Participants WCCC 2007

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Daniel Mehrmann
Posts: 855
Joined: Wed Mar 08, 2006 8:24 pm
Location: Germany
Full name: Daniel Mehrmann

Re: Participants WCCC 2007

Post by Daniel Mehrmann » Thu May 31, 2007 7:47 am

Kirill Kryukov wrote:
pijl wrote:The thing that got me angry (and perhaps I should have made that more explicit) is that Thomas wrote that people should accept that some people were not that silly to do all the work themselves (at least, this is how I read his post).

The title of this thread still says 'Participants WCCC'.
The WCCC should only be open to those people that are silly enough to create everything themselves (or at least in a cooperating team).

I have no problem with Toga as a GPL deriviate of Fruit which is obviously making some people happy.
I have no problem with gridchess as a research tool/subject

I do have a problem with either of them entering in the WCCC.
Richard.
If WCCC was conducted this way you like, no one would be allowed to use EGTB there, except a single team which has Eugene Nalimov as member. Do you realize this? And also which team it should be if he does not answer the mails?
Hmmm, i don't understand you. I don't understand a lot of people here. How can you compare EGTB code with a complete engine ?? Is this the same for you ? Also the EGTB code was never released on GPL, that's the reason why no GPL engine can use it, it's easy not allowed through copyrights by Eugene.

I don't understand the drift of this thread, direction Toga and Thomas, as well. Nobody said negative things about Toga or Thomas, but some people thinking he was attacked again. Why ? It seems some people want read, what they like to read, but not what was written. :roll:

Finally i agree 100% with Richard's view of points and i believe that Gridchess is on the participants list says nothing yet.
I think it wasn't smart from the ICGA to do that, but they can't reject Gridchess yet, because there is no proof for the ICGA that it could be some patchwork.

I'm sure they will ask for the sources before the tourney going to be start and hopefully decide the right things. :wink:

Best,
Daniel

User avatar
hgm
Posts: 23381
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Participants WCCC 2007

Post by hgm » Thu May 31, 2007 8:10 am

I agree with Kirill. It is ourtrageous that engines are allowed to use EGTBs generated by others. Especially if they use a standard probing code, this clearly violates the rules of being an original work.

This should clearly be outlawed.,

User avatar
Thomas Mayer
Posts: 383
Joined: Thu Mar 09, 2006 5:45 pm
Location: Nellmersbach, Germany

Re: Participants WCCC 2007

Post by Thomas Mayer » Thu May 31, 2007 9:40 am

Hi Harm,
hgm wrote:I agree with Kirill. It is ourtrageous that engines are allowed to use EGTBs generated by others. Especially if they use a standard probing code, this clearly violates the rules of being an original work.

This should clearly be outlawed.,
well, myself I would have no problems with that, Quark has already it's own bitbases. As consequence I think every engine should also have their own bookcode, the usage of GUI books shouldn't be allowed. And only GUIs should be allowed that do not interfer into the game play (like WinBoard) or be original for the engine... But well, it's doubtful that this all will happen - especially there are so many that love the Fritz GUI book features so much...

Greets, Thomas

User avatar
Thomas Mayer
Posts: 383
Joined: Thu Mar 09, 2006 5:45 pm
Location: Nellmersbach, Germany

Re: Participants WCCC 2007

Post by Thomas Mayer » Thu May 31, 2007 10:07 am

Hi Thomas,
Thomas Cutter wrote:
Thomas Gaksch wrote: I quit chess programming.
I hope, you will not drop this hobby entirely. The improvement of Fruit was considerably and surely you have more ideas to progress.

Keep on!
well, as far as I know the reason for Thomas to quit computerchess is the lack of time because he has a child now. And I understand him very well, when there is the question whether to use the time for computer chess or for your kid it's definitely better to chose the kid.
But he will come back, I am very sure about that, he is infected with the computer chess virus and that one may sleep for some years but never keep totally silent... :)

Greets, Thomas

Theo van der Storm

Re: Participants WCCC 2007

Post by Theo van der Storm » Thu May 31, 2007 8:37 pm

pijl wrote:
Kirill Kryukov wrote:
pijl wrote:The thing that got me angry (and perhaps I should have made that more explicit) is that Thomas wrote that people should accept that some people were not that silly to do all the work themselves (at least, this is how I read his post).

