split 3 from Zach technical thread

[GenoM]

Well, I guess we can expect Theron, Zach and co. to start disassembling Rybka 3 and posting that here now. After all, their mission to take down Vas won't amount to much if they don't.

Albert

You're a bad fortune-teller.

[tiger]

Boy, not much meat in these threads.

Anyway, I personally find it minorly useful to distinguish whether "important" parts of the code (such as search / eval) of Rybka 1.0 vs. "unimportant" (such as input processing and string compares) are from fruit in part because of implications regarding other versions. If the only overlap between Rybka 1 and Fruit were things like string compare and input processing, it would be quite easy to believe that versions 2.0 + have no overlap as those parts would be easy to rewrite without harming performance. If other parts (say, the hash table, see function, move generation) were copied, I would expect they would be harder to replace without effecting performance and would be possibly passed on to future versions. This distinction is unimportant to those who just care about whether there was a GPL violation (probably more of an emotional issue to those that have made their code public than those of us who haven't [not sure why it seems so emotional to many without a codebase]).

Just out of curiosity, were there similar size ruckuses in the past regarding other programs? It is kind of depressing to think that if I found some amazing new algorithm that took my engine over the top that I would be reticent to make my engine public for fear of a string_equal() hunt finding something hidden somewhere in my code. Luckily, odds of me coming up with enough big breakthroughs to make this relevant is small (and hopefully the odds of someone finding something appearing to be a GPL violation is also small).

-Sam

There have been such cases in the past, many of them actually.

However it was generally more clear-cut. The cases were about programs that had been almost completely "cloned". That is the reason why the word "clone" is still being (incorrectly) used even in this discussion.

The discussions were softer and shorter because they involved mainly people who could technically argue. And as I said it was easier to decide. In the current discussion there are two problems:
- it's harder to decide
- many people who cannot understand the technical aspects react emotionally and create a lot of noise



// Christophe