Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The PGN specification, and attempts to change it

Author: Alan Cowderoy

Date: 13:23:55 09/11/01

Go up one level in this thread


>There are many topics for discussion here:

yes I entirely agree.

>
>1) Is Steven J. Edwards out, and if so, is this how ownership of the standard is supposed to be passed?

Good question. Before even considering publishing anything at all I attempted to
contact Steve Edwards. I was unable to do so. I discussed this with Mark
Crowther who had been in contact with him while writing his book but the address
he had for him no longer worked either. I posted a request for information to
rgcc without success.

I wrote in my post to rgcc that the document is intended as a **draft
proposition**. It was written in the spirit of trying, however imperfectly, to
put together some concrete proposals for discussion. If we have not submitted
these correctly to the relevant authority then I appologise, and I come back to
your own question - Who owns the standard? To whom should i (or anyone else for
that matter) submit proposals for changes to it?

>2) Is this a fait-accompli since Chessbase and Chess Assistant have apparently
>signed on to this?

Why would we be publishing **draft proposals** to rgcc if it was a fait
accompli? That makes no sense.

As I understand it Chessbase and ChessAssistant's position is that they would
like to see **something** implemented. If it is possible to agree an extension
then they are willing to support it. They are essentialy saying "agree something
and we will support it".

This then constitutes a genuine possibility to advance the standard with support
from the major chess database manufacturers.

The questions then are:
"Is there now any way to propose changes to the pgn standard?"
"If so how should this be done and who arbitrates?"
"Is encoding clock times in pgn a good idea?"
"If it is a good idea how should it be done?"

My guess is that if there is no way of extending pgn then DGT, CB and CA will
agree on some proprietary format or fork the standard.

>
>3) Does anyone have anything to say about the changes proposed?  Are they good
>changes?

Anders Thulin has made a fairly detailed set of criticisms of the proposals as
they stand which I will try and answer as soon as I can. While i have put a good
deal of effort into the proposals I realise that others are far more expert than
I in producing this sort of rigorous specification. I am genuinely gratefull for
help in getting this right

The question though is not only "Are they good changes?" but also "Are there
good changes we need to make? "Are there people who are willing to help get them
right" and "Who decides?"

>
>I don't know who started this effort, but the main point seems to be to design a
>way that the time on the clock after each move can be recorded in PGN.  There is
>no sensible way to do this now, so they use pseudo-comments to do this.

I did. It started after we used our pgn to html/js software Palview to cover the
Swedish championships live. This coverage worked a treat apart from the fact
that we were unable to display any clock times for the simple reason that as our
pages were generated from pgn we didn't have any clock times to display. (the
proprietary lost boys software does not have this problem of course)

As a result of that Axel Fritz from the German news site schach.de put me in
touch with Ben Bulsink of DGT who make the boards.  He is in contact with
Mathias Feist and Viktor Zakharov and the discussion rapidly turned to the
possibility of proposing an extension to pgn.

>
>A possible problem is that they are overloading their clock-time tag already.
>It is both a command to the viewer, which presumably could be displayed however
>the viewer wants to display it, and as an embedded argument (I can't think of
>another way to say this) to a normal comment.  Meaning that we have this:
>{[%clk 1:03:23]}
>and this:
>{White is in time pressure since he only has [%clk 0:00:22] left.}

Yes. There is a section of the document which seems to suggest that the commands
should somehow be replaced by text in order to complete the comment. I agree
that this is wrong and it has caused some confusion.
The clk command is intended to encode the clock time shown on a digital clock
(ie the time remaining to the next control.) and nothing else. We do not realy
imagine this command being written by anything other than software, and normaly
by software connected to electronic boards.
The emt command, which is the one used in the paragraph causing the confusion is
intended as a command to encode the time taken for the move in question. There
is no intention that this be a command to the viewer nor that the result
returned needs to somehow slot into the surrounding comment as the example does
indeed suggest. I stand corrected.

><snip></snip>I'm still concerned that there could be a
>rush to change the spec so that DGT and CB/CA can solve a technical problem, and everyone else could pay for this due to a broken or poorly designed set of PGN
>extensions.

Personaly i'm not rushing anywhere.

I'm just trying to see if we can solve the problem that you rightly pointed out
yourself in this thread in 1996
http://groups.google.com/groups?hl=en&safe=off&ic=1&th=1fe06718d5ec89a5,20&seekm=4psfnl%24i5m%40darkstar.UCSC.EDU

read on and Steve Edwards replies that the solution was the BIF format which was
never implemented. He even specificaly uses clock times as an example.

Someone else immediately suggested embedding the code in comments in order to
avoid breaking existing software - so we come full circle because five years
later that is what we are proposing and for exactly the same reason.

>
>Is this the way the PGN standard should go?
>
and if it isn't where should it go and who decides?

Regards
Alan Cowderoy
Palamede



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.