What do you folks make of this ?

[Chris Whittington]

There's two 'reports' written. One, anonymous or semi-anomymous, by BB and another by the known programmer Zach Wegner. BB's report is written in a relatively neutral style, the reader is left to form his own conclusions, it comes across as a good neutral expert witness style report, difficult to criticise for bias or pre-formed conclusions.

I might point out that, in the infancy of its preparation, my report on Rybka/IPPOLIT was oriented at being disseminated on the Rybka forum, and thus I couched some of it to rebut some of the preconceived opinions commonly seen there. With the publishing of the report at a place where the sentiments run oppositely, I think I would have made a few things more clear. For instance, in Section 1.2, I now add: It should be obvious from a perusal of this document that there are many similarities in ideas, and that the IPPOLIT maker(s) knew the Rybka workings quite well.

Well, in a sense you didn't need to say it because it came across through reading. IIRC I wrote on first and second reading that the top down (ideas) similarities seemed high and even postulated they might be written by the same person

[Chris Whittington]

Feel free to quote this "ad hominem" you refer to. I have no idea what you are talking about, otherwise.

who cares? read your own post at CCC. I'm more interested in this .....


your argument is all very fine and dandy, were it to fit the facts of the case I referred to

your case (a) above (What matters is that (a) the two sets of values are similar or identical) is not met as the two sets of values are very different both in scale and in relation to each other. Since they are not the comparison between the code in Fruit and the lookuptable data in Rybka is completely misleading, he shouldn't have even tabled up the comparison without a huge "THIS DOES NOT PROVE ANYTHING" written next to it. I can see that skim-reading Zach's report will easily lead to the conclusion of rip-off, but actually reading it and cross-checking brings about an entirely different conclusion, namely that this report appears to have been written by the prosecution with one intent in mind. Guilt and conviction.

I agree with you that we need to stay on track, and keep it on topic. That said, I completely disagree with respect to two sets of values being identical data not met. Going on Vas' own words:

Generally, code theft is easy to show - just show the two sections of identical code, side-by-side. There isn't much to debate in such cases.

Rybka 1.0 beta is free. Take a look at the binary starting from address 004092E0 and compare it to Fruit protocol.cpp beginning from line 430 onwards -- there are lines there copied verbatim, line for line. And it does not stop there of course. Not that am breaking new grounds here by any means, but I think the case is pretty much shut. It doesn't mean that we take anything away from his accomplishments. He stood on shoulders of those before him, and it's only right for him to carry a little weight.

was this block the UI code we all argued about a few years ago? Or something else?

Anyway, some more on the ZW expert witness report ... I'll try and summarise, removing the guilt assumptions that litter the report and trying to get to some form of neutral summary, also suitable for non-experts. Please be aware that I am assuming ZW's report to be essentially honest and truthful (if biased).

For brevity later, here are some possible methods of development.

A. Vas never looked at Fruit and wrote Rybka all by himself
B. Vas looked at Fruit, absorbed ideas and then went away and wrote Rybka
C. Vas looked at Fruit, wrote down the ideas, went away and reimplemented them himself according to a check list quite likely plus other ideas plus optimisations on the method of coding. He might, or he might now, have used Fruit evaluation weights as holding values, to be changed later.
D. Vas used Fruit as a test bad. He wrote Rybka as a parallel (meaning alongside) engine and tested sub-sections bit by bit within the Fruit structure, testing that it performed identically to remove bugs. Vas then optimised/tuned/changed the code in various places and optimised/tuned/changed the weights. Final product is then called Rybka. Side by side development followed by optimisation, additon of new ideas etc.
E. Vas started with Fruit and hacked at it bit by bit until he had changed it sufficiently for his purposes and called it Rybka. Development within Fruit environment. More akin to a derivative work, although the final product would be all Vas.
F. Vas has ended up with a product which is so far removed from Fruit (or anything else) that nobody can have any idea or proof of how the development took place - essentially such a product would be quite clean.

Distinguishing between D and E is not easy, especially if the final product has lost all traces of the original. Intermediate versions (Rybka 1beta for example) might contain traces that give away the approach.

1. King_Eval
the data structure used is different
the code is different
the weights that affect the evaluations are different
the ideas are similar although many are quite obvious to use (for example a car has to have some wheels etc.)