The title of this thread still says 'Participants WCCC'.
The WCCC should only be open to those people that are silly enough to create everything themselves (or at least in a cooperating team).

I have no problem with Toga as a GPL deriviate of Fruit which is obviously making some people happy.
I have no problem with gridchess as a research tool/subject

I do have a problem with either of them entering in the WCCC.
Richard.
If WCCC was conducted this way you like, no one would be allowed to use EGTB there, except a single team which has Eugene Nalimov as member. Do you realize this? And also which team it should be if he does not answer the mails?
Yes, I know. However, for EGTB and GUIs (as especially in UCI that is part of the game playing code also!) usually an exception is made.

If the ICGA or the players (in the players meeting) would decide that this is no longer acceptable I have no problem with that. It would stimulate programmers to create their own formats which many are already doing anyway.

I'm not using a GUI during the WCCC, and play/test often without EGTB's anyway.
Richard.
The data (book and EGTB) is not an issue as long as the use is not illegal.

The code for the screen lay-out and user interaction would not be considered "game-playing", because it has no influence on the move played by the program.
As an aside "GUIs" in the computer chess world tend to be more than this, e.g. also opening book access.

Strictly speaking the code to retrieve opening moves from a book and endgame position properties from an EGTB should be considered "game-playing code", which is the central phrase in the rule.

However allowing this code qualifies as a compromise, because the programmers gain the same improvement. There's nothing wrong with having a good compromise, but it should be made explicit.

User avatar
Thomas Mayer
Posts: 383
Joined: Thu Mar 09, 2006 5:45 pm
Location: Nellmersbach, Germany

Re: Participants WCCC 2007

Post by Thomas Mayer » Thu May 31, 2007 8:59 pm

Hi Theo,
Theo van der Storm wrote:
pijl wrote:
Kirill Kryukov wrote:
pijl wrote:The thing that got me angry (and perhaps I should have made that more explicit) is that Thomas wrote that people should accept that some people were not that silly to do all the work themselves (at least, this is how I read his post).

The title of this thread still says 'Participants WCCC'.
The WCCC should only be open to those people that are silly enough to create everything themselves (or at least in a cooperating team).

I have no problem with Toga as a GPL deriviate of Fruit which is obviously making some people happy.
I have no problem with gridchess as a research tool/subject

I do have a problem with either of them entering in the WCCC.
Richard.
If WCCC was conducted this way you like, no one would be allowed to use EGTB there, except a single team which has Eugene Nalimov as member. Do you realize this? And also which team it should be if he does not answer the mails?
Yes, I know. However, for EGTB and GUIs (as especially in UCI that is part of the game playing code also!) usually an exception is made.

If the ICGA or the players (in the players meeting) would decide that this is no longer acceptable I have no problem with that. It would stimulate programmers to create their own formats which many are already doing anyway.

I'm not using a GUI during the WCCC, and play/test often without EGTB's anyway.
Richard.
The data (book and EGTB) is not an issue as long as the use is not illegal.

The code for the screen lay-out and user interaction would not be considered "game-playing", because it has no influence on the move played by the program.
As an aside "GUIs" in the computer chess world tend to be more than this, e.g. also opening book access.

Strictly speaking the code to retrieve opening moves from a book and endgame position properties from an EGTB should be considered "game-playing code", which is the central phrase in the rule.

However allowing this code qualifies as a compromise, because the programmers gain the same improvement. There's nothing wrong with having a good compromise, but it should be made explicit.
well, there is definitely a difference between EGTB-probing-code - which just gives a score out of a database which what you can do what you want - and bookcode. The latter is definitely chessplaying itself. And you can do a lot in your bookcode, that starts with how you statistically use the basis-database, how you program learning etc. etc.

A lot of other programmers think as well that bookcode should be definitely something internal of an engine and shouldn't be shared. But of course I agree to Harm that also the egtb-probing code shouldn't be shared even when it isn't chessplaying in itself. As you said, I also believe that it was some kind of compromise especially because the pro's did use more and more the same GUI. But in the spirit of the WCCC it isn't the best compromise though.

