Rybka 1.0 vs. Strelka

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Rolf wrote:
bob wrote:
Rolf wrote:
bob wrote:
Rolf wrote:
bob wrote:You are assuming too much. For example, would you file suit against someone that was making a claim that hurt you, if you_knew_ that the claim was true? Because to file the suit, you have to make a sworn statement that the claim is false in order to seek damages. And if the claim is later proven true, you just committed perjury and are now looking at prison time rather than seeking financial redress from someone else. The sword of justice cuts both ways so caution is required.
I could cut out the other stuff because this here already shows your bias. You simply argue always from the position that Vas has done something wrong. I already wrote a message to write my astonishment how experts could be biased. If you at least would find arguments in favor of Vas, just out of principle. Or also in case you knew how something might have no legal relevance.
Has Vas responded in any way about this? How can one find points in his favor if there is nothing but a deafening silence from his side of the table???
I think he can rely on a purely psychological standpoint for the moment. When did you talk to a commercial programmer collegue during the last 50 years? Now that's going too far. You are looking upon everything like the guy with a hammer. Everything looks to him like a nail while hidden nails frighten him. A commercial programmer cannot discuss what he does, Bob, he lets his program speak. He's in chess what you are on ICC in Bullet. Simply the best!
Sorry, but that boat won't float. When I was accused of cheating several years ago, about something that happened back in 1986. I chose to not sit idly by and let the accusations reverberate around r.g.c.c... I replied factually, quoted a specific letter from David Levy which described the investigation he did and the conclusion that absolutely nothing wrong was done, and so forth. It would be easy enough to simply post "Rybka is my own unique work, I didn't borrow any code form any GPL or open-source programs, so I don't know why this kind of discussion has come up." I can think of only one reason why _I_ would not write that were the discussion about me, I'll leave discovering that reason as an exercise for the reader. I think it is obvious enough that anyone will "get it."
Well, didnt we two discuss more than once the differences between mere lynching justice without fair court trials and a justice after the Roman system? Innocent until proven guilty but NOT guilty until having defended oneself against all kind of campaign like CT&BobH allegations.
What "allegations" have I made? My participation started when someone asked "why would one obfuscate node counts, depths, and PV output?" My response was not any sort of allegation, it was a simple and direct explanation of why that might be done. Then another alternative was provided by CT and it does carry some weight.



It also strikes me that you sit in a boat with people who admittedly have cheated with clones and or output but you cant leave someone innocent his innocense until he was proven guilty.
I can't parse that and have no idea what you mean.
Your newest law of Friday morning is that someone is guilty until he has not asked David Levy ('Satisfying sex with robots', Dissert. Maastricht 2007!) and for all didnt declare that he has never slapped his, sorry, stolen Fruit code nor from Crafty.
I said nothing of the kind. When Berliner first made his claim, David quickly followed up on it. We did not play at the 1986 ACM event because we had just played in the 1986 WCCC and won it. But Harry went because he wanted to attend the conference. David asked him to have Cray restore the same program we used a month or two earlier at the WCCC, and play over some disputed moves in that game. We had already provided the log file showing all analysis, and David/Harry replayed the game to verify that the analysis matched the log, and the actual game. David then wrote a lengthy letter explaining how he evaluated each claim by Berliner, and dismissed each one in succession in several ways, from the replay mentioned above, to "here are other programs that play the same move that you said _no_ computer would play."

As the president of the ICCA, he had an obligation to maintain the integrity of the ICCA, and he investigated thoroughly. I mentioned him, not because of who he was, but because he did a careful analysis of each claim in turn and verified that they were _all_ 100% unfounded. If someone else had done that rather than David, I would have cited their findings instead.


Again, although you attack the actually best programmer worldwide except Alabama as already guilty because he doesnt speak you buddy with the Tiger author who was indirectly accused of having obfuscated the output of his Tiger versions 12, 13, 14. At least he never answered the decent questions of former cooperator Jeroen Noomen. But this is no problem for you. Although CT has obviously cheated his users you still support CT and attack Vas. Furthermore you buddy with someone who was caught red-handed with a clone what he denied for a long time until he confessed his guilt. So you buddy with such famous programmers but you hate to grant Vas the same innocense. Vas is already guilty for you with unknown reasons.