Conclusion. This section could be used to suggest use of ideas. But no more. It doesn't exclude D or E, but there's no evidence there to make a case for D or E.
Caveat to ZW.s report. ZW found some code relating to pawn storms in Fruit. He claims this is replaced by a lookuptable in Rybka, and he presents a possible coding to emulate this which is very similar to code in Fruit doing the same thing without lookuptables. Hyatt argues that this is perfectly acceptable as a proof in cases where the lookuptable and the claimed original Fruit code produce identical results. But they don't. The evaluations coming from Fruit's codes and Rybka's lookup are entirely different. That section of the ZW report is misleading.


2. Passed_Pawn
data structure is different
code is different
there's one case where suspiciously identical weights are used, albeit in a different way. The weights are four decimal digits long, so the chance of this being random are low.
ideas are similar but again fairly obvious

Conclusion. This section can be used to suggest similar use of ideas and no more, EXCEPT for the suspicious two numbers data match. That provides evidence for possible use of D or E with a 'forgetting' to optimise/change the particular pair of weights.

3. Piece_Evaluation
code is different
weights are different
data structure is different
ideas are similar

Conclusion. Same as case 1. above.


Overall ......
With the caveat that I didn't look with cross-checking at everything, one can probably say:

Rybka 1beta owes some (indeterminate quantity) ideas to Fruit. These ideas are not new, in that others also do the same, but it looks as if (according to ZW presentation) that the Rybka author was very familiar with the workings of Fruit and implemented some or most or all (indeterminate). Case (A) is probably knocked out.

Rybka 1 beta(in the randomly selected eval segments above) shows major differences in data structure, code and weights, certainly more than enough to describe it as Vas's work and not Fabien's work. If ideas were copied, it was not a crass copying verbatim of code, it was a total re-write, changed data structure, use of lookups, optimisations etc. etc. However, the existence of the suspicious two-number data match precludes giving Rybka 1 beta the clean status F and also probably knocks out status B.

Which leaves us with C, D or E (and anything else I never thought of).

Perhaps more contentiously, I see no copyright infingements, the code and data structure differences are too great. a differentiation between C, D and E are probably impossible at this stage and I suggest that in the last analysis one has to ask "What was the intent of Vas?"

Was it to produce a very strong program that was all his own work, in that it contained code and data that he alone had produced? If he gets there in the end, the how of how he gets there is perhaps not relevent? Rybka 1 beta may have had some residual weights in it (the suspicious values above), are they or anything like it still in R3 or R4? Unlikely.

[CPU]

In the end Vas doesn't care if your happy, he wants your money.

Then why is he issuing a Rybka 4.1 free update for his customers?

Because without it, he will not be receiving our money in the future.

I think that Vas gets a pretty bad rap most of the time, but he's made it pretty clear that his customers are not his top priority, at least not at the moment.

Jeremy

Honestly, as a Rybka customer I get the opposite impression. The customer support from Vas is great IMO.

[Sentinel]

Honestly, as a Rybka customer I get the opposite impression. The customer support from Vas is great IMO.

Sure it is, if you count free Rybka sample to the most faithful believers as the customer support .

[Jeremy Bernstein]

In the end Vas doesn't care if your happy, he wants your money.

Then why is he issuing a Rybka 4.1 free update for his customers?

Because without it, he will not be receiving our money in the future.

I think that Vas gets a pretty bad rap most of the time, but he's made it pretty clear that his customers are not his top priority, at least not at the moment.

Jeremy

Honestly, as a Rybka customer I get the opposite impression. The customer support from Vas is great IMO.

If complete incommunicado for 6+ months, and the non-delivery of a bug-fix update for R3, is great, well, then, I suppose you have a point. If R4+ materializes, I'll be pleased, but he should probably have started working on R4 more than 10 days before the final deadline. Don't get me wrong -- I own Rybkas 3&4 and am happy with them (more with 3 than with 4, until the analysis bugs in 4 are fixed), but I don't consider his recent decisions to be in line with the needs of his customers.

Jeremy

[CPU]

Honestly, as a Rybka customer I get the opposite impression. The customer support from Vas is great IMO.

Sure it is, if you count free Rybka sample to the most faithful believers as the customer support .

There is a free Rybka for all, but no, I was not counting that. Thanks for reminding me, though.

