Source tags on CG (slight return)

There isn’t threading on CG, so let’s try to make one here for this discussion… some posts in the thread have been omitted at this editor’s (i.e. my) discretion:

Apr-22-16 zanzibar: <chessgames> can we get a status call on the Source tag, and maybe a prognosis, svp?
Premium Chessgames Member
luftforlife: <>: Will inclusion at this time of a source tag in a PGN submission for upload (specifically, in the PGN header thereof) run afoul of any of the policies, guidelines, or practices here on <>? Will inclusion of a source tag in a PGN submission for upload cause such a submission, even if it be otherwise acceptable in all other respects, to be rejected automatically, or reflexively, for that reason alone?I sincerely want to follow site policies, guidelines, and practices, and I don’t want my PGN upload-submissions to be rejected merely because I’ve included source tags in them. I apologize in advance if my having included source tags hitherto in my submissions has caused me unwittingly to run afoul of any site policies, guidelines, or practices, or has caused any difficulty or inconvenience in any wise to those who vet such submissions (who may, perhaps, have had to strip out the source tags because they have not yet been approved and instituted as regular parts of PGN upload submissions).

Thank you in advance for your kind consideration of my inquiries. If you’re unable to reply at this time, I would be grateful for a reply from an editor who might be authorized, in your stead, to provide official and authoritative answers to my inquiries. Best regards.

Premium Chessgames Member <Source Tags>

I have a couple of things to say. First, in chessforum (kibitz #24817) Zanzibar says

<What is our goal?
Our goal is to develop a system that encourages biographers who actually contribute new PGN to use a <Source> tag.>

I agree completely.

<luftforlife> You seem to think that a source tag would be a reason to reject your PGN at this time. That’s ridiculous. You can make a PGN tags called WhiteColorOfShoes and BlackColorOfShoes and we’ll just happily ignore them, but an unknown source tag is never reason for alarm.

However, for right now, it won’t be used. That gets back to Zanzibar’s goal.

I brought up the issue of small citations vs long ones, and I mentioned a 32 character limit which I think some people took too seriously. It was mostly a thought experiment. 32 would be painfully short, we could surely get away with 60 to 70. I brought up the ultra-trim notion because, one: it’s a good mental exercise, and 2: shorter fields really are better. If we decide to go with 64 we still should try to keep it short.

Some problems are begging to be addressed, like “Can you put quotation marks inside your citation and if so, how?” And, “Can you put a web link as a source?” I don’t the answers to these off the top of my head.

Next, it should be noted that the new “Editor Notes” can be used exactly for the purpose of showing the game citation when known. So if there are multiple sources and web links and all sorts of fancy stuff that would never fit in a 64 character field, we can always fall back on that for a truly comprehensive treatment. It won’t crop up a lot, but nice to know we have it when we need it.

On a related note, there is something that <zanzibar> said which I take exception to. I can’t find the post offhand, but he said words to the effect that “the more data we have in the PGN, the better.”

I think is not true, and it’s somewhat important to realize that. It was never intended to be a single encapsulated blob of data which covers everything from annotation to web links to photographs. I even take exception with the practice of including clock-times, in spite of its uncanny rise in popularity.

The ChessBase Corporation has data files for chess gams which are a lot like what zanzibar envisions for PGN: not only can it have annotation, but video annotation, and clock times and photographs and a complete interactive experience. It’s great for their format, but it would stretch PGN far beyond its intended purpose.

OK, you might say “You’re being silly. When I said it should have as much data as possible I didn’t mean multimedia, I meant simple ASCII facts in discrete tags.” Even then, I have to say to consider the worth of each field carefully. I enjoyed (I think it was) MissScarlet’s quip that we could have PGN tags that show what color shoes each player had on that day.

So all I’m saying is, new tags are great, but let’s keep the new tags useful.


I know that people are champing on the bit to see the new source tag inclusion and we’re working on it now. In fact, I am tempted to not process luftforlife’s submissions on purpose in order to use his games as the first round of testing.

The discussion continued over on the Bistro…

zanzibar: RE: <Source tags>
This discussion is being brought over from the <chessgames>forum. Here’s a post to begin the chessforum (kibitz #24952)

I hope to keep an edited thread of the discussion here:…

(Provided <CG> doesn’t object…)

zanzibar: RE: <Source tags>
Ok, we agree on the goal, that’s really the most important thing.But let’s roll up the shirtsleeves and get down to some details. Let’s use <chessgames> post as a starting point, and make general remarks to begin:

<(1) Mechanical details.>

The length limit of 64 characters strikes me as too short, given that PGN allows 255.

What is the rationale for being more restrictive?

The quote concern is good to raise early. I don’t use them, but I imagine that single quotes could be freely allowed.

