Strelka 2.0B + sources

[Tony]

Sorry Uri,

but if you think Strelka only contains small parts of Fruit then you're wrong.

Tony

Sorry Tony,

but if you think Strelka contains any code of Fruit then you're wrong. Ask V.Rajlich He said Strelka contains disassembled code of Rybka, not of Fruit.

Regards,
Geno

My guess is that this is the motivation of the "author" of Strelka.

Tony

[GenoM]

Yes. You're absolutely right.

With Strelka Osipov shows to all of us that Rybka beta was re-written Fruit + added re-written tables of material disbalance of Larry Kaufman.

[Daniel Mehrmann]

Yes. You're absolutely right.

With Strelka Osipov shows to all of us that Rybka beta was re-written Fruit + added re-written tables of material disbalance of Larry Kaufman.

You should be very very carefully writing such stuff in public. Maybe it's possible, but so long it's not proven i wouldn't write it.

If you ask me i would say i don't believe so.

And by the way a lot of engines are a little bit fruity. That's the computerchess of today.

Borrow ideas of Fruit isn't a crime.

Best,
Daniel

[Thomas Gaksch]

Hi Daniel,
that is a very unfair statement from you. Dann (and the others) had only seen the old Strelka source code. So this code could be different.
By the way this is again a very strange discussion. Concerning Fruit i think Strelka is legal. Yes it is true that there are a lot of ideas copied from Fruit but not the CODE DIRECTLY. As far as i understand the GPL it is not wrong to do this.

And what do you think how many engine authors have fruit ideas in their closed sources? The most engine authors have studied the fruit code very accurate and have implemented these ideas. And thats absolutely ok! Thats open source.

But of course it is not correct that the Strelka author has copied the data tables from rybka. thats illegal.

Thomas

[GenoM]

Yes. You're absolutely right.

With Strelka Osipov shows to all of us that Rybka beta was re-written Fruit + added re-written tables of material disbalance of Larry Kaufman.

You should be very very carefully writing such stuff in public.

Hi Daniel.

When I'm saying in public that 2+2=4, I am perfectly calm.

Regards,
Geno

[Uri Blass]

From one side we have opinions here at the CCC that Strelka is an obvious "translation" of Fruit, or that it is at least "very, very Fruit like".

On the other hand Vas has posted on the Rybka forum that Strelka is Rybka but he does not "see obvious signs of other code usage" so he claims Strelka as his own!

Surely there is a contradiction between those views? My head is now spinning...

Well, i believe that strelka has a lot of fruit code inside. I don't know if it was created from a full fruit, but the "author" claimed it.

However, you can see it very easy by just looking on the coding style. How things are written, e.g case sensitive or using"_" or only big letters and so on.
You will find a lot of parts in the sourcecode which doesn't fit in the style of the complete source...

Anyway, Dann Corbit is a very very big disapointment for me and at least he should see that what i explained above. He should no longer be asked in clone cases....sorry

Best,
Daniel

You may also consider Tord as disappointment for you because Tord did not consider strelka to be a clone.

I prefer to assume that a person is not quilty in case of a doubt and some similiarity does not prove that a person is quilty.

I think that strelka has only a small part of fruit code and small part does not prove cloning(otherwise based on the same logic every 2 programs that have i++; inside their code can be considered as clones)

Uri

Well, let me tell you some things about coding. I don't know if you're working as a professional progammer like me over years in IT/software houses. I worked in big projects (some hundreds of programmer together) and now in a little project (just 30 people). You'll get experince of code style. You'll learn that every programmer has a own signature of coding.

Specialy if it a private one person project, like chessprogramming, the signature is pretty clear. Basicly you're looking to other code and try to understand and write it down in your own signature of code.
In the case of Strelka i found serval lines of code which isn't written in the signature of the other parts.