[orgfert]

Feel free to quote this "ad hominem" you refer to. I have no idea what you are talking about, otherwise.

who cares? read your own post at CCC. I'm more interested in this .....


your argument is all very fine and dandy, were it to fit the facts of the case I referred to

your case (a) above (What matters is that (a) the two sets of values are similar or identical) is not met as the two sets of values are very different both in scale and in relation to each other. Since they are not the comparison between the code in Fruit and the lookuptable data in Rybka is completely misleading, he shouldn't have even tabled up the comparison without a huge "THIS DOES NOT PROVE ANYTHING" written next to it. I can see that skim-reading Zach's report will easily lead to the conclusion of rip-off, but actually reading it and cross-checking brings about an entirely different conclusion, namely that this report appears to have been written by the prosecution with one intent in mind. Guilt and conviction.

I agree with you that we need to stay on track, and keep it on topic. That said, I completely disagree with respect to two sets of values being identical data not met. Going on Vas' own words:

Generally, code theft is easy to show - just show the two sections of identical code, side-by-side. There isn't much to debate in such cases.

Rybka 1.0 beta is free. Take a look at the binary starting from address 004092E0 and compare it to Fruit protocol.cpp beginning from line 430 onwards -- there are lines there copied verbatim, line for line. And it does not stop there of course. Not that am breaking new grounds here by any means, but I think the case is pretty much shut. It doesn't mean that we take anything away from his accomplishments. He stood on shoulders of those before him, and it's only right for him to carry a little weight.

was this block the UI code we all argued about a few years ago? Or something else?

Anyway, some more on the ZW expert witness report ...

It is just one example of the clear evidence of verbatim copying. Your post tries to deflect the facts and obfuscate the issue with a lot of words that aren't about the cloned code just illustrated. You must admit you are wrong as everyone else has moved on to the consequences of violating the GPL license.

[Sentinel]

There is a free Rybka for all, but no, I was not counting that. Thanks for reminding me, though.

Really??? Rybka 3 and 4 for free? Where? I hope you are not suggesting piracy...

[CPU]

There is a free Rybka for all, but no, I was not counting that. Thanks for reminding me, though.

Really??? Rybka 3 and 4 for free? Where? I hope you are not suggesting piracy...

No, version 3 or 4 is not free. The free versions are from the Rybka 2 series, but they are pretty strong too. But my favorite free engine is Stockfish! (I really admire Tord & Co for giving us that great engine, but now I'm going slightly off topic here... )

[Chris Whittington]

Feel free to quote this "ad hominem" you refer to. I have no idea what you are talking about, otherwise.

who cares? read your own post at CCC. I'm more interested in this .....


your argument is all very fine and dandy, were it to fit the facts of the case I referred to

your case (a) above (What matters is that (a) the two sets of values are similar or identical) is not met as the two sets of values are very different both in scale and in relation to each other. Since they are not the comparison between the code in Fruit and the lookuptable data in Rybka is completely misleading, he shouldn't have even tabled up the comparison without a huge "THIS DOES NOT PROVE ANYTHING" written next to it. I can see that skim-reading Zach's report will easily lead to the conclusion of rip-off, but actually reading it and cross-checking brings about an entirely different conclusion, namely that this report appears to have been written by the prosecution with one intent in mind. Guilt and conviction.

I agree with you that we need to stay on track, and keep it on topic. That said, I completely disagree with respect to two sets of values being identical data not met. Going on Vas' own words:

Generally, code theft is easy to show - just show the two sections of identical code, side-by-side. There isn't much to debate in such cases.

Rybka 1.0 beta is free. Take a look at the binary starting from address 004092E0 and compare it to Fruit protocol.cpp beginning from line 430 onwards -- there are lines there copied verbatim, line for line. And it does not stop there of course. Not that am breaking new grounds here by any means, but I think the case is pretty much shut. It doesn't mean that we take anything away from his accomplishments. He stood on shoulders of those before him, and it's only right for him to carry a little weight.

was this block the UI code we all argued about a few years ago? Or something else?

Anyway, some more on the ZW expert witness report ...

It is just one example of the clear evidence of verbatim copying. Your post tries to deflect the facts and obfuscate the issue with a lot of words that aren't about the cloned code just illustrated. You must admit you are wrong as everyone else has moved on to the consequences of violating the GPL license.

You confuse me with all the other fans.