next two characters are digits that give the month, the next
character is a period, and the final two characters are digits that give the
day of the month.  If the any of the digit fields are not known, then question
marks are used in place of the digits.

Examples:

[Date "1992.08.31"]

[Date "1993.??.??"]

[Date "2001.01.01"]

I could argue 15?? either way; it's the year "field" so if you don't know it, you don't know it and it should be "????".

An obviously more useful (to a human reader ...) notation is 15?? for a sixteenth century game for which you don't know the precise year.

I would probably lean toward the former ("????") if it had crossed my desk when I had to worry about standards compliance, but it's really one of those awkward things: something the people who wrote the standard didn't consider, leaving everyone to make "least worst" choices. A program that couldn't handle ????.??.?? could unequivocally be said to be broken. 15??.??.?? is arguable either way.

Hmm, I see... I suppose I could try creating such a file and testing it in all of the major database programs. If it's read and sorted properly, then I guess you could say that that's about as 'correct' as you can get given the lack of clarity in the specifications. (I suspect your programmer's instincts may be right about this, though.)

This came to my mind recently while I was transcribing a game in the Oxford Encyclopedia of Chess Games played by two unknown players and taken from a sixteenth-century manuscript. It's listed as ? vs. ?, played in the year 15?? (truly the Match of the Century!). I think I've also seen games where the year is only given approximately (e.g. 1964?) or within a certain decade (e.g. 193x).

Anyway, thanks again, skelos!

It's difficult; you could use a non-standard header but you really want to put what you know of the date into the [Date: "..."] header.

If the standard were being maintained, it would be reasonable to submit a question or proposal for clarification. I don't think the PGN specification is maintained, however, it just is. I'd like to be wrong on that point; I would like PGN format to be extended to use Unicode to represent names more accurately. They might want to be provided twice: a Unicode version to use if you can, and a fallback ASCII version.

I'm sure I can think of other things too: FIDE ID comes immediately to mind.

Yeah, that's the problem... It hasn't been revised since 1994, which isn't yesterday, and certain programs have implemented their own additions to it, some of which have been widely adopted (e.g. arrows and highlighted squares), others, not so much.

I suppose I could just either enter an approximate date (e.g. "1500") or enter a text comment at the beginning of the game string, like "{15?? Europe}". Lots of ways to get around it, I'm sure.

It could be revised, but it'd take agreement from some of the major players (ChessBase et al) and preferably the cooperation or at least approval of the original authors.

Fixing without doing "too much" or making it backwardly incompatible ... challenging.

chess.com (at least, possibly others) have started using structured comments for clock information per-move. chess.com would be a useful sponsor of any change too.

FIDE has an interest as they define algebraic notation in the laws of chess. 0-0 vs O-O might not seem like a large thing but it would be nice for FIDE to bless both. (PGN does want to keep O-O rather than 0-0.)

I'd have to check if "=" for promotion is in the current FIDE laws but whether it is or not it's commonly seen. (OK, generalising here from scoresheets I've seen, but I'll bet it's common.)

I'm not up for trying to get the necessary people involved started on a multi-year project to get us to PGN 1.1 (or 2.0 if there's enough change). [One of the lacks in PGN is a version number.]

P.S. I'd be up for the job if someone were willing to sponsor it, but it would take multiple meetings, lots of email, many hours and likely a couple of years before a new version would get signed off.

Then would the working group shut down or continue ...