Author: Vincent Diepeveen
Date: 04:16:45 09/25/02
Go up one level in this thread
On September 23, 2002 at 23:25:41, Dann Corbit wrote: >On September 23, 2002 at 20:47:38, Peter McKenzie wrote: > >>On September 23, 2002 at 17:59:35, Dann Corbit wrote: >> >>>There is zero chance that Ruffian is a copy of Fritz 6. None at all. >>> >>>Fritz 6 does not play under Winboard. Ruffian does. Nobody has the Fritz 6 >>>code except the original author. >>> >>>Not even any sort of slight, remote chance is involved. You are barking up the >>>wrong tree. >> >>Dan, you are getting carried away here. Just because you say something 4 times >>doesn't make it true. > >Worked for a German political party in the 30's and 40's. OK, OK, -- you've got >me there. > >>It is of course possible to disassemble Fritz or any other chess program, >>thereby gaining access to everything you need to steal the engine. > >That could happen. I don't see a connection to the claim that this is a >modified version of a commercial engine. If you reverse engineer something in >that way, I also expect it to be a lot harder than writing it from scratch. > >>Impossible? Absolutely not, no way, nada, not on your life. >>Unlikely? Yes. > >I have the binary. It's not a copy of Fritz 6. In fact, I will say that it is >impossible. Is it a reverse engineered Fritz? Well, an asteroid could strike >this area before I hit the post button. I think it is about the same liklihood. > >>But I've spoken to at least one chess programer who has had a peek inside a >>commercial engine using a disassembler. I believe that you can easily find out >>quite a bit if you have the right skills. > >I could do that too. To reverse engineer a chess program would take many years, >though. Then you are at the biggest beginner level of reverse engineering. I hope you realize that reverse engineering is very easy for search. A few hours at most. C programs are harder to reverse engineer than well written assembly programs. The better optimized programs are and the better written, the easier. Crap programmers are harder to reverse engineer, simply because of the many instructions needed to do a single thing. Yet the progress of the vaste majority of commercial programs, especially in the past times, i'm 100% convinced it was by reverse engineering. This is how great things have been 'shared' among many programmers. Nothing as easy as to dissassemble something or read out a chip. No i'm not saying i'm doing it, i even lack the time for it currently. So i'm not saying ruffian is designed by reverse engineering, no that would be way too complicated. Much easier is if a commercial programmer wanted to do an experiment and called that ruffian. However he can't hide its evaluation function. Amazingly that no one ever compared games of ruffian start of this year (februari 2002 or something) at standard level with how commercial programs play. At that time no one cared shit about ruffian, but it was 2700+ rated and the only program that positionally won a few games from DIEP. Which in itself is amazing. So i had a good look to what it is. I'm amazed about the many threads here. Is everyone that stupid to not see it? People look for tactics right now. that's dumb. if i turn of SE in DIEP then things get solved at other plies. Try instead games it played at februari 2002 or something and compare that with commercial versions of programs. A generation ago of course... >>Of course I'm not at all suggesting that Ruffian is the result of disassembling >>Fritz, I wouldn't have the slightest idea about that although it seems unlikely. >> I'm just saying that in theory (and it would be quite alot of work) it is >>possible. > >Lots of things are possible. This could be a copy of Fritz. But it isn't. It >could be a supercharged TSCP. But it's not. > >If it were a copy of anything, I would say that Pepito as a base is the only >chance of it. It has about the same branching factor and both Winboard and UCI >support -- in both ways like Pepito. That's on the one hand. On the other >hand, I don't think it is a clone of Pepito. I think it is an original program. > But it isn't a commercial program, which has been ran through a string editor. >It just isn't. The probability is zero. The chance that it is a reverse >engineered Fritz? I'll grant you one in a trillion here.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.