A question [about Strelka and Rybka sources]

[GenoM]

Hi

I want to ask a question. Let's experts speculate on it.

If already published by Osipov sources of Strelka are identical to Rybka beta's ones then why Rajlich don't make them public so everyone is assured that this is true? The damage is already done (if we accept the hypotesis of Rajlich that sources of Strelka and Rybka beta are identical).

If the sources aren't identical then Strelka is not a clone of Rybka. Am I right? In this case Rajlich has to stop claiming that Strelka is his own. And the whole talk about "Strelka=clone" has to stop too. And Strelka has to be
acknowledged as independent engine.

Please, stay calm. It's just a question. Share your views.

Regards,
Geno

[mjlef]

Hi

I want to ask a question. Let's experts speculate on it.

If already published by Osipov sources of Strelka are identical to Rybka beta's ones then why Rajlich don't make them public so everyone is assured that this is true? The damage is already done (if we accept the hypotesis of Rajlich that sources of Strelka and Rybka beta are identical).

If the sources aren't identical then Strelka is not a clone of Rybka. Am I right? In this case Rajlich has to stop claiming that Strelka is his own. And the whole talk about "Strelka=clone" has to stop too. And Strelka has to be
acknowledged as independent engine.

Please, stay calm. It's just a question. Share your views.

Regards,
Geno

I suggest we get away from the word "clone". In biology, a clone is something gentically identical (like identical twins). So no one whoe reverese engineers a complex program will ever get it completely identical. The issue to me is if one person took ideas from another, without the original author giving permission. That would be wrong. It is unfortunate, and perhaps foolish that Rybka 1.0 beta was distributed with no license. If it had restrictions on use, then it would be pretty easy to sue others taking the ideas. But as I see it, stealing those ideas is just immoral and slimy.

[LRoberts]

Hi

I want to ask a question. Let's experts speculate on it.

If already published by Osipov sources of Strelka are identical to Rybka beta's ones then why Rajlich don't make them public so everyone is assured that this is true? The damage is already done (if we accept the hypotesis of Rajlich that sources of Strelka and Rybka beta are identical).

If the sources aren't identical then Strelka is not a clone of Rybka. Am I right? In this case Rajlich has to stop claiming that Strelka is his own. And the whole talk about "Strelka=clone" has to stop too. And Strelka has to be
acknowledged as independent engine.

Please, stay calm. It's just a question. Share your views.

Regards,
Geno

No, you're not right. Because if Strelka has taken any of Rybka's code (and I am not pre-judging one way or the other) Strelka might have only taken some of Rybka's code and there may be other elements of Rybka 1.0 beta that have not been made public

[GenoM]

But as I see it, stealing those ideas is just immoral and slimy.

Thanks, I see your point. Someone would argue that misleading people is immoral too. I can imagine that Rajlich decision to hide the real depth and nodes per second made many people suspicious.
Regards,
Geno

PS:
BTW, what would you say about words of Anthony C (Rybka forum http://rybkaforum.net/cgi-bin/rybkaforu ... =3006;pg=3):
"Anyway, in my professional opinion Strelka is basically Fritz5 + history reductions + bigger (but not more) mobility/passed pawn terms + super-vasik-material evaluation."
Isn't it a brand new view at the case?

[GenoM]

Strelka might have only taken some of Rybka's code and there may be other elements of Rybka 1.0 beta that have not been made public

If these (purely hypotetical, of course) unpublished in Strelka sources elements of Rybka are unimportant then the damage is already done. So publicising Rybka beta code wouldn't make more damage than is already done.
If they are important, then why not publish them?
Just speculation, of course. Please, do not understand me wrong -- I'm not interrested in reading Rybka sources. I'm not a programmer.
Regards,
Geno

[mjlef]

BTW, what would you say about words of Anthony C (Rybka forum http://rybkaforum.net/cgi-bin/rybkaforu ... =3006;pg=3):
"Anyway, in my professional opinion Strelka is basically Fritz5 + history reductions + bigger (but not more) mobility/passed pawn terms + super-vasik-material evaluation."
Isn't it a brand new view at the case?

