1
2
![[a2]](/mwf/v3/flags/a2.png "Satellite Provider")
www.talkchess.com/forum/viewtopic.php?t=37762
come on!!! said no 2dadys source
come on!!! said no 2dadys source
![[us]](/mwf/v3/flags/us.png "United States")
Hopefully Fabien will say more about it. As is these are rather non-committal, vague statements.
We need him to give some opinion on Rybka directly, and if he considers there to be any violation, and if so whether or not he cares.
We need him to give some opinion on Rybka directly, and if he considers there to be any violation, and if so whether or not he cares.
![[za]](/mwf/v3/flags/za.png "South Africa")
Fabien's statements are less vague than any of Vas's on the matter.
From the tone of Fabien's letter I gather he does care.
From the tone of Fabien's letter I gather he does care.
Will this have any effect on the release date of Rybka 4.1?
![[mx]](/mwf/v3/flags/mx.png "Mexico")
No.
> Will this have any effect on the release date of Rybka 4.1?
It shouldn't, unless Vas gets cheesed off and decides to retire from chess programming
![[ro]](/mwf/v3/flags/ro.png "Romania")
Hello Inspectore G ( point ) !
Here's an unique image with Rybka-cluster after Letouzey affair:
Regards,
S
Here's an unique image with Rybka-cluster after Letouzey affair:
Regards,
S
>> Here's an unique image with Rybka-cluster after Letouzey affair:
Yeah, but if you read carefully, Fabien says "I assumed it was some closed-source free engine and that people wanted to know whether it was based on the Fruit source code. The short answer was "no", it was not a verbatim copy of the source code."
But who am I to say anything on the matter? I'm not chess a programmer, though it is one of my plans to write a chess program that I can beat
.
> I assumed it was some closed-source free engine and that people wanted to know whether it was based on the Fruit source code. The short answer was "no", it was not a verbatim copy of the source code."
>
Jawohl !
Fabien says also :
It was however a whole re-write (copy with different words if you like, similar to a translation) of the algorithms. Not just an extraction of a couple of ideas as is common, and normal.
S
Cool
![[it]](/mwf/v3/flags/it.png "Italy")
I don't think that "never" can be postponed further :)
![[de]](/mwf/v3/flags/de.png "Germany")
Mmh, I wonder why Fabien doesn't simply ask Vas.
As far as I know, the 1.0 version of that engine was the Rybka clone.
As far as I know, the 1.0 version of that engine was the Rybka clone.
the problem is that Fabien said that this engine is a bitboard translation of Fruit. Years ago Vasik said that this engine was a clone of Rybka.
So, if A = B and B = C....... A = C ?
So, if A = B and B = C....... A = C ?
That's the zeroth law of thermodynamics.
But that would mean Rybka = Fruit.
No, Rybka != Fruit but Rybka = Fruit+
It's not clear what is the case. One should wait for Vas' answer to that.
![[nz]](/mwf/v3/flags/nz.png "New Zealand")
> It's not clear what is the case. One should wait for Vas' answer to that.
I really do hope that Vas replies this time.
![[no]](/mwf/v3/flags/no.png "Norway")
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=3006
Vas clearly stated that 2.0 was a Rybka clone.
Vas clearly stated that 2.0 was a Rybka clone.
Vas mentioned very specific items that were taken verbatim from Rybka. As anyone who looked at the 2.0 code, there was also a lot copied from Fruit. This does not mean that the parts copied from Rybka and 2.0 were not disjoint (in fact, Vas only mentions things taken from Rybka).
The problem is the word "clone" which leads people to believe that 2.0 would be a copy of Rybka when this is not the case. The description "derivative work" is both more accurate and less likely to lead to these types of misunderstanding. Of course the Rybka forum moderators are largely responsible for creating this confusion by pushing the clone story...
The problem is the word "clone" which leads people to believe that 2.0 would be a copy of Rybka when this is not the case. The description "derivative work" is both more accurate and less likely to lead to these types of misunderstanding. Of course the Rybka forum moderators are largely responsible for creating this confusion by pushing the clone story...
"I do not see obvious signs of other code usage, but perhaps this deserves a closer look. Some of the transplanted ideas, such as the null verification search, are rather naive implementations of the approach in Fruit/Toga, although my first impression is that that code itself is original."
Are you saying that he basically lied when he said that? Especially taken into account that he is very familiar with the Fruit code.