Please reconsider all that, Bob.
I don't see anything to reconsider. He offered to explain how his search worked, and how his node count was somewhat different as a result. I've already posed several "options" for counting nodes. But you notice that the "other side" (your words) never offers any answers. Only counter-claims and counter-questions. Which do nothing to help resolve the questions being asked.
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Enir wrote:
bob wrote:
Rolf wrote:
bob wrote:
Rolf wrote:
bob wrote:You are assuming too much. For example, would you file suit against someone that was making a claim that hurt you, if you_knew_ that the claim was true? Because to file the suit, you have to make a sworn statement that the claim is false in order to seek damages. And if the claim is later proven true, you just committed perjury and are now looking at prison time rather than seeking financial redress from someone else. The sword of justice cuts both ways so caution is required.
I could cut out the other stuff because this here already shows your bias. You simply argue always from the position that Vas has done something wrong. I already wrote a message to write my astonishment how experts could be biased. If you at least would find arguments in favor of Vas, just out of principle. Or also in case you knew how something might have no legal relevance.
Has Vas responded in any way about this? How can one find points in his favor if there is nothing but a deafening silence from his side of the table???
I think he can rely on a purely psychological standpoint for the moment. When did you talk to a commercial programmer collegue during the last 50 years? Now that's going too far. You are looking upon everything like the guy with a hammer. Everything looks to him like a nail while hidden nails frighten him. A commercial programmer cannot discuss what he does, Bob, he lets his program speak. He's in chess what you are on ICC in Bullet. Simply the best!
Sorry, but that boat won't float. When I was accused of cheating several years ago, about something that happened back in 1986. I chose to not sit idly by and let the accusations reverberate around r.g.c.c... I replied factually, quoted a specific letter from David Levy which described the investigation he did and the conclusion that absolutely nothing wrong was done, and so forth. It would be easy enough to simply post "Rybka is my own unique work, I didn't borrow any code form any GPL or open-source programs, so I don't know why this kind of discussion has come up." I can think of only one reason why _I_ would not write that were the discussion about me, I'll leave discovering that reason as an exercise for the reader. I think it is obvious enough that anyone will "get it."
Maybe so. Maybe not. CCC is not ICGA and no one has the obligation, moral or not, to defend from accusations on a forum. It is also possible to look down to them and tell oneself "they bark, so I ride". 2001: been there, done that.

Several people in the last days implicitly accused Vas of copyright breaching. Implicitly. I still have to see an open, unambiguous accusation like "Vas copied Fruit, is a plagiarist and breached Fabien's copyright." Maybe Vas is waiting for this?

Enrique
Have you read this entire thread? At least 2-3 people have made that exact statement, except the issue is GPL'ed code. They have provided source code from both programs showing the very improbable exact matching, and more is coming out every day. Makes you think, for sure.
User avatar
tiger
Posts: 819
Joined: Sat Mar 11, 2006 3:15 am
Location: Guadeloupe (french caribbean island)

Double standards

Post by tiger »

Graham Banks wrote:
chrisw wrote:
Graham Banks wrote:
Tony Thomas wrote:
Graham Banks wrote:
Tony wrote: Uri,

who are you defending ? Vasik or yourself ?

Tony
Apparently Vas isn't under attack here, so why would he need defending?
Vas isnt under attack, only the origin of his program is.
That's the excuse constantly being used - yes.
His integrity and honesty are under attack through the insinuations being made.
Graham,

How do you propose moderation deals with this different to the way it is doing?

You were noderator last time, please inform us which posts you would have deleted out of this thread .......
Boy - that's the million dollar question! :wink:
The problem with CCC is that there's this thing called the charter which unfortunately severely inhibits this sort of discussion.

I would have thought that questioning the legality of an engine falls under the guise of a legally questionable post.

Also, by questioning the legality of an engine in such a manner, one is attacking the honesty and integrity of the engine author, so you're on the fringe with regards to the charter that prohibits personal or libellous attacks.

Unfortunately, if you enforce the charter, such discussions would therefore not be allowed. I'm not saying whether that's a good or bad thing, but that's the fact of the matter.

Hope that explains the predicament that faces the moderators in such situations.

The reason I didn't stand again was so that I no longer had to make such decisions, so I have to leave that to the current team.
Be warned though that you'll cop abuse no matter what you do, although I know that you're already aware of this.

Cheers, Graham.

PS - I think that your post on clones at the top of the page makes the position of the moderators clear.


Questionning the legality of an engine has been done several times in the recent past of CCC. How many times has it happened? 10 times? More?

I have never seen anybody standing up and saying it was against the charter.