Well, I think Anthony is saying is that Strelka is a fast program (like Fritz5), by "history reductions" I think he means Late Move Reductions (since Strelka does not use history counters), and the rest is right. I would add several search enhancements as well. What everyone is missing is the magic, and hard work it took to come up with the new ideas. Merely using someone's data does not add anything to computer chess. Others have commented that the data in Strelka was taken from Rybka for things like the material imbalances and I assume other evalution terms. The Strelka author did not do the hard work and research to come up with these values. And no one else has as far as I know. Strelka just proves you can reverse engineer a 2 year old program. It does not show how to make a better program, like Vas has done. Although I can admire how skillfully a forger copies the Mona Lisa, the copy will never inspire me like a new work of art.

People have been trying to change the argument saying that Strelka uses a lot of Fruit ideas, and if Strelk is like Rybka, then Rybka used Fruit ideas too. Well Duh! All programmers use published ideas from open source programs. Fruit used many ideas from older programs , and probably every author except Shannon has done the same. We learn from those that came before us. Yes, I am sure Vas used older ideas, but he refined them and made them better and added to them. Has the Strelka author?

Mark

[Dann Corbit]

Hi

I want to ask a question. Let's experts speculate on it.

If already published by Osipov sources of Strelka are identical to Rybka beta's ones then why Rajlich don't make them public so everyone is assured that this is true? The damage is already done (if we accept the hypotesis of Rajlich that sources of Strelka and Rybka beta are identical).

If the sources aren't identical then Strelka is not a clone of Rybka. Am I right? In this case Rajlich has to stop claiming that Strelka is his own. And the whole talk about "Strelka=clone" has to stop too. And Strelka has to be
acknowledged as independent engine.

Please, stay calm. It's just a question. Share your views.

Regards,
Geno

I share his moral outrage, but I do not know if anything illegal has been done or if anything in Strelka legally belongs to Vas.

In the long run, I have no idea what will come of it.

Of course, inevitably, the material imbalance ideas of Larry Kaufman would become committed to code somewhere.

P.S.
"Paradox
THE MAGAZINE OF THE MELBOURNE UNIVERSITY MATHEMATICS AND STATISTICS SOCIETY, Issue 3 2007"

[Dann Corbit]

But as I see it, stealing those ideas is just immoral and slimy.

Thanks, I see your point. Someone would argue that misleading people is immoral too. I can imagine that Rajlich decision to hide the real depth and nodes per second made many people suspicious.
Regards,
Geno

PS:
BTW, what would you say about words of Anthony C (Rybka forum http://rybkaforum.net/cgi-bin/rybkaforu ... =3006;pg=3):
"Anyway, in my professional opinion Strelka is basically Fritz5 + history reductions + bigger (but not more) mobility/passed pawn terms + super-vasik-material evaluation."
Isn't it a brand new view at the case?

I don't think that we should try people based upon opinions.
Judgements should be made on the basis of facts.

[GenoM]

But as I see it, stealing those ideas is just immoral and slimy.

Thanks, I see your point. Someone would argue that misleading people is immoral too. I can imagine that Rajlich decision to hide the real depth and nodes per second made many people suspicious.
Regards,
Geno

PS:
BTW, what would you say about words of Anthony C (Rybka forum http://rybkaforum.net/cgi-bin/rybkaforu ... =3006;pg=3):
"Anyway, in my professional opinion Strelka is basically Fritz5 + history reductions + bigger (but not more) mobility/passed pawn terms + super-vasik-material evaluation."
Isn't it a brand new view at the case?

I don't think that we should try people based upon opinions.
Judgements should be made on the basis of facts.

hi Dann
I was thinking that professional opinion of leading programmer is based on facts, that's why I asked to comment it.
Regards,
Geno

[GenoM]

I just saw I made a mistake in my previous post. So it's edited text:

Strelka might have only taken some of Rybka's code and there may be other elements of Rybka 1.0 beta that have not been made public

If these (purely hypotetical, of course) unpublished in Strelka sources elements of Rybka are unimportant then the damage is already done. So publicising Rybka beta code wouldn't make more damage than is already done.
If they are important, then Strelka is not Rybka and V. Rajlich has not moral right to claim it his own.
Just speculation, of course. Please, do not understand me wrong -- I'm not interrested in reading Rybka sources. I'm not a programmer.
Regards,
Geno