First. It could be that these programs were derived from Rybka through some sort ofhyatt wrote:I don't have an opinion on Rybka 3 at all. I am convinced of just one thing in the ip*/robo*/etc + rybka 3 discussions.
ip*/robo* are reverse-engineered from _something_. Based solely on 40+ years of programming and dealing with compiler optimizations in a compiler course I taught for years. Humans don't write code that looks like that. Not ever. Whether it was modified _after_ the RE was done, I can't say, because I have not tried to ascertain where the original code came from. But I am _certain_ that the "family" is the produce of RE first. Whether something was modified after that is unknown. Whether the source was Rybka 2, Rybka 3, or something else is also unknown. So for those, I have no opinion.
I have a strong suspicion that they came from Rybka. Based on the author's comments. But whether it is a clone or a derivative I have no idea, and it really doesn't matter enough to cause me to want to step through that very messy code and compare it to something that is more naturally written.
Do you believe that ippolit and friends to _not_ come from Rybka at all? Because they certainly came from somewhere, and do not represent original programming effort. At least, of that I am certain. Unless it was written by aliens that think like an optimizing compiler when they write code, that is...
'cleanroom reverse engineering'-like process. When I read the comparison document
last year, I noticed that all concepts had not just been reimplemented, but also reconsidered,
changed, refined, expanded or simplified (deliberate or not, it is hard to guess).
Second, the sources that -we- see might just as well be compiled+decompiled versions
of itself for obfuscation reasons. Either that or there really was a lot of Vodka involved
when writing it.