Are you saying that he basically lied when he said that? Especially taken into account that he is very familiar with the Fruit code.
Lied? No. Vas is way too smart to lie about obvious things like this.
A much, much more likely scenario is that in his initial quick review, he only looked at areas that were original to Rybka, found them in 2.0, and put out his clearly erroneous statement. I see no contradiction to the narrative that 2.0 is based on Fruit with the substantive Rybka derived improvements listed by Vas. It is clearly a derivative work of both Fruit and Rybka. Whether it is legal or not is another matter...
A much, much more likely scenario is that in his initial quick review, he only looked at areas that were original to Rybka, found them in 2.0, and put out his clearly erroneous statement. I see no contradiction to the narrative that 2.0 is based on Fruit with the substantive Rybka derived improvements listed by Vas. It is clearly a derivative work of both Fruit and Rybka. Whether it is legal or not is another matter...
Never mind. Apparently he changed his mind later in the thread and his statements seems to contradict each other.
>The description "derivative work" is both more accurate and less likely to lead to these types of misunderstanding.
As you point out, 2.0 is a derivative work; however, there are degrees of derivation. In the case of 2.0, Vas says:
"In light of the above, I am claiming Strelka 2.0 as my own and will release it in the next few days under my own name."
I think this speaks for itself...
![[ru]](/mwf/v3/flags/ru.png "Russian Federation")
Hi Jury,
You are the person who knows the truth as this is your derived engine. Did you mix the code of Fruit and of Rybka to produce 2.0?. I think you are the person who can give us more information
You are the person who knows the truth as this is your derived engine. Did you mix the code of Fruit and of Rybka to produce 2.0?
. I think you are the person who can give us more information
I better be quiet now.
You should write a book. There would be at least a few dozen people who would be lining up to buy it!
I'll give you this book for free and without waiting in line.
Hi Jury,
You are right. I had forgotten that he later retracted the statement. However, after reading both threads, it is clear to me that he never actually changed his mind about the close similarity between the two programs. Rather, it was about what actions he could take legally without getting a court judgment.
All his posts on the subject show a strong belief that the codes of the 2 programs are so close that 2.0 cannot be considered an original work, even if it does not fit the technical definition of a clone.
You are right. I had forgotten that he later retracted the statement. However, after reading both threads, it is clear to me that he never actually changed his mind about the close similarity between the two programs. Rather, it was about what actions he could take legally without getting a court judgment.
All his posts on the subject show a strong belief that the codes of the 2 programs are so close that 2.0 cannot be considered an original work, even if it does not fit the technical definition of a clone.
![[fi]](/mwf/v3/flags/fi.png "Finland")
Ninja'd.
ah ok, i stand corrected.
![[ca]](/mwf/v3/flags/ca.png "Canada")
but now my question is this
What kind of moral rectitude can blame me when I located a publication as "nytimes"? The argument on that occasion was that the clones and their environment, not deserved to appear in such a forum as it is honorable (rybkaforum) but I I meant only to participation in this contest among other "rybka" and "stockfish", which is well known by everyone enjoys a reputation very original, well now maybe only stockfish
Felix Kling said 2 m
"I don't care what other people think and publish. Cloning and piracy will never be allowed in our forum. If you are not able to understand that, go somewhere else. "
I really feel cheated by a cloner that promised me a champion, and as if this were not enough, his vassal, his squire, has the audacity to dispense to a customer those unfortunate words with extreme arrogance
on my land they said "aqui hasta el mas chimuelo masca tuercas" here 2
What kind of moral rectitude can blame me when I located a publication as "nytimes"? The argument on that occasion was that the clones and their environment, not deserved to appear in such a forum as it is honorable (rybkaforum) but I I meant only to participation in this contest among other "rybka" and "stockfish", which is well known by everyone enjoys a reputation very original, well now maybe only stockfish
Felix Kling said 2 m
"I don't care what other people think and publish. Cloning and piracy will never be allowed in our forum. If you are not able to understand that, go somewhere else. "
I really feel cheated by a cloner that promised me a champion, and as if this were not enough, his vassal, his squire, has the audacity to dispense to a customer those unfortunate words with extreme arrogance
on my land they said "aqui hasta el mas chimuelo masca tuercas" here 2
![[gb]](/mwf/v3/flags/gb.png "United Kingdom")
As I stated in the 'Edge' when someone, {I think Graham Banks - I could be wrong so if I am then my apologies Graham} said {and I paraphrase} we only want to test 'original' engines. I followed up with a definition of 'Original' and challenged anyone to tell me who in the programming world has a complete set of code that was created entirely from their sole efforts, i.e., that nowhere in the code would there be any ideas and/or code of others. No one responded.