Thanks to the discussions, a number of clones or illegal derived works have been discovered.

CCC is the place for these discussions to happen. The moderators' job is to let them happen, but under reasonable control.



// Christophe
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Graham Banks wrote:
trojanfoe wrote:
Rolf wrote: I was asking for the reverse situation. Programmer A has created Fruit code. And he doesnt want to make charges against programmer B (by chance the here heavily attacked Vas, author of Rybka) what is exactly what happened as reported by Dann Corbit who communicated with Fabien Letouzy. Just for your information. In case you needed it. <g>
Oh I see; you are asking if the original author is 'guilty' if he doesn't persue the issue of the GPL infringement. No he's not guilty of anything is he; Use of the GPL by an author doesn't mean he has to persue every infringement he finds, or even encourages him to go looking for infringements.

Again, it doesn't stop others from looking for infringements and discussing them on public forums.

Cheers,
Andy
Trouble is that there's a fine line between discussion and character assassination. :(
Only if there is a question of accuracy and motive. here, the accuracy looks to be pretty convincing, so motive would be irrelevant. Crimes get discussed in the newspaper, on the news and everywhere else _before_ a trial is even held. That is not abnormal anywhere in the world. Trying to squash discussions does not work.
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Anil wrote:
bob wrote:The correct term is "semantically equivalent". That is a common term in discussing student assignments and plagiarism/copying. The probability of even 800 syntactically identical lines is vanishingly small, however 800 lines semantically identical is almost as small. About as likely as two authors writing two different mystery novels, and the storylines are identical, with only the character names and locations being different. It just doesn't happen. And when a couple of chapters are identical, and word-for-word, the probability of "simultaneous invention" is zero.
Taking the example of a programming assignment and its similarity to Fruit vs Strelka / Rybka:

Assuming 4 days for completion of assignment. A smart and intelligent student (Fruit) completes the assignment first (say in 1 day) and now he wants to help his 3 friends (but without allowing them to copy his work). So, he explains his understanding and his algortithm to resolve the problems encountered during completion of his own assignment. Luckily, as he is pretty smart and intelligent, he proved to be a very good teacher and explains his idea to the very last detail.

Now, his friends start working on their own assignments with the first student's algorithm as the base. Also, as they are from the same school/college and have been taught programming by the same teacher, their code is bound to be somewhat similar (as the algorithm that they work on is same). 1-2 friends may have some innate programming talent and optimize their own work.

What do we have now? Assignment by a smart student (Fruit) and so many other similar assignments which do not constitute plagiarism. Its a win-win situation for teacher and the students as the studnets have learned from this experience.
It doesn't work that way. Suppose I explain some architectural topic to a group, say tournament branch prediction. And then ask them to write a one page explanation. Do you think they will be similar in terms of exact sentences written? Same topics covered in the same order in the same paragraphs? Same diagrams?

I can answer that question and the answer is "no". They will each describe the same topic, but their writing will look nothing like each others'. And that is a _far_ smaller amount of writing than what is done in a chess engine.
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Uri Blass wrote:
bob wrote:The correct term is "semantically equivalent". That is a common term in discussing student assignments and plagiarism/copying. The probability of even 800 syntactically identical lines is vanishingly small, however 800 lines semantically identical is almost as small. About as likely as two authors writing two different mystery novels, and the storylines are identical, with only the character names and locations being different. It just doesn't happen. And when a couple of chapters are identical, and word-for-word, the probability of "simultaneous invention" is zero.
I think that things are dependent on the task.

If you have a bitboard program that one bitboard dictate many of the other bitboards so one bitboard that is identical may cause many other lines to be identical.

I do not talk about the case of strelka and fruit here (I did not learn all the similiarities) but talk about the general case.

Uri
There are _too_ many ways to do things. Some have a "FirstOne()" function. Some have a "FIrstOneandClear()" function. Some use magic, some use rotated, some use direct computation for attacks. Some have a mailbox board also, some do not. We are already deep enough to produce 100 different bitboard move generators. And we haven't gotten to search or evaluation. where the chance for divergence goes far higher.

This kind of "matching" does not happen through chance alone.
User avatar
tiger
Posts: 819
Joined: Sat Mar 11, 2006 3:15 am
Location: Guadeloupe (french caribbean island)

Re: Wanted: some opposition to the provided evidence

Post by tiger »

Uri Blass wrote:
Tony wrote:
Graham Banks wrote:
Tony wrote: Uri,

who are you defending ? Vasik or yourself ?

