FIDE Ethics Commission ruling on ICGA/Rybka complaint

[Roger Brown]

And the primary authors do not wish to participate to the ICGA tournament. Case closed! Yet this thread (which is more about SF than Rybka by now) never ends. Why? You guys really have nothing better to do or what?

Hello Lucas,

There is a lot of good information in the thread which has admittedly morphed a number of times.

The fact that the primary authors have expressed their firm position - a position which has been mentioned, regretted and respected throughout the thread - does not preclude discussion about what are really fascinating issues.

The participation, or non-participation of Stockfish, has been overtaken by a consideration of who is the author of a project of the size and nature of Stockfish etc.

Gone are the days when an engine was easily traced to one author who was usually the tester as well.

It really is interesting stuff if you take a step back from viewing it as an extended attempt to "coerce" anyone to go against their principles.

Later.

[hgm]

I would expect the ICGA to do due diligence in investigating the provenance of an engine. There are plenty of people to ask.

As far as this example goes:

This is scattered throughout the StockFish6 source:

search.cpp: Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad

I'd think that would be the definitive list of primary authors.

Why would you want to burden the ICGA with even bothering about this?

What exactly is gained by imposing such a rule, and what is lost? It seems to me that the only 'merit' of such a rule is that it is ultra-coservative: it resembles most closely the rules that fail to obtain the desired results in the current environment. I see no upside whatsoever. Just more work for ICGA, more disagreement, fewer participants...

Also note that copyright does not necessarily coincide with authorship. They are tradable goods. For instance, the copyrights of Fruit are now held by the Free Software Foundation. By your guidelines that would mean Fabien no longer has the right to enter it, because the FSF now should be recognized as 'primary author'.

Also note that what the source code says need not actually be correct. In the case of XBoard the copyrights of any contribution that goes into the official version (at GNU Savannah) has to be transferred to the FSF, by signing official documents to that effect. (You sell your copyrights to FSF for a symbolic $1, which you never receive in practice...) So the FSF remains the copyright holder, and the contributors never appear in the copyright messages in the source files. I doubt if contributors for Stockfish have to do the same paper work, which would mean they retain the copyrights on their contribution, no matter whether the actual messages in the files fail to mention that.

And even if all contributors would sell their copyrights to Tord or Marco, so that the message is correct, I don't think the ICGA should mistake that for authorship. If I would buy the latest Rybka sources from Vas, making him an "offer he can't refuse", should the ICGA then recognize me as 'principal author' of Rybka, and allow me to enter it in the WCCC? (Note that Vas is banned from the WCCC for life, not me or Rybka!) When Vas would sell me legal ownership of the Rybka 4 copyrights, keeping the "economic ownership" (meaning our deal is shuch that any proceeds from selling the source keep going to him), should the ICGA allow me to enter Rybka 4 (assuming it is 'clean') as 'principal author'?

[bob]

Sure. The originators of the program. You can figure this out for most any program. Slate/Atkin for chess 3.x/4.x... Ken for the various flavors of Belle. Hsu/Campbell for Deep Thought. Hyatt/Gower/Nelson for Cray Blitz. Me for crafty. Fabien for Fruit. (and if anyone's name is misspelled, I simply get tired of matching wits with Apple and their auto-correction stuff.)

Well with this definition SF has one primary author and that is Tord....

Would you question who the primary SF authors are? Three names (and only three names) come to my mind instantly.

What "comes to mind instantly" is not a valid definition since it differs from person to person. I do not know what would come to mind instantly to the ICGA people having to deal with a potential SF entry

Joking aside, I for one would add Gary Linscott as a major author for his creation of fishtest which has allowed SF to gain hundreds of elo in a short period of time.

I'd bet if you asked a dozen people here, you would get the SAME three names every time.

Even if that is true you surely understand that this cannot be part of a formal policy of ICGA... You wouldn't expect the ICGA to base its rules on the results of a TalkChess poll, would you?

You have not given an unambiguous definition of "primary author" and I am sure you can't. The only unambiguous concept is "copyright holder" (informally: author). SF has about 80 of them and that number will continue to grow.

I would expect the ICGA to do due diligence in investigating the provenance of an engine. There are plenty of people to ask.

