Vas about IPP engine

[Christopher Conkie]

Jury is still considering if the crashes are caused by Arena, or your buggy compiles.

That would be difficult considering he does not have our compiles.

The latest version is .084, and it does seem to be very stable and bug free, and perhaps 50 elo stronger than 0080a.

If you can read in Russian somehow, you might alter that opinion.

After 30 games at long time controls, it has yet to lose a game. It's evaluation/playing style reminds me very much of Fruit/Toga.

Set the contempt value to 0 for Rybka 3 when it plays it.

Then it will get a thumping.

Christopher

[Anton]

The latest version is .084, and it does seem to be very stable and bug free, and perhaps 50 elo stronger than 0080a.

If you can read in Russian somehow, you might alter that opinion.

Do you have a link, or the text please?

[Christopher Conkie]

To compare what specifically Jury? The code has stayed the same on the site since we found it. Any alterations that we made to make it work or add certain features, were done with the 0080a code.

This is very strange.
The first version had a number 0080a.
In the first version made a lot of changes compared to Rybka 3.
For six months nothing added.

Yes, correct. It is very strange. The only things that were done were an attempt at translation to English and then Italian (0.084)

It is the months before 0.080a that really matter. It would be very interesting to know what 0.001 looked like.

It still does not look written by a human, it looks like a human has amended something created by software..

What do you in your experience of these things think of the code Jury?

Your point of view would be valued on this if you can say something about what you personally see.

Christopher

[Michael Sherwin]

To compare what specifically Jury? The code has stayed the same on the site since we found it. Any alterations that we made to make it work or add certain features, were done with the 0080a code.

This is very strange.
The first version had a number 0080a.
In the first version made a lot of changes compared to Rybka 3.
For six months nothing added.

Yes, correct. It is very strange. The only things that were done were an attempt at translation to English and then Italian (0.084)

It is the months before 0.080a that really matter. It would be very interesting to know what 0.001 looked like.

It still does not look written by a human, it looks like a human has amended something created by software..

What do you in your experience of these things think of the code Jury?

Your point of view would be valued on this if you can say something about what you personally see.

Christopher

And. Does 'b'-code really exist or is it just being used as a ruse to hide a decompile from r3? If 'b'-code does exist can the complete original 'b'-code be obtained? If Ippolit was originally in 'b'-code then why a decompile from r3 to assembler then a rewrite to 'b'-code and a further decompile to C from the 'b'-code. This is beyond incredible. I am overwhelmed!

[Osipov Jury]

I compared the code of Ippolit and Rybka 3. There are a lot of matches. But in general the code very seriously rewrited. I think that this could be done only by author of Rybka. I think that Ippolit - intermediate version of Rybka 4.
About "nonhuman" parts of the code I can not say anything. Rybka 3 do not have these parts.

[Christopher Conkie]

To compare what specifically Jury? The code has stayed the same on the site since we found it. Any alterations that we made to make it work or add certain features, were done with the 0080a code.

This is very strange.
The first version had a number 0080a.
In the first version made a lot of changes compared to Rybka 3.
For six months nothing added.

Yes, correct. It is very strange. The only things that were done were an attempt at translation to English and then Italian (0.084)

It is the months before 0.080a that really matter. It would be very interesting to know what 0.001 looked like.

It still does not look written by a human, it looks like a human has amended something created by software..

What do you in your experience of these things think of the code Jury?

Your point of view would be valued on this if you can say something about what you personally see.

Christopher

And. Does 'b'-code really exist or is it just being used as a ruse to hide a decompile from r3? If 'b'-code does exist can the complete original 'b'-code be obtained? If Ippolit was originally in 'b'-code then why a decompile from r3 to assembler then a rewrite to 'b'-code and a further decompile to C from the 'b'-code. This is beyond incredible. I am overwhelmed!

ь code is pseudocode written in Russian so that "Yakov" can "understand" things better. That is what it says at least. So you would suppose that correlates to what is in the actual source code (even faintly).

There is plenty there about search (forget the eval page) in "ь code" on the website.

Your mission (should you choose to accept it) is to find any ь code (once translated into English and not unreasonable C code) in the actual source code.

I mean the theory behind what is written in the ь code. You could use the values to do a search or look for functions that look (or are) the same.

Maybe we missed something in all the excitement....in May.



Christopher

[Osipov Jury]

And also.
I do not think that Ippolit was obtained by the reverse.
For example, it has a lot of constant expressions. The compiler evaluates such expressions, and the binary contain the numbers. Reset the expression from these numbers is quite difficult.

[Christopher Conkie]

I compared the code of Ippolit and Rybka 3. There are a lot of matches. But in general the code very seriously rewrited. I think that this could be done only by author of Rybka. I think that Ippolit - intermediate version of Rybka 4.
About "nonhuman" parts of the code I can not say anything. Rybka 3 do not have these parts.

And also.
I do not think that Ippolit was obtained by the reverse.
For example, it has a lot of constant expressions. The compiler evaluates such expressions, and the binary contain the numbers. Reset the expression from these numbers is quite difficult.

I put the two posts together, it is easier like that.

Correct me if I am wrong about what you say.

1. There are a lot of matches.

In what way? What do you see that are matches? Too many matches with Rybka so that it icould only be based on Rybka itself (in your opinion)? I know other people have said this as well.

2. The code is seriously rewritten.

Yes, it is. Most if not all of those we spoke to said the same thing, but also they said that they could not imagine a person would write code in this way. Mostly they said "it's a mess Chris".

3. You think that the code could only be made/done by Vasik.

Could you tell us why? I mean, what is it about the code that make you think that. Is it because you see the same code in Rybka?

4. You think that Ippolit - intermediate version of Rybka 4.

It will be very interesting to see what the finished Rybka 4 is like, so we can see if there are similarities between Ippolit and it. Whether or not the conspiracy theory that this is a marketing ploy sticks or not, the fact remains that it would be questionable to give away a source (even partial and mangled) of a commercial product. You can't discount the idea but it is questionable.

5. I do not think that Ippolit was obtained by the reverse.
For example, it has a lot of constant expressions. The compiler evaluates such expressions, and the binary contain the numbers. Reset the expression from these numbers is quite difficult

Some time ago there was a version of Rybka 2 that appeared that did not have symbols stripped. Would that help someone who wanted to do that?

Sorry, if there are a few questions here. It is only about what you think and your opinion. We also asked quite a few people. Adding such opinions (for that is what they are) together helps to gain a good overall picture.

Christopher

[Gian-Carlo Pascutto]

And also.
I do not think that Ippolit was obtained by the reverse.
For example, it has a lot of constant expressions. The compiler evaluates such expressions, and the binary contain the numbers. Reset the expression from these numbers is quite difficult.

The same could be said of the entire reverse engineering effort.

It is interesting to compare the Ippolit source with the RobboLitto source. The amount of refactoring between the two is already tremendous.

With Ippolit it is obvious it was reverse engineered. It is not so obvious in RobboLitto.

But the code is the same.

[mcostalba]

With Ippolit it is obvious it was reverse engineered.

There are a lot of missing field names in the data structures, and also the variable names are "strange" to say the least, but the sources as they are presented seems to me a preprocessed file more then anything else.

Sorry, this is a bit technical for non programmers, but Gian Carlo has understood perfectly what I mean.

The reason of "why" such sources are publicly avaiable remains a mystery, perhaps linked with the smaller but still mistery of why the author of the supposed cloned engine has not said a word until now. I am not talking of what happened months ago, I am talking of these latest days, I don't think he doesn't read this forum, or also if he doesn't almost for sure someone of his people should have pointed out to him this thread.