Tony
Apparently Vas isn't under attack here, so why would he need defending?
Sorry, should have been Rybka or Movei

Tony
If I wanted to defend Movei I could say nothing in the first place because movei was not under direct attack.

I simply things that what you do is unfair and Rybka may be more original than part of the commercial programs.

I think that by justice Rybka should not be considered as derivative inspite of the similiarity when toga is clearly a derivative.

If we talk about the law I think that it is unclear otherwise you can make the court to do something against Vas.

Uri


I can't believe for one second that you do not understand the problem.

Toga is more similar to Fruit 2.1 than Rybka 1.0, most probably. But Toga has been released under the GPL, which is perfectly in line with the desire of the author of Fruit.

Rybka 1.0 is maybe a derivative of Fruit 2.1 (this point is still under analysis) and if it is the case then it is an illegal derived work because it does not respect the licence intended by the author of the original work.

Nobody has pointed a gun at Rybka's author forcing him to use GPL code in his program. It's very clear what you are not allowed to do with GPL code: if you want to release a program under a proprietary licence, then just don't use GPL code. No excuse.

Nobody will ever believe either that the author of Rybka was not able to write his own program from scratch without using GPL code. No excuse.

If you think that parts of commercial programs are illegal, then please tell us which programs and provide evidence or at least some pointers so that others can help you prove your point.

But at this time there is only one fish trapped in the net.



// Christophe
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Graham Banks wrote:
Tony wrote: Uri,

who are you defending ? Vasik or yourself ?

Tony
Apparently Vas isn't under attack here, so why would he need defending?
Some of these comments are going _beyond_ stupid. Do you not understand the difference between attacking a person personally, and attacking something he has done as being wrong? It _really_ is a simple concept...
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Wanted: some opposition to the provided evidence

Post by bob »

Uri Blass wrote:
Tony wrote:
Uri Blass wrote:
bob wrote:The correct term is "semantically equivalent". That is a common term in discussing student assignments and plagiarism/copying. The probability of even 800 syntactically identical lines is vanishingly small, however 800 lines semantically identical is almost as small. About as likely as two authors writing two different mystery novels, and the storylines are identical, with only the character names and locations being different. It just doesn't happen. And when a couple of chapters are identical, and word-for-word, the probability of "simultaneous invention" is zero.
I think that things are dependent on the task.

If you have a bitboard program that one bitboard dictate many of the other bitboards so one bitboard that is identical may cause many other lines to be identical.

I do not talk about the case of strelka and fruit here (I did not learn all the similiarities) but talk about the general case.

Uri
Uri,

who are you defending ? Vasik or yourself ?

Tony
I defend the justice.
The probability of
800 lines semantically identical is not something constant but depending on the task that you do.

If you write a code to calculate many simple mathematical functions you can expect easily 800 lines that are semantically identical.

One function may be to calculate if a number is prime or not prime when you can have also many different easy tasks in chess program and
I cannot say speicific number of lines that more than it means that one is derivative of the other.

I also do not think that copying a single line can break the gpl(like people suggest) and I think that some conndition like it is against justice so the condition should be illegal in the first place.

Uri
This is simply wrong. I have been giving assignments for 38 years. Major chunks of identical code do _not_ happen. In simple programs. In compilers and assemblers written by students. In numerical simulations I use regularly. It just doesn't happen, any more than two people will write the same book using the same words, as there are too many ways to express algorithms. Spock's IDIC comes to mind...

We should get off this non-starter topic. Nobody that programs nor teaches programming techniques will buy this explanation any more than they will buy ocean-front properties in Kansas...
User avatar
tiger
Posts: 819
Joined: Sat Mar 11, 2006 3:15 am
Location: Guadeloupe (french caribbean island)

Re: Wanted: some opposition to the provided evidence

Post by tiger »

Uri Blass wrote:<snipped>
trojanfoe wrote: That's not true, as Bob has repeatedly said. The chance that 800 lines of code will be the same from 2 different programmers, even when given the same assignment at the same time, is very very small.

Cheers,
Andy
I think that things are dependent on the assignment.
If the assignment contain 1000 small assignements that everyone of them is 10 lines then it is possible to have 800 lines that are the same without
copying.

Uri


This has nothing to do with the current case and you know it. Your quibble can only confuse people who are not familiar with the field of programming.

I'm really starting to believe that all you want is to divert the attention from the real topic.



// Christophe