Perhaps going forward we can put all this to one side as it is extremely likely that very few, if any, engines are original and therefore applying an intuitive stab at whether an engine has or has not taken ideas from someone else seems pointless as can be seen from the fact that the Fruit programmer has revealed what yesterday we all thought was not true, to now not in fact be true.
Going forward we should just keep it simple, if software exists today, whatever its source {claimed or otherwise}, then test it for what it is, i.e., a piece of software that tries to play chess. Keeping the facts separated from the opinions is a much better way to deal with this.
Perhaps going forward we can put all this to one side as it is extremely likely that very few, if any, engines are original and therefore applying an intuitive stab at whether an engine has or has not taken ideas from someone else seems pointless as can be seen from the fact that the Fruit programmer has revealed what yesterday we all thought was not true, to now not in fact be true.
Going forward we should just keep it simple, if software exists today, whatever its source {claimed or otherwise}, then test it for what it is, i.e., a piece of software that tries to play chess. Keeping the facts separated from the opinions is a much better way to deal with this.
In the talkchess post, Fabien states, "Bear in mind that I am mostly unaware of what has happened for five years though."
What has he been doing for the last 5 years? I curious as I find it odd, that after his must acclaimed fruit code was published, he completely left the chess programming community with no interest in the progress of the science and is just coming back now.
What has he been doing for the last 5 years? I curious as I find it odd, that after his must acclaimed fruit code was published, he completely left the chess programming community with no interest in the progress of the science and is just coming back now.
He probably finally made a certain decision...
This will be an interesting debate...
but we know that all(!) are clones of very known algorithms&tweaks(bitboards)... just with little tweaks/piece-value-adjusts and goodies.
Believe me, i know this feeling too well...
but we know that all(!) are clones of very known algorithms&tweaks(bitboards)... just with little tweaks/piece-value-adjusts and goodies.
> he completely left the chess programming community with no interest in the progress of the science and is just coming back now.
Believe me, i know this feeling too well...
In the US, there is absolutely no prohibition to using an algorithm that is not patented, no matter how much the developer may not like it. The bottom line here is that any ideas that were gleaned from reverse engineering Rybka can be legitimately used by other developers. People who can't understand this simple concept invariably end up sounding foolish. They may feel they have the moral high ground, but in reality, they are just modern day Troglodytes. Vas has unwittingly advanced the computer engine field by many years through the reverse engineering of his code. There is no way to get the genie back into the bottle. Vas understands this and has moved on. It would be nice if his supporters did likewise...
Yes, i'm absolutely with you, the whole process of the develope of chess-engines is learning from each other, like in formula one or other ares... and you all know the f1-accueses "the have stolen/copied"... haha
> Vas understands this and has moved on. It would be nice if his supporters did likewise...
Yes it would be good to move on except for one thing; Vas made the claim Ip***** & Ro******* are Ry*** 3 clone and since then, derived chess engines from the 2 engines have been banned, censored, not allowed in tournaments etc.
So the problem is no longer local. The after-effects of the initial accusation/clone-claim is now creating a cascading effect in computer chess community and any further progress in future engines. Moving-on, as good as it sounds, will not just solve & vaporize these problems on its own.
Mr. Letouzy is helping clear things-up in CCC forum. IMO, this approach by Fabien Letouzy is more logical. Especially on the current unresolved 'clone issues & status' and any derived works/engines from them.
Clone has no legal meaning, and outside of cell replication, it doesn't have a clear technical definition either, so I'll leave arguing about whether engine A is a clone of engine B to people who are comfortable arguing about this kind of meaningless stuff.
It is perfectly natural for Vas to be upset about the fact that people have reverse engineered his product and discovered his ideas for building better chess engines. I am 100% sure that if this hadn't happened, Rybka would be at least 150 Elo better than any competitor on equal hardware. It's certainly his prerogative to say that anyone involved in this effort is a dirty SOB.
But if you step back and look around, you will see that this is the normal method that technology advances. For every innovative guy like Vas, there are ten less innovative guys that have other things to offer. Maybe some will build products with fewer bugs, or a better interface. Some may be better at marketing and sales. Having innovators being overtaken and even pushed out by less innovative people or companies is the norm, not the exception.