As far as this example goes:

This is scattered throughout the StockFish6 source:

search.cpp: Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad

I'd think that would be the definitive list of primary authors.

And the primary authors do not wish to participate to the ICGA tournament. Case closed! Yet this thread (which is more about SF than Rybka by now) never ends. Why? You guys really have nothing better to do or what?

Why are YOU chiming in? You are not one of the primary authors. Yet YOU simply add noise to the discussion for no reason???

[bob]

I would expect the ICGA to do due diligence in investigating the provenance of an engine. There are plenty of people to ask.

As far as this example goes:

This is scattered throughout the StockFish6 source:

search.cpp: Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad

I'd think that would be the definitive list of primary authors.

Why would you want to burden the ICGA with even bothering about this?

What exactly is gained by imposing such a rule, and what is lost? It seems to me that the only 'merit' of such a rule is that it is ultra-coservative: it resembles most closely the rules that fail to obtain the desired results in the current environment. I see no upside whatsoever. Just more work for ICGA, more disagreement, fewer participants...

Also note that copyright does not necessarily coincide with authorship. They are tradable goods. For instance, the copyrights of Fruit are now held by the Free Software Foundation. By your guidelines that would mean Fabien no longer has the right to enter it, because the FSF now should be recognized as 'primary author'.

Also note that what the source code says need not actually be correct. In the case of XBoard the copyrights of any contribution that goes into the official version (at GNU Savannah) has to be transferred to the FSF, by signing official documents to that effect. (You sell your copyrights to FSF for a symbolic $1, which you never receive in practice...) So the FSF remains the copyright holder, and the contributors never appear in the copyright messages in the source files. I doubt if contributors for Stockfish have to do the same paper work, which would mean they retain the copyrights on their contribution, no matter whether the actual messages in the files fail to mention that.

And even if all contributors would sell their copyrights to Tord or Marco, so that the message is correct, I don't think the ICGA should mistake that for authorship. If I would buy the latest Rybka sources from Vas, making him an "offer he can't refuse", should the ICGA then recognize me as 'principal author' of Rybka, and allow me to enter it in the WCCC? (Note that Vas is banned from the WCCC for life, not me or Rybka!) When Vas would sell me legal ownership of the Rybka 4 copyrights, keeping the "economic ownership" (meaning our deal is shuch that any proceeds from selling the source keep going to him), should the ICGA allow me to enter Rybka 4 (assuming it is 'clean') as 'principal author'?

If you don't, you reach another breaking point. What happens if program X is entered, and then someone else wants to enter X also. Which one takes precedence? For me, it would be the "primary author(s)" if they are one of the two entrants..

So you have to be ready to handle this. But I don't think it is that difficult. I have seen more than one list on the internet giving most every program and the primary author(s).

You can't enter a FIDE event without ID. It makes sense (to me) to at least be prepared for when this actually comes up.

[hgm]

If you don't, you reach another breaking point. What happens if program X is entered, and then someone else wants to enter X also. Which one takes precedence? For me, it would be the "primary author(s)" if they are one of the two entrants..

So you have to be ready to handle this. But I don't think it is that difficult. I have seen more than one list on the internet giving most every program and the primary author(s).

You can't enter a FIDE event without ID. It makes sense (to me) to at least be prepared for when this actually comes up.

I don't see that as a problem, and the 'primary-author' concept isn't really a solution:

If two co-authors want to enter exactly the same program X, it will be entered once on behalf of both of them. If they cannot agree amongst each other who should actually operate it... Well, I don't think the ICGA should put much effort in solving such childish quibbles, (e.g. by launching an investigation on on how much code exactly each contributed), it could just make them draw straws.

If there is a conflict between programs Y and Z that share code X, this could very well happen when the 'principal authors' of X and Y are different from any author of X, e.g. because X makes up only 40% of X or Y. So the principal-author rule could not solve this anyway. But you can simply make Y and Z play a knock-out qualifier match against each other.

Not that this is only a worst-case scenario, where they indeed would both like to enter. Much more likely is that one of them (say Y) wants to enter, and the other one doesn't, but just exists. In that case you don't have to do anything at all, and there would not be any reason to refuse Y.