how far is too far: houdini for sell?

[Graham Banks]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Graham we've been down this road and hearsay is meaningless. No one looked at the source of 3 and 4 and even if they had it would be meaningless as a defence. If they take these engines apart completely it's likely they're not clean regardless.

Why do you people persist? You should be ashamed to defend a liar and a thief like Vas!

Please show me where I've defended Vas since the action or proffered an opinion as to whether the ICGA verdict was right or wrong.

[bob]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Totally false. What was looked for was for a big eval change to figure out where Larry Kaufman's influence entered the equation. Mark found it to be in version 3...

To the best of MY knowledge, no one has done the detailed RE comparison between Rybka 3 and Fruit that was done with the pre-3.0 versions... I am not sure this will be done unless Vas asks for it and provides source to prove that the later ones actually are original...

[Terry McCracken]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Graham we've been down this road and hearsay is meaningless. No one looked at the source of 3 and 4 and even if they had it would be meaningless as a defence. If they take these engines apart completely it's likely they're not clean regardless.

Why do you people persist? You should be ashamed to defend a liar and a thief like Vas!

Please show me where I've defended Vas since the action or proffered an opinion as to whether the ICGA verdict was right or wrong.

You're hopeless. You don't give a damn what the ICGA has ruled.
You won't stop defending this idiot until a court decides. You've told me this countless times. You hate Houdini or so you say but continue to test it along with Rybka and you say you're not taking sides, give me a break.

[bhlangonijr]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Graham, why do you think it would change a thing in this case?
After so many years of refactoring and changing the code it would be completely normal that eventually you won't easily find exact matches when comparing the code of the derivative with the original one. It doesn't mean necessarily the program is "clean", but that it may have changed so much that it would be virtually impossible to compare it with the original program. Take Glaurung 2.1 and Stockfish 2.1.1 for example. I'd bet if you RE both programs you won't find much similarities. But it doesn't change the fact Stockfish is derived from Glaurung, and the name of Tord must keep among the authors...

[bob]

But you know him well enough to suggest he stole code......Bob you think he stole code. So you will use your power as moderator to punish him. Even if this means overriding the other moderators .... Yes or No? Your actions say YES.

Welcome to the Hyattian Oligarchy, this what the majority of CCC members voted for, tyrany!

Cordially,

Sean

Sean,

Bob Hyatt enjoys a level of prestige and honor in the computer chess community, but he does not have nearly the power that people seem to ascribe to him. I laugh every time I hear someone imply that Bob is in charge of everything that happens in computer chess. He deserves honor and respect because of his contributions to computer chess and his longevity but it does not go any farther than that. Not to mention his willingness to share ideas and help others.

However he is no different from any other computer chess author when it comes to decisions made by the ICGA or anyone else. None of us speaks for the others or is specially looked to for guidance or direction. There is no power structure here. The ICGA probably is probably the defacto governing body for computer chess, but even their power is seriously limited - they don't make decisions for other organization.

The kind of talk I keep hearing is the result of an overactive imagination. The scenario being sold is that Bob Hyatt is a Mafia Don with all the chess authors who are his "gumbatta" and he is controlling the decisions of the ICGA. And he has put out a contract against Vas who must have crossed him. And you had better be careful about what you say about him or you are likely to wake up with a horse head in your bed.

I see that previously reasonable people seem to be buying into this at least to a degree.

You just burst my bubble. By the way, until I read your post, last time I saw someone write about me "what a dog" I thought it was a compliment. Apparently I have dyslexia as I though he wrote "what a God".

In any case, people will say strange things in time of stress. I'm old enough to ignore most of it anyway...

[Don]

Trying to find a perfect definition for what a derivative is, is a fools errand. It is always going to come down to human judgement. If such a document is created it will be just as highly subject to argumentation as we are currently experiencing.

And as soon as a decision is rendered that someone does not agree with, the document will come under question as being biased, or wrongly interpreted. It will face the usual criticism for not conforming to the standard of making the decision come out the way someone wanted it to.

The truth is that it would take a really poorly constructed document to exonerate Rybka or Houdini/Robbo, but I don't believe that is really what people here care about. It's just a smokescreen. To me this talk is the same as O.J. Simpson's offer of $500,000.00 reward for information leading to arrest of the "real killer or killers."

[Sean Evans]

Trying to find a perfect definition for what a derivative is, is a fools errand.

Judging by the mounting evidence, the synonym du jour for derivative is "Crafty".

Cordially,

Sean

[rodolfoleoni]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Totally false. What was looked for was for a big eval change to figure out where Larry Kaufman's influence entered the equation. Mark found it to be in version 3...

To the best of MY knowledge, no one has done the detailed RE comparison between Rybka 3 and Fruit that was done with the pre-3.0 versions... I am not sure this will be done unless Vas asks for it and provides source to prove that the later ones actually are original...

These words seem to represent a way out for Vas. "To prove that the later ones actually are ORIGINAL", I mean. Could they really be considered original?

[Graham Banks]

You're hopeless. You don't give a damn what the ICGA has ruled.
You won't stop defending this idiot until a court decides. You've told me this countless times. You hate Houdini or so you say but continue to test it along with Rybka and you say you're not taking sides, give me a break.

Which programs the ICGA decides to exclude from their tournaments is none of my business. I am not a programmer, so I'm not in a position to question their findings.

Yes, I'd like to see an FSF ruling on Rybka because they are seen as being independent of the computer chess scene. I am hopeful that such a ruling would bring some finality to the issue for those who are still uneasy or unhappy what's happened. After all, look at the mess that our hobby is currently in.

Yes, both Rybka and Houdini (along with a few other "controversial" engines marked as such) are in the CCRL rating lists, so I include them in my testing. I can see how you might perceive this to be my defending Vas, but the group's stance is made clear in the notes at the start of each update report that gets posted.

[michiguel]

Let me point out, Rybka versions prior to version 4 used Code from Crafty (rotated bitboard stuff).

How do you know? Up to now only versions up to 2.3.2a have been examined.

Because Vas stated that version 4 moved from rotated bitboards to magic multiply. How easy was that to answer???

Bob - I'd heard that a couple of members of the panel examined Rybka 3 and Rybka 4 and found them clean?

Graham, why do you think it would change a thing in this case?
After so many years of refactoring and changing the code it would be completely normal that eventually you won't easily find exact matches when comparing the code of the derivative with the original one. It doesn't mean necessarily the program is "clean",

If we apply the same criteria to Ferret, Comet, and...
They were not the only ones that started as a different program and everybody considered them clean.

but that it may have changed so much that it would be virtually impossible to compare it with the original program. Take Glaurung 2.1 and Stockfish 2.1.1 for example. I'd bet if you RE both programs you won't find much similarities. But it doesn't change the fact Stockfish is derived from Glaurung, and the name of Tord must keep among the authors...

According to this investigation, and if we follow similar criteria, you really risk that Fabien should be included in Glaurung's list of authors (and SF, of course).

Miguel