About the last move highlighting. I don't like highlighting in general, ever (I find it distracting), but here I can barely notice it since it is so faint. So, in this case, it doesn't bother me. But if it were brighter or more obvious, it would.

I'm confused--those are sorted perfectly by round. (Maybe somebody added the date information after your post?)>

Well, there may be some confusion from the fact that <Buenos Aires (1970)> had about 50% of the games missing round number - until yesterday.

That may thrown the sorting off from what I was expecting, leading to my inquiry.

<In any case, you are right that if the date is known but the rounds are known, it won't sort properly. In fact it won't sort at all, except by the year.>

Before commenting on this, is there a typo?

<I am contemplating a kind of compound field the incorporates the date (if known), and the tournament (even if the date is unknown), and the round number, so that every game can have an integer that represents its proper order in a sequence. I think that would work better than simply sorting on date, especially in cases where we have only partial information. If nothing else, that would force it to group like tournaments together in general searches.>

Yes, grouping a player's tournament games together in a generic list would be the top-priority.

(Normally a tournament would be Event/Site/EventDate, but CG could/should have a tournament index for each game).

Then the next ordering priority should be by round number - if known.

(Often tournaments have delayed games played later - rarely does anybody list the game chronologically. In fact, does any publication do so?)

* * * * *

The concept of each game having a tournament index (which could be -1, if unassigned) is important for maintenance of the integrity of the database.

Ah, a man (or kitten) after my own heart.

<About the last move highlighting. I don't like highlighting in general, ever (I find it distracting)>

We've received two emails asking us how to turn it off, as well.

Before commenting on this, is there a typo?>>

Yes, I meant to say "if the date is <not> known but the rounds are known."

<The concept of each game having a tournament index (which could be -1, if unassigned) is important for maintenance of the integrity of the database. A topic to which I'll return at some point, I'm sure.>

http://www.uschesschamps.com/sinque...

We can revisit the sorting issue again later. It's not a big priority - though it really would be helpful to group by Event/Site (and then sort by Round/Date) when looking at just a single player's games.

I recall our previous discussions with the tournament number, i.e., some internal index <CG> uses.

The tournament index I've talking about is the tid used to organize the promoted tournaments.

E.g. Staunton vs Saint Amant (1843) being tid=79243

The tournaments have a forward link to the games, I think the games should have a backwards link to the tournaments.

This backlink would be quite useful in preserving tournament integrity, since it could be used to enforce more rigorous/cautious updating of the PGN.

You may well be correct. Since intonations aren't shown on player names here, I may have picked the wrong form of one or more elements of her name.

But I have to admit that I'd never thought of "Marquis" as part of a Chinese name. :-D

Braun Knows?

Is it legal for PGN to use '[' instead of '{'

I don't think so, even though SCID and pgn4web both seem smart enough to handle it, but my program chokes on it.

If it isn't legal, then I don't think CG should use it.

I wanted to double-check here first before submitting a correction slip though.

Example games are fortunately few, so far I have: 1042345, 1071311

<chessgames> thanks, that's what I thought.

And you're right, "[" and "]" tokens are allowed inside comments.

I might have been chasing a red herring - I think I found the problem but have to "unwind the stack" to see how it came about. Those two games probably shouldn't have been cited. But thanks for looking.

And while you're here- there might be a related problem(?), missing <EventDate> tags.

I don't think <EventDate> is mandated by the standard, but for tournaments it probably should have been. My experience is that the tag is always present (and my sentiment is that it should be).

However these three games don't seem to have it:

;; <Viktor Korchnoi -- Efim Samoilovich Stoliar 1957 (R1) 28 1-0>

;; <Daniel Stellwagen -- Friso Nijboer 2003 (R5) 69 1-0>

;; <Daniel Stellwagen -- Erwin L'Ami 2003 (R9) 26 1/2-1/2>

Shouldn't they have it?

(Out of 230005 <CG> games I looked at, these are the only ones missing it.)

I would think that pgnfix.pl would right such wrongs, but apparently it doesn't. That would be the best way to address the issue.

For example:

http://www.chessgames.com/perl/ches...

shows as a clickable partial URL and not as <Monte Carlo (1903)>.

Monte Carlo (1903)

Monte Carlo (1903)

It's been my experience that it's very difficult to set up code-fixing algorithms/search-and-replace strings that anticipate all the ingenious ways people can unwittingly mess up the code. :-D

Just as a side note, I tracked down the "[" inside a "{" problem.

My parser normally would handle it OK - but I put in a special case to handle potentially corrupted PGN, where the input parser resynched if a "&lbrack" was found in column 1.

That rationale was to allow for things like a missing "}", or a missing end-result in particular. The idea is to confine the damage to the one game, and pick up fresh on the next.

The trouble now is that legal PGN gets rejected (just the one game). An example is this game:

M Umansky vs Berliner, 2002