So now Vas' ideas are out and others are using them. Some would like you to believe that only a small group of upstarts are smart enough to take advantage of these new ideas. Of course, only an imbecile would believe this, but there are no shortage of these types out there. In reality, all good engine designers will incorporate lessons learned from Rybka and use them to strengthen their own engines (just as they have done with ideas from Crafty and Fruit and Stockfish).
In this context, did Mr. Letouzy's recent statement really tell us anything we didn't already know? I really don't think it did. Vas has already stated that he studied everything available when he was first developing Rybka. This explicitly included Fruit. Along with being one of the strongest available engines, Fruit was also far and away the most cleanly written program, so it inspired many to use a lot of its architecture. Did Vas use anything from Fruit in the Beta? It's possible, but do we really need to care about this? It really is time to move on to bigger and better things...
It is perfectly natural for Vas to be upset about the fact that people have reverse engineered his product and discovered his ideas for building better chess engines. I am 100% sure that if this hadn't happened, Rybka would be at least 150 Elo better than any competitor on equal hardware. It's certainly his prerogative to say that anyone involved in this effort is a dirty SOB.
But if you step back and look around, you will see that this is the normal method that technology advances. For every innovative guy like Vas, there are ten less innovative guys that have other things to offer. Maybe some will build products with fewer bugs, or a better interface. Some may be better at marketing and sales. Having innovators being overtaken and even pushed out by less innovative people or companies is the norm, not the exception.
So now Vas' ideas are out and others are using them. Some would like you to believe that only a small group of upstarts are smart enough to take advantage of these new ideas. Of course, only an imbecile would believe this, but there are no shortage of these types out there. In reality, all good engine designers will incorporate lessons learned from Rybka and use them to strengthen their own engines (just as they have done with ideas from Crafty and Fruit and Stockfish).
In this context, did Mr. Letouzy's recent statement really tell us anything we didn't already know? I really don't think it did. Vas has already stated that he studied everything available when he was first developing Rybka. This explicitly included Fruit. Along with being one of the strongest available engines, Fruit was also far and away the most cleanly written program, so it inspired many to use a lot of its architecture. Did Vas use anything from Fruit in the Beta? It's possible, but do we really need to care about this? It really is time to move on to bigger and better things...
In the EU there is clear protection against reverse engineering. You can only reverse engineer and take ideas/code for the purposes of interoperability, something that doesn't apply here.
If Rybka 1 or Fruit was published in the US, the whole thing could be a legal mess with difficulties of juristiction and conflicting law. If (as I assume) Fruit and Rybka 1 were published somewhere in the EU then the law re reverse-engineering is pretty clear-cut, and in practice it is likely to be pretty clear if they were published almost anywhere in Europe.
Anyway, why is reverse enginerring anything to do with the debate about Rybka and Fruit? (sure it applies to Streika and Rybka). Fruit's source was published under GPL. Even US law has recognised that the terms of software licences override fair use exemptions to copyright law, and as such the GPL would apply regardless of whether the source was used as published or reverse engineered out of the executable.
Assuming Fabien's statements to be true, the real legal question is whether tweaking every line in a programme so that is very slightly different but performs the same function gets round the GPL licence. I would be almost certain that a UK court would say it doesn't and would be pretty certain that the court of any major European country would not either.
If Rybka 1 or Fruit was published in the US, the whole thing could be a legal mess with difficulties of juristiction and conflicting law. If (as I assume) Fruit and Rybka 1 were published somewhere in the EU then the law re reverse-engineering is pretty clear-cut, and in practice it is likely to be pretty clear if they were published almost anywhere in Europe.
Anyway, why is reverse enginerring anything to do with the debate about Rybka and Fruit? (sure it applies to Streika and Rybka). Fruit's source was published under GPL. Even US law has recognised that the terms of software licences override fair use exemptions to copyright law, and as such the GPL would apply regardless of whether the source was used as published or reverse engineered out of the executable.
Assuming Fabien's statements to be true, the real legal question is whether tweaking every line in a programme so that is very slightly different but performs the same function gets round the GPL licence. I would be almost certain that a UK court would say it doesn't and would be pretty certain that the court of any major European country would not either.
> In the talkchess post, Fabien states, "Bear in mind that I am mostly unaware of what has happened for five years though."