Back to your arguments i just can repeat that the author claimed he was starting from a full fruit and if i see the code i would agree with him, because as i pointed out above you would write it in your own sginature of code style instead of copying the stuff. That helps you later on to understand your own program better. If you would using other styles and just start to mix your code you'll get a dirty picture.
I know that programmers, this is based on my experince of my work as professional programmer, would avoid this case if it possible anyway.

And keepout Tord please. This is just a dirty try.

Best,
Daniel

Strelka has clearly less files than fruit so starting from the full source of fruit does not seem logical to me.

It seems that it is less work to start from your own code based on understanding the ideas of fruit when it is possible that small parts of fruit were copied in the process but no big parts.

Uri

[Tony]

No.

One of the biggest things of Fruit is its stability. Starting with this as a basis is a serious improvement over starting from scratch.

Tony

[Uri Blass]

From one side we have opinions here at the CCC that Strelka is an obvious "translation" of Fruit, or that it is at least "very, very Fruit like".

On the other hand Vas has posted on the Rybka forum that Strelka is Rybka but he does not "see obvious signs of other code usage" so he claims Strelka as his own!

Surely there is a contradiction between those views? My head is now spinning...

Well, i believe that strelka has a lot of fruit code inside. I don't know if it was created from a full fruit, but the "author" claimed it.

However, you can see it very easy by just looking on the coding style. How things are written, e.g case sensitive or using"_" or only big letters and so on.
You will find a lot of parts in the sourcecode which doesn't fit in the style of the complete source...

Anyway, Dann Corbit is a very very big disapointment for me and at least he should see that what i explained above. He should no longer be asked in clone cases....sorry

Best,
Daniel

You may also consider Tord as disappointment for you because Tord did not consider strelka to be a clone.

I prefer to assume that a person is not quilty in case of a doubt and some similiarity does not prove that a person is quilty.

I think that strelka has only a small part of fruit code and small part does not prove cloning(otherwise based on the same logic every 2 programs that have i++; inside their code can be considered as clones)

Uri

Sorry Uri,

but if you think Strelka only contains small parts of Fruit then you're wrong.

It seems you ( and others) are not looking at the functionality of the code, but only at the looks.

The Fruit way

Code: Select all

mob=0

mob+= AttackValue[from-31];
mob+= AttackValue[from-33];
mob+= AttackValue[from-14];
mob+= AttackValue[from-18];

etc

score+=mob*376

is off coarse exactly the same as Strelka's

Code: Select all

BB=AttacksKnight(from) & ~myPieces

mob=PopCount(BB);

score+=mob*376

It just looks different.

If you don't see this, it doesn't matter, just don't bother with your opinion in the clone debate.

If someone steals your bike and paints it, it's still your bike.

The only problem that arises here is that the bike seems to have been stolen before ...

Tony

I simply disagree with you about definition of stealing.
ideas of fruit are not protected when code is protected.

If you translate ideas of evaluation to different structure then I do not consider it as cloning.

Note that based on my memory the constants are in most cases not the same as fruit and the idea to calculate knight mobility in that way is an idea that can be used by everybody.

Uri

[Uri Blass]

No.

One of the biggest things of Fruit is its stability. Starting with this as a basis is a serious improvement over starting from scratch.

Tony

Fruit for me has too many files and I feel that if I try to start from it I cannot find nothing and do not know what to change.

Understanding fruit should be the first step and if you really understand fruit(something that I never did except very small parts) then I think that starting from scratch is easier than changing fruit.

Uri

[Tony]

..

If you translate ideas of evaluation to different structure then I do not consider it as cloning.


Uri

That's the point, I guess. I do. There is just too many different ways to do exactly the same.

Mind you, I'm not talking about "almost doing the same".

If one person loops from 0 to 10 and the other from 10 to 0, it can still be exactly the same. ( Though not literally the same)

I guess for me it's about the design decisions that are taken, not about the implementation.

Exactly the same decisions with a different implementation, would still be a clone for me. ( So yes, I consider reverse enginering, cloning and therefore stealing )

For me, this is common sense, regardless of what laws are saying.

Tony