If a user insisted on using double-quotes the PGN standard allows for that via backslash escape (\”):

<A quote inside a tag value is represented by the backslash immediately followed by a quote.>…

The <CG> input processor can ensure conformance during a pre-screening parsing.

Using the “Editor Notes (EN)” as a backup is a good idea – in fact I could envision autoloading an EN from a PGN submission with the source tag info.

zanzibar: RE: <Source tags>
From <chessgames> post:<On a related note, there is something that <zanzibar> said which I take exception to. I can’t find the post offhand, but he said words to the effect that “the more data we have in the PGN, the better.”>

<(2) The PGN should reflect as much information about the game as possible.>

This is a matter of degree, which <chessgames> turned into a reductio absurdum with shoe colors.

We all know that shoes have nothing to do whatsoever with the quality of chessplay…

On the other hand, certain fashion features apparently do affect play, and therefore the PGN should accurately record this information:…

zanzibar: RE: <Source tags>
Let’s be a little more serious, and use some real-world examples from my work to illustrate the choices and compromises made.<(2) The PGN should reflect as much information about the game as possible.>

When I wrote that (or whatever was the original), I’m sure I was thinking more about stubs than I was about fashion.

In fact, I’m sure that I wasn’t thinking about clock times. That’s actually a great example of some information one might like to keep, but which is more likely just cluttering up the PGN.

Let’s agree that it should be stripped out at this stage – and return to it in the next round.

Remember – PGN is primarily designed for programs to read, even if we all know how to read and write it by hand (or we all should, being responsible biographers).

Let’s consider some actual games from <Hereford> whose “proper” treatment involves several of the same issues raised by stubs.

(Next post)

zanzibar: For a first example, consider the following game:
< <1885.08.08 A02 28 (R7) 1-0 Mason — Owen>[ECO “A02”]
[Stub “incomplete”]
[Source “TB G-03 p15/20”]
[PlyCount “55”]

1.f4 f5 2.Nf3 Nf6 3.e3 e6 4.b3 b6 5.Bb2 Bb7 6.Be2 Be7 7.O-O O-O 8.Qe1 Qe8 9.Qg3 Nc6 10.a3 Nd8 11.Nc3 a6 12.Nd1 Nf7 13.Nf2 Kh8 14.Qh3 Ne4 15.Nxe4 fxe4 16.Nh4 Nh6 17.g4 Bf6 18.Bxf6 gxf6 19.f5 exf5 20.gxf5 Rg8+ 21.Kh1 Rg5 22.Rg1 Nf7 23.Ng6+ Kg7 24.Nf4 Qe5 25.Rxg5+ Nxg5 26.Rg1 Rg8 27.Qg3 Qxf5 28.h4

{ @stub – incomplete “(h) The winning of a piece is now immediately assured ; and though play was continued for upwards of a dozen moves more, Black found no resource, and so resigned.”}


I stripped out the canonical tags for conciseness.

So, I want the PGN to accurately reflect as much information as it can about the game, right? That’s the design goal.

The Source tag informs us that the TB (Tournament Book) was used for the game. But the TB only gave a truncated version of the game.

The PGN should reflect this, and indeed it does, categorizing the game as an “incomplete” Stub.

You’ll notice the convention I’ve adopted, having a concise Stub tag, and a concise PGN comment header, but allowing for more elaboration in the comment. In this case the TB ending note is copied wholesale.

Thus, a text search for @stub would turn up the game (or a search for a Stub tag). Stubs come in various varieties; <CG> doesn’t use 0-move stubs (at the moment) but other databases do.

<CG> certainly does use “incomplete” stubs, often. And more often than not without any indication of such.

But why brother?

Well, as a biographer, the desire for accuracy is intrinsic. So the answer there is just ’cause it’s the “right-thing” to do.

But there are practical reasons. For instance, in another similar game I first used a truncated game from the TB. But it turned out that the full score of the game was published in another contemporaneous periodical.

I therefore (for convenience) grafted the two versions together. Which required me to append two sources together in the Source tag:

[Source “BCM v5 (Oct 1885) G-367 p 347/362 (1-34) + C-M v7 (1885) G-590 p23 (34-48)”] >

Moves (1-34) from the 1st source, (34-48) from the 2nd.

Of course, in this case, I could just go back and type in all the moves from the 2nd source (or cheat and just delete the first source? – no never!). But hypothetically, one could find the beginning of a game in one source, and only the finish in a second.

This argues for being as generous as possible with the length restriction.

And having a game clearly marked as incomplete alerts the diligent biographer to be on the lookout for the complete score.

Congratulations if you’ve followed all these arguments. I’ve tried to be as cogent as possible. But in these detailed matters, some focus is naturally required.

(Enough for now…)



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.