Should downloadlinks for Ippo/Robbo be allowed again?

[gerold]

Is the separate search function the reason it has a time problem.

Best.

Gerold.

[Matthias Gemuh]

Bottom line is that the evidence for cloning is rather thin and the evidence for a lot of originality is rather high.

I agree 100%.

Matthias.

[michiguel]

Even without any evidence I am sure that Ippo/Robbo is clone: why publish superstrong engine without getting any money or at least fame for
your real name? Additionally if your first engine is 3100 rated it is not
credible at all!

Jouni

Someone that believes in a Marxist philosophy would do what the author of Ippolit--according to the Ippolit website--has allegedly done. It makes perfect sense. As far as the strength is concerned, well that is not proof.

Evidence:

1.) Vas says that it is a Rybka 3 clone.

2.) Jury says that it is not Rybka 3 but rather it is Rybka 4.

1 and 2 are mutually exclusive.

3.) There is strong evidence that it was developed from many sources and MAY have some reversed engineered ideas from Rybka.

1 and 2 are greatly discounted by 3.

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

5.) There are some differences and some similarities in its play compared to Rybka, but the differences are much more prominent.

...

How many versions of Fruit was there before it became the strongest program in the world 1.0, 1.5, 2.0, 2.1? What if there were several versions of Ippolit, in "chess language b", before it was translated into C to make it available to the west?

Bottom line is that the evidence for cloning is rather thin and the evidence for a lot of originality is rather high. So, unless someone that is trusted disassembles Rybka 3 and finds a nearly 1 to 1 correlation in search functions, evaluation functions or evaluation constants then I for one choose to err on the side of innocent until proven guilty rather than joining in on a witch hunt.

Miguel

[Gian-Carlo Pascutto]

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

Rybka 3 does have separate search functions for white and black. No need to reverse engineer anything. You can find this trivially with a disassembler that gives stack dumps.

I don't think it's a particularly good idea but very clearly both Rybka 3 and Ippolit have this, and I know in fact of no other program that does.

[michiguel]

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

Rybka 3 does have separate search functions for white and black. No need to reverse engineer anything. You can find this trivially with a disassembler that gives stack dumps.

I don't think it's a particularly good idea but very clearly both Rybka 3 and Ippolit have this, and I know in fact of no other program that does.

I did not say that... the message I quoted did...

Miguel

[Dr.Wael Deeb]

Even without any evidence I am sure that Ippo/Robbo is clone: why publish superstrong engine without getting any money or at least fame for
your real name? Additionally if your first engine is 3100 rated it is not
credible at all!

Jouni

Someone that believes in a Marxist philosophy would do what the author of Ippolit--according to the Ippolit website--has allegedly done. It makes perfect sense. As far as the strength is concerned, well that is not proof.

Evidence:

1.) Vas says that it is a Rybka 3 clone.

2.) Jury says that it is not Rybka 3 but rather it is Rybka 4.

1 and 2 are mutually exclusive.

3.) There is strong evidence that it was developed from many sources and MAY have some reversed engineered ideas from Rybka.

1 and 2 are greatly discounted by 3.

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

5.) There are some differences and some similarities in its play compared to Rybka, but the differences are much more prominent.

...

How many versions of Fruit was there before it became the strongest program in the world 1.0, 1.5, 2.0, 2.1? What if there were several versions of Ippolit, in "chess language b", before it was translated into C to make it available to the west?

Bottom line is that the evidence for cloning is rather thin and the evidence for a lot of originality is rather high. So, unless someone that is trusted disassembles Rybka 3 and finds a nearly 1 to 1 correlation in search functions, evaluation functions or evaluation constants then I for one choose to err on the side of innocent until proven guilty rather than joining in on a witch hunt.

A great presentation of the issue Michael,thank you
Dr.D

P.S.Making your way back home,be carful not to be assaulted by one of Rybka's sectarians

[Gian-Carlo Pascutto]

A great presentation of the issue

...that is factually just flawed.

[Alexander Schmidt]

1.) Vas says that it is a Rybka 3 clone.

2.) Jury says that it is not Rybka 3 but rather it is Rybka 4.

1 and 2 are mutually exclusive.

3.) There is strong evidence that it was developed from many sources and MAY have some reversed engineered ideas from Rybka.

1 and 2 are greatly discounted by 3.

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

5.) There are some differences and some similarities in its play compared to Rybka, but the differences are much more prominent.

Just one word: Yes.

How many versions of Fruit was there before it became the strongest program in the world 1.0, 1.5, 2.0, 2.1? What if there were several versions of Ippolit, in "chess language b", before it was translated into C to make it available to the west?

And not to forget that there where at least one other engine that came from nowhere to the top... Remember?

[gerold]

O.T.

Take this time to thank you for a book you wrote.
Its for Shredder but i use it for a special engine.

Best.

Gerold.

[Zach Wegner]

4.) The programming style is very different and rather amateurish. I doubt that Vas would start from an example like that of Fruit and devolve the code into separate search functions for white and black. It would be naive to think that it would be faster.

Rybka 3 does have separate search functions for white and black. No need to reverse engineer anything. You can find this trivially with a disassembler that gives stack dumps.

I don't think it's a particularly good idea but very clearly both Rybka 3 and Ippolit have this, and I know in fact of no other program that does.

Of course, this doesn't have to be THAT ugly, if you use templates. I tried something like this for evalution, but it ended up being pretty ridiculous (having a whole file of stuff like shift_up(), etc), and I wanted to minimize the code size.

And given what Vas has said about his development, it seems that most of Rybka's code is just a big custom preprocessor (a b compiler? ) in order to create lots of code changes easily. Even back in the Rybka 1 days he was doing something almost like this, as all of the eval constants were in big hardcoded tables.