Mr. Fabien Letouzy also said this, after catching up-to-date http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=389931&t=37762
![[kw]](/mwf/v3/flags/kw.png "Kuwait")
I don't really like this sort of cryptic comments. Why not come out and say clearly what he means instead of running behind some comments that are purposely written in a way that can mean a different thing for different people. If he thinks that Rybka 1.0 beta violated GPL then he should come out and say it ... for some reason he has not done that and doesn't seem to want to do that. He seems incredibly oblivious to all the controversy, and I find that hard to understand that he only recently became aware of it. He should come out and flatly say whether he thinks Rybka 1.0 did violate GPL code, or if he thinks that since the code was re-written then things are OK. Apparently he seems to also be oblivious to the read me file that was included in Rybka 1.0 beta, where Vas made it plenty clear that he did use stuff from open source Fruit.
> where Vas made it plenty clear that he did use stuff from open source Fruit.
Then surely R 1.0 should be open source with a GPL licence?
Well, not if code was re-written ... which it seems that Fabien accepts happened. To be honest I wouldn't know what would constitute a GPL violation, and if re-writing ideas is enough of a change to be considered a non GPL violation. Obviously Vas never thought it was a violation and seemed to think that giving credit in a readme file was enough ethical "cover" of giving the Fruit author some credit in his creation. It is possible that Vas was wrong and there was a GPL violation, but to be honest I find it incredibly difficult to get either side to get to the bottom of this. I guess it must be the "nerd" factor in computer chess and we will just have to live with it!
If I translate the latest, best selling, novel into say French and sell it, do you think I would get away with it?
I am not so sure that analogy stands. Why not this analogy:
You are visiting a neighboring tribe in the bronze age and you happen to witness an effective way of killing wild game by using a long straight piece of wood, which is thrown with some velocity by a hunter. You go back to your village and remember that on a recent trip to another tribe, you witnessed how they were able to hurl large rocks a long distance by using several laminates of wood tied together by animal guts. Bingo ... you make a miniature of the laminated wood and shape it in a curve (bow) then you twist the cat gut that you managed to pull out of your wife's cat (damn cat was always in heat so you solved another problem there) and make a string out of it ... Then in a moment of true inspiration, you find that your miniaturized long straight pieces of wood fly further and straighter and more accurately if you add a few feathers in the back. You have now used several innovative ideas to create your own solution.
You are visiting a neighboring tribe in the bronze age and you happen to witness an effective way of killing wild game by using a long straight piece of wood, which is thrown with some velocity by a hunter. You go back to your village and remember that on a recent trip to another tribe, you witnessed how they were able to hurl large rocks a long distance by using several laminates of wood tied together by animal guts. Bingo ... you make a miniature of the laminated wood and shape it in a curve (bow) then you twist the cat gut that you managed to pull out of your wife's cat (damn cat was always in heat so you solved another problem there) and make a string out of it ... Then in a moment of true inspiration, you find that your miniaturized long straight pieces of wood fly further and straighter and more accurately if you add a few feathers in the back. You have now used several innovative ideas to create your own solution.
Because my analogy makes sense and you are looking for any reason you can find to defend Vas. I can understand that but I think better to stay quiet and see how far fabien is willling to go!?
Unfortunately the thread on CCC has become impossible to navigate, and if I cannot find time to go through it all, I highly doubt that Fabien would want to go through it. From what I have managed to glean from Fabien's posts is that Strelka code had a lot of stuff that was re-written but used same ideas as Fruit. I have yet to see him mention a GPL violation although he seems to be hinting of an ethical problem if Strelka and Rybka 1.0 beta were directly related since he seems to think no credit was given to Fruit (however maybe if he reads the readme file in Rybka 1 beta he would change his mind). Now we are assuming that Strelka code is Rybka 1.0 code ... which it is not. Osipov clearly mentioned that in his disassembly or Rybka 1.0 beta, he was more interested in the LK evaluation tables (which he copied verbatim) and some other inner workings. He seemed to have patched up some other items from Fruit (which was also pointed out by Vas) and the rest from stuff he already had. A lot has been made of Vas's declaration that Strelka is a Rybka 1.0 clone ... but most likely it had to do with code that was copied verbatim from Rybka rather than meaning that Strelka is an identical copy of Rybka 1.0 beta. Anyway, the sane way to deal with this is to have Fabien and Vas post about it in one post. That most likely will never happen and so unfortunately the computer chess community and its forums will be full of crap for the near future.
1
2
Powered by mwForum 2.23.0 © 1999-2010 Markus Wichitill