Well, and in the case GridChess is really playing at Amsterdam future events by the ICGA should be boycotted, maybe one should even try to form something like a ECCA and try to hold our own world championships e.g. together with the ACCA.

Greets, Thomas

User avatar
hgm
Posts: 23381
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Participants WCCC 2007

Post by hgm » Thu May 31, 2007 9:17 pm

Theo van der Storm wrote:However allowing this code qualifies as a compromise, because the programmers gain the same improvement. There's nothing wrong with having a good compromise, but it should be made explicit.
The point is that this is not true. As I understand it, use of Nalimov probing code is illegal unless you have explicit permission from Nalimov himself for it. And from what I heard he has lost interest in answering requests for such permission. Plus I think there are conditions attached to such a permission, which you might not want to satisfy...

It is cheating plain and simple!

User avatar
Thomas Mayer
Posts: 383
Joined: Thu Mar 09, 2006 5:45 pm
Location: Nellmersbach, Germany

Re: Participants WCCC 2007

Post by Thomas Mayer » Thu May 31, 2007 9:36 pm

Hi Harm,
hgm wrote:
Theo van der Storm wrote:However allowing this code qualifies as a compromise, because the programmers gain the same improvement. There's nothing wrong with having a good compromise, but it should be made explicit.
The point is that this is not true. As I understand it, use of Nalimov probing code is illegal unless you have explicit permission from Nalimov himself for it. And from what I heard he has lost interest in answering requests for such permission. Plus I think there are conditions attached to such a permission, which you might not want to satisfy...

It is cheating plain and simple!
finally agreed. It might be okay to offer a version for download with nalimov code so that users can analyse with it, but for championships it should be disabled. Either use your own code there with maybe own databases or just don't use EGTBs. I am one of the lucky guys who have a permission by Eugene but you are right that it wasn't getting easier these days. (By the way, you need also permission by Andrew Kadatch for his compression code) Also open source engines with e.g. GPL are out of business because Eugene didn't want to have his code GPLed. So for next tournament I will either use only my own bitbases or some own written tablebase code.

Greets, Thomas

Theo van der Storm

Re: Participants WCCC 2007

Post by Theo van der Storm » Thu May 31, 2007 10:58 pm

Thomas Mayer wrote:Hi Harm,
hgm wrote:
Theo van der Storm wrote:However allowing this code qualifies as a compromise, because the programmers gain the same improvement. There's nothing wrong with having a good compromise, but it should be made explicit.
The point is that this is not true. As I understand it, use of Nalimov probing code is illegal unless you have explicit permission from Nalimov himself for it. And from what I heard he has lost interest in answering requests for such permission. Plus I think there are conditions attached to such a permission, which you might not want to satisfy...

It is cheating plain and simple!
well, there is definitely a difference between EGTB-probing-code - which just gives a score out of a database which what you can do what you want - and bookcode. The latter is definitely chessplaying itself. And you can do a lot in your bookcode, that starts with how you statistically use the basis-database, how you program learning etc. etc.

finally agreed. It might be okay to offer a version for download with nalimov code so that users can analyse with it, but for championships it should be disabled. Either use your own code there with maybe own databases or just don't use EGTBs. I am one of the lucky guys who have a permission by Eugene but you are right that it wasn't getting easier these days. (By the way, you need also permission by Andrew Kadatch for his compression code) Also open source engines with e.g. GPL are out of business because Eugene didn't want to have his code GPLed. So for next tournament I will either use only my own bitbases or some own written tablebase code.

Greets, Thomas
Thomas,

Again, this is all my personal opinion.
I agree what you wrote about the bookcode and what I wrote about it remains valid.

In critical game positions bad performance of home-grown EGTB access code can slow down the program so that the move selection by the program is affected, so it falls into the "game-playing code" category as I wrote.

Harm's statement about EGTB has merit. Of course cheating should be avoided. Maybe for the compromise w.r.t. EGTBs to work the ICGA should ask Nalimov and Kadatch permission on behalf of all participants.

If you feel so strongly about the Gridchess entry, why don't you prepare a rule amendment (just a few lines) to make things totally clear? Then it could be discussed and adopted for next year or maybe even this year.

Post Reply