Rybka disqualified and banned from World Computer Chess Cham

[Laskos]

And certainly if Loop gets awarded anything that would be a very bad joke.

Based on what is reported here, it seems that after Rybka's disqualification, Loop is now considered to be in 2nd place at the 2007 WCCC ?

I guess it remains to see whether that will stand, since Loop appears to contain a Fruit-derived eval...

Can you prove that?

I believe I saw Adam Hair's similarity results, based purely on output, which seem to indicate beyond any doubt that Loop is a Strelka (if I remember well).

Kai

[bob]

Mr Hyatt.

You are a scientist and i expected a different approach.

For example You prove A using a1,a2,a3 math arguments to do so.

If i take those a1,a2,a3 add b1,b2,b3 to prove A+ which is a better and bigger generalization of Statement A, what does that mean?

You can say that is derived. But everything is derived from something else.

I mean. I start from crafty. I start changinging code to make it a different engine. I am changing everything. Search, evaluation, pruning everthing.
At what point does the transformed code is NOT crafty anymore?

At the point where there is no copied code left, for starters. Although one can discuss the idea of taking a C program and translating it to some other language, such as (heaven forbid) Cobol. Then there would be no remaining code, but the programs would be semantically identical.

If 2 programs share 5% of code are they "linked"?

Yes.

I do not think someone can compare Fruit and Rybka 3 as engines. They may share some code parts, but i am sure those parts were considered insignificant and thus where not changed. (i am sure everyone starts coding his chess program that way until you have a program up and running)

I am sure everyone doesn't. I know I didn't as there was no program to copy in 1968 except for MacHack written in PDP-10 assembly language...

[oreopoulos]

I am sure everyone doesn't. I know I didn't as there was no program to copy in 1968 except for MacHack written in PDP-10 assembly language...

Mr Hyatt, i am sure no one can argue that you are a pioneer in the field and you where "obliged" to start from scratch, but i think your point of view is somewhat "too american"

Science is NOT copyrighted. You create tools. I use them to go forward. You use code someone else wrote.

There is

a) % of code that is similar
b) weight of significance of that code

Maybe 20 years ago, movegeneration algorithm was important for an engine performance. Nowadays, can you claim something like that??

I think this decision is bad for chess engines and not for Rybka alone.

[wgarvin]

Maybe 20 years ago, movegeneration algorithm was important for an engine performance. Nowadays, can you claim something like that??

I think this decision is bad for chess engines and not for Rybka alone.

Its only bad for chess engines if their authors want to enter them into tournaments such as the WCCC. If you want to enter the tournament, you have to write the program yourself. You can't copy code from others and then pretend it was all your own work. Pretty simple, really.


The issue of licensing is completely separate from the issue of ICGA tournament rules. If you want to sell a commercial engine (or even distribute an open-source one), you have to either avoid using other people's code, OR simply respect whatever license agreement they have attached to their code.

Copyright law in most countries, protects the rights of whoever owns the copyright to that code (for chess engines, that is usually the author). If you want to use someone else's copyrighted code in your engine, you need their permission. This permission usually comes in the form of a license. As long as you follow the conditions of the license, you can use it. If you break the license conditions, then you lose your permission and you are now infringing the owner's copyright if you distribute their code (or a compiled engine containing their code, or some code derived from their code, or a compiled engine containing some code derived from their code...)

Rybka is a commercial program that (at least in some versions) apparently contained code derived from Fruit 2.1. Rajlich did not follow the terms of the GPL license of Fruit 2.1, so he may be liable for copyright infringement because he sold a derived work based on that code.

[oreopoulos]

There is copyrighted code and algorithm.

What part of the code is "copyrighted"? You have a loop for example that sorts an array. If i copy paste it, is that copyrighted code?

Is a loop that prints an array "copyrighted code"?

I mean as i see it it is very difficult to define and identify copyrighted code.
Because behind code there is an algorithm and that is what is important. Implementation is important when the complexity of the algorithm increases

I cannot accept , that if you copy paste 3 lines of code, from a "copyrighted code" , that is violation of copyright. Lines of code are not out of the context.

Its another story stealing someones code, make some changes and claiming its yours , and something else, starting from someone elses code, change 95% of the code and leave 5% of things that really do not matter.

[Terry McCracken]

I am sure everyone doesn't. I know I didn't as there was no program to copy in 1968 except for MacHack written in PDP-10 assembly language...

Mr Hyatt, i am sure no one can argue that you are a pioneer in the field and you where "obliged" to start from scratch, but i think your point of view is somewhat "too american"

Science is NOT copyrighted. You create tools. I use them to go forward. You use code someone else wrote.

There is

a) % of code that is similar
b) weight of significance of that code

Maybe 20 years ago, movegeneration algorithm was important for an engine performance. Nowadays, can you claim something like that??

I think this decision is bad for chess engines and not for Rybka alone.

Chris is that you? Mr. Too American..... Regardless this is a nonsensical argument.

No not Chris....wrong Jack-in-the-Box.

I hate surprises.

[noctiferus]

<(heaven forbid) Cobol>
Robert, sorry for outposting, but you reminded some troubles in my youth (30 years ago).
We were consulting to an important car factory. We had to implement MIL STD105 statistical plans for quality control on some parts, and had to do it in RPGII .
I still remember it as a nightmare

[bob]

I am sure everyone doesn't. I know I didn't as there was no program to copy in 1968 except for MacHack written in PDP-10 assembly language...

Mr Hyatt, i am sure no one can argue that you are a pioneer in the field and you where "obliged" to start from scratch, but i think your point of view is somewhat "too american"

Science is NOT copyrighted. You create tools. I use them to go forward. You use code someone else wrote.

There is

a) % of code that is similar
b) weight of significance of that code

Maybe 20 years ago, movegeneration algorithm was important for an engine performance. Nowadays, can you claim something like that??

I think this decision is bad for chess engines and not for Rybka alone.

In your world, books would not be worth reading since they would all be very similar. Chess tournaments would simply be a random coin-toss with everybody using the strongest robo* version they could find. Etc. When I race, I want to use _my_ legs, or a car that _I_ built myself. Then I am an active participant in the race. That is what makes the competition fun...

[bob]

There is copyrighted code and algorithm.

What part of the code is "copyrighted"? You have a loop for example that sorts an array. If i copy paste it, is that copyrighted code?

Is a loop that prints an array "copyrighted code"?

I mean as i see it it is very difficult to define and identify copyrighted code.
Because behind code there is an algorithm and that is what is important. Implementation is important when the complexity of the algorithm increases

I cannot accept , that if you copy paste 3 lines of code, from a "copyrighted code" , that is violation of copyright. Lines of code are not out of the context.

Its another story stealing someones code, make some changes and claiming its yours , and something else, starting from someone elses code, change 95% of the code and leave 5% of things that really do not matter.

Whatever you copy is a violation. Hence the term "copyright". I don't think anyone raises a red flag if you copy a random number generator, or a sort, or a tablebase access procedure, so long as the sources are attributed. Here we are talking about copying _much_ more than that...

[bob]

<(heaven forbid) Cobol>
Robert, sorry for outposting, but you reminded some troubles in my youth (30 years ago).
We were consulting to an important car factory. We had to implement MIL STD105 statistical plans for quality control on some parts, and had to do it in RPGII .
I still remember it as a nightmare

report program generator -- another nightmare for a real programmer.