Strelka -- clone status

[Daniel Mehrmann]

I think we do not know what the author claimed and maybe the person who made the claims is not the author of the program.

Assume you're right and a another person claimed wrong things in the name of the author of Strelka, do you really believe the author of Strelka wouldn't correct it ??

The author of Strelka has now so many contacts to other programmers. If there would be something wrong with this statement, he would correct it for sure. But months ago now nothing happend yet. So, you're greenly if you still believe on it.

Suppose tord claims tomorrow that he started from a fruit code base .
Is it enough to claim that Glaurung is a fruit clone without looking at the source.

I have no problem with it and Tord is wellcome to work on fruit anyway, because Tord is fighting for open source and he would follows his manner. Uri, that is the most bad example case you could ever choose here.

Best,
Daniel

[Uri Blass]

I think we do not know what the author claimed and maybe the person who made the claims is not the author of the program.

Assume you're right and a another person claimed wrong things in the name of the author of Strelka, do you really believe the author of Strelka wouldn't correct it ??

The author of Strelka has now so many contacts to other programmers. If there would be something wrong with this statement, he would correct it for sure. But months ago now nothing happend yet. So, you're greenly if you still believe on it.

Suppose tord claims tomorrow that he started from a fruit code base .
Is it enough to claim that Glaurung is a fruit clone without looking at the source.

I have no problem with it and Tord is wellcome to work on fruit anyway, because Tord is fighting for open source and he would follows his manner. Uri, that is the most bad example case you could ever choose here.

Best,
Daniel

My point was not that Tord is not free to work on fruit but that people are not supposed to believe everything that they read.

If somebody claims that code X is based on code Y when people who saw both codes have a different opinion then I will not say that it is obvious that X is based on Y.

Recently I try to work on movei in order to improve some pruning code so I have no time for strelka but I plan to go back to strelka later.

Uri

[Dann Corbit]

All of the programmers who examined Strelka are relatively familiar with Fruit. But the real expert and the one who's opinion matters most is Fabian. Fabian is making a careful examination now. I think it is his opinion that will matter more than anyone else's.
I think that the best course to take at this point is to see what those authors have to say.

No.

People have suspected it to be a Rybka clone(or an algorithm theft, w/e), Fruit issues was bought later as more suspicion aroused.

Therefore, the only person/programmer whom opinion matters, is Vasik Rajlich, coder of Rybka, who is also the only person with access to it's code.

Other programmer's opinions is useless.

And considering Strelka's sources are already in too many hands.. I very much doubt if he would like to do that.

So the gloomy identity of Strelka's sources shall remain well concealed for the time being, and quite honestly - no one should care, as already the offspring strength lags an epoch in front his big brother's (whoever that might be..) shade

Strelka 1.8 2866 / Rybka 2.3.2a 32-bit 3013

Because the sources for Rybka have never been made public, this claim is literally impossible.

It is possible to study the algorithms by stepping through in assembly.
The Strelka code has no assembly.
Therefore, at a bare minimum, the Strelka author had to rewrite the algorithms in C or C++.
Hence, he has not produced an illegal strelka clone. It is literally impossible for him to have done so.

On the other hand, there is a question about large tables of numbers. It is possible that this could constitute a copyright violation (though I am not sure about it).

But an illegal program clone without source code is literally impossible unless the clone program itself were entirely assembly.

This claim (the clone status) has not merit whatsoever.

Again, copyright infringement may be possible, but I am not expert enough to know (even if the numbers are identical).

[Ryan Benitez]

His name is Fabien. Anyway I hope this puts some closure to this topic so everyone can move on. For now I maintain my original stance about Strelka that it is a very well done study released in a dubious way.

*sign*

I'm just want to add on that it looks like to me everybody is ignoring what the author of Strelka claimed. He said that he started from a fruit code base and that's enought to call it a clone in my eyes (don't forget the GPL violation) or at least a dirty start developing a own chessengine.

There is no need to prove something.
There is no need to explain something.

However, i wrote so much about this topic and now it's just annoying. I hope the end of this story comes fast.
But one thing is pretty clear for me, the computerchess community itself got and get a big image damage anyway.

Best,
Daniel

In most cases I would say you are absolutely right. In this case I assume that the author could not communicate what he wanted to say because he did not speak a western language such and French, German, English, or Spanish that others could easily understand. I offered to verify the code was not derived from Fruit but received no response. Fabien is a better source to confirm this assuming he has time to do so of course.

[Dann Corbit]

All of the programmers who examined Strelka are relatively familiar with Fruit. But the real expert and the one who's opinion matters most is Fabian. Fabian is making a careful examination now. I think it is his opinion that will matter more than anyone else's.
I think that the best course to take at this point is to see what those authors have to say.

No.

People have suspected it to be a Rybka clone(or an algorithm theft, w/e), Fruit issues was bought later as more suspicion aroused.

Therefore, the only person/programmer whom opinion matters, is Vasik Rajlich, coder of Rybka, who is also the only person with access to it's code.

Other programmer's opinions is useless.

And considering Strelka's sources are already in too many hands.. I very much doubt if he would like to do that.

So the gloomy identity of Strelka's sources shall remain well concealed for the time being, and quite honestly - no one should care, as already the offspring strength lags an epoch in front his big brother's (whoever that might be..) shade

Strelka 1.8 2866 / Rybka 2.3.2a 32-bit 3013

This is just weird.

Nobody but Vas has Rybka's source, and so a non-assembly clone of Rybka is literally impossible.

[mjlef]

Not true. There are commercial tools that generate C code from assembly sources.

[Nid Hogge]

All of the programmers who examined Strelka are relatively familiar with Fruit. But the real expert and the one who's opinion matters most is Fabian. Fabian is making a careful examination now. I think it is his opinion that will matter more than anyone else's.
I think that the best course to take at this point is to see what those authors have to say.

