I don't see why we should harass the 'source-code experts', or even doubt their judgement.
Strelka was not a cloning job, the source code was largely original. That verdict stands. That it was largely inspired by Fruit, doesn't make it a clone. Using partially the same data structures does not make it a clone. To convert an engine from mailbox to bitboard requires an almost complete re-write.
Using the same, publicly known algorithms is not cloning, or all engines on the WBEC list would be clones (perhaps with the exception of Pos).
Exactly copying the evaluation parameters of another program is not cloning. If Joker were to use separate piece-square tables for different piece types, I might be tempted to simply copy them from Ed's Rebel pages, and I would not have the feeling of any wrong doing. Why else would they be up there except as providing an easy source of that kind of data.
If I were to write a very fast engine, that would evaluate positions exactly like TSCP, but through completely different code, (say updating all required info differentially, rather than through counting on the spot, and using piece lists for fast access, using 0x88 in stead of 8x8 board, etc., but using the same scores for passers, backward pawns, etc.)... Would I even have to acknowledge that? Source-code inspection would classify it as an original work. And correctly so.
How much we want to allow people using each others eval parametrization and parameter values, is an entirely different discussion. And I am afraid that this will always remain a gray area, as to a certain extent almost everyone does it.
Daniel Mehrmann wrote: Also we don't know how much code was changed. I suppose it's near 99% anyway.
Somewhere you stated, that the violation of GPL was a very severe act. In the above qoute, you suppose, that 99% of the code was changed.
Is 1% a severe violation?
Regards Thomas
No, once it is GPL, it will be always GPL. You can change 100% of the code, its still GPL.
If it makes sense is another question to change 100%, because it's might be better then to start with your own files and an empty project.
But they provided by a "out-of-the-box" running engine and this is a huge advantage. (versus all other programmers).
So, the GPL tries to avoid this case that it says it will be always GPL state, also if you change 100% code of the project.
How many Fruit based engines would be come if the GPL wouldn't protect this state ?
Best,
Daniel
Last edited by Daniel Mehrmann on Thu Jul 12, 2007 1:35 pm, edited 1 time in total.
Rolf wrote:<...>Was that the main purpose of Osipov? Why didnt he analyse FRITZ or JUNIOR out of actual reasons? Because we all would lose our belief in the myths of commercial computerchess programming?
Who one would "normally" suppose standing behind Osipov? Surely not friends of RYBKA all IMO.
Hi Rolf
what is the main purpose of Osipov knows noone but (may be) him. I was trying to explain how I see the situation regarding the facts and statements I knew at this moment. No much no less.
Why he didnt analyze Fritz or Junior? I can only speculate about this. May be they are more original works in his eyes than Rybka is. May be he cant write a copy of Fritz or Junior rewriting to bitboards some program with open source. I really don't know.
And finally I don't buy your "friends of Rybka" Are you a friend of cheesburgers you eat? are you a friend of Coca-Cola you drink? are you a friend of the car you drive? I haven't such a friends, Rolf
We have a basic misunderstanding. If I follow the associations you made I cant see Vas Rajlich in the COCA COLA line, since that one is actually in Hamburg, Germany. Therefore I dont get all this hype of the discussion. I didnt ask why in particular Ossipov didnt doctor FRITZ or JUNIOR, I asked for the motivation to chase RYBKA and its author Vas! Everything else makes no sense to me.
The motivation may be because of the fact that Rybka gave wrong output of depth and hided the evaluation of it at small depth so it seems to him a bigger challange to create something that behaves the same as rybka.
Daniel Mehrmann wrote:No, once it is GPL, it will be always GPL. You can change 100% of the code, its still GPL.
This is total nonsense. GPL is a copyright condition. The law specifies what is considered a copy, and what not. Nowhere in the world would something that is 100% different be considered a copy. So it is irrelevant what GPL claims about this case.
Even a 1% identity would be not enough to have the law recognize something as a copy. The fact that Joker contains the line
int main(int argc, char **argv)
which is an exact match with the vast majority of GPLd programs, does not make it an infringement on any copyright.
Daniel Mehrmann wrote:No, once it is GPL, it will be always GPL. You can change 100% of the code, its still GPL.
This is total nonsense. GPL is a copyright condition. The law specifies what is considered a copy, and what not. Nowhere in the world would something that is 100% different be considered a copy. So it is irrelevant what GPL claims about this case.
Even a 1% identity would be not enough to have the law recognize something as a copy. The fact that Joker contains the line
int main(int argc, char **argv)
which is an exact match with the vast majority of GPLd programs, does not make it an infringement on any copyright.
I'm talking in case only if you're using sources which is already under GPL released and i'm not talking about a single function, i'm talking about a whole program. I think you misunderstood my last posting.
However, ask the FSF if you're don't believe me. This is my last posting in this case.
Why should I ask the FSF? Their opinion is worth zilch! They are subject to the law, and if they would make claims that are counter to the law, the claims are void. No matter how often or loud they make them, how wide they would spread them or how eager they would be to tell me about them...
In particular, changing something beyond recognition does not fall under the legal definition of copying. Everything they would say about it is thus null and void.
hgm wrote:Why should I ask the FSF? Their opinion is worth zilch! They are subject to the law, and if they would make claims that are counter to the law, the claims are void. No matter how often or loud they make them, how wide they would spread them or how eager they would be to tell me about them...
In particular, changing something beyond recognition does not fall under the legal definition of copying. Everything they would say about it is thus null and void.
Sorry, i must reply here. This crap can't be standing without a comment.
Maybe you should study court decisions, topic GPL, around the world before you writing these nonsense. Of course there are no court decisions in each country, but for sure there is one for germany and russia.
Are you a friend of cheesburgers you eat? are you a friend of Coca-Cola you drink? are you a friend of the car you drive?