No.

People have suspected it to be a Rybka clone(or an algorithm theft, w/e), Fruit issues was bought later as more suspicion aroused.

--snip--

Because the sources for Rybka have never been made public, this claim is literally impossible.

It is possible to study the algorithms by stepping through in assembly.
The Strelka code has no assembly.
Therefore, at a bare minimum, the Strelka author had to rewrite the algorithms in C or C++.
Hence, he has not produced an illegal strelka clone. It is literally impossible for him to have done so.

On the other hand, there is a question about large tables of numbers. It is possible that this could constitute a copyright violation (though I am not sure about it).

But an illegal program clone without source code is literally impossible unless the clone program itself were entirely assembly.

This claim (the clone status) has not merit whatsoever.

Again, copyright infringement may be possible, but I am not expert enough to know (even if the numbers are identical).

This is just weird.

Nobody but Vas has Rybka's source, and so a non-assembly clone of Rybka is literally impossible.

Our definition of "Clone" differs. But I've said it once and I'll say it again:

Copying source code from prog X ======= Decompling prog X and stealing it's ideas. Just as bad.

Whether he backtraced 1 piece of code or a thousand, the outcome is equal.

I don't know about the whole legal issue. But I honestly don't care.
It's just plain wrong.
The semantics are irrelevant. It's a theft.

Not true. There are commercial tools that generate C code from assembly sources.

Yep.
I don't care how insufferable a task like this would be. I just don't rule such a thing out.

[Dann Corbit]

Not true. There are commercial tools that generate C code from assembly sources.

They don't work.

The link to the commercial tool you found is probably a company that went out of business.

There are some tools in sourceforge for that purpose, but they don't work either.

Something as large as a chess program is not going be be reverse engineered.

.NET and Java (on the other hand) can easily be turned back into source from CLR. Even at that, it seems grey to me if the algorithm encoded in CLR is protected, since the fundamental algorithm has no copy protection.

On the other hand, with conversion of CLR back to .NET or Java classes back into Java, even if it were completely legal, it seems underhanded.

[Dann Corbit]

All of the programmers who examined Strelka are relatively familiar with Fruit. But the real expert and the one who's opinion matters most is Fabian. Fabian is making a careful examination now. I think it is his opinion that will matter more than anyone else's.
I think that the best course to take at this point is to see what those authors have to say.

No.

People have suspected it to be a Rybka clone(or an algorithm theft, w/e), Fruit issues was bought later as more suspicion aroused.

--snip--

Because the sources for Rybka have never been made public, this claim is literally impossible.

It is possible to study the algorithms by stepping through in assembly.
The Strelka code has no assembly.
Therefore, at a bare minimum, the Strelka author had to rewrite the algorithms in C or C++.
Hence, he has not produced an illegal strelka clone. It is literally impossible for him to have done so.

On the other hand, there is a question about large tables of numbers. It is possible that this could constitute a copyright violation (though I am not sure about it).

But an illegal program clone without source code is literally impossible unless the clone program itself were entirely assembly.

This claim (the clone status) has not merit whatsoever.

Again, copyright infringement may be possible, but I am not expert enough to know (even if the numbers are identical).

This is just weird.

Nobody but Vas has Rybka's source, and so a non-assembly clone of Rybka is literally impossible.

Our definition of "Clone" differs. But I've said it once and I'll say it again:

Copying source code from prog X ======= Decompling prog X and stealing it's ideas. Just as bad.

Whether he backtraced 1 piece of code or a thousand, the outcome is equal.

I don't know about the whole legal issue. But I honestly don't care.
It's just plain wrong.
The semantics are irrelevant. It's a theft.

Not true. There are commercial tools that generate C code from assembly sources.

Yep.
I don't care how insufferable a task like this would be. I just don't rule such a thing out.

It's not legally a theft to trace through an algorithm, figure out what it does and then do it yourself.

I also do not think that it is as cut and dried as you imagine. Imagine (for instance) that mathematics is kept from people because money can be made from the mathematics. Would that seem good to you?

Imagine if C.A.R. Hoare had never published the quick sort algorithm would that seem good to you?

Reverse engineering for the purpose of discovery does not seem bad to me, in the same way that it does to you.

The notion of turning assembly code back into C is a pure fantasy.

While on the one hand, it is provably impossible to do it, I do realize that in a pragmatic sense people have managed little snippets from particular compilers using custom tools that they have built. Personally, I think it would be a lot easier to just trace through the assembly to understand what it is doing.

That's on the one hand. But if someone does gather information using reverse engineering, I think there is a clear obligation to divulge that. Otherwise, it seems like intellectual dishonesty to me. (Really, it's a form of plagiarism).

[Ryan Benitez]

Not true. There are commercial tools that generate C code from assembly sources.

They don't work.

The link to the commercial tool you found is probably a company that went out of business.

There are some tools in sourceforge for that purpose, but they don't work either.

Something as large as a chess program is not going be be reverse engineered.

.NET and Java (on the other hand) can easily be turned back into source from CLR. Even at that, it seems grey to me if the algorithm encoded in CLR is protected, since the fundamental algorithm has no copy protection.

On the other hand, with conversion of CLR back to .NET or Java classes back into Java, even if it were completely legal, it seems underhanded.

I have never seen a tool that turns an executable into useful c code. I saw a tool that made ugly c mixed with assembly many years ago but I can’t seem to find it any more. As for reverse engineering a chess program, this is very possible only it is tedious and ugly. I am sure that in the time it would take to reverse engineer a chess program you could make a far better one in the same time spent. I am sure most programmers have done at least a little poking around in Rybka though.