Michael Sherwin wrote:Sedat Canbaz wrote:Hello,
I have no patience for final conclusion from the experts that Strelka 1.8 UCI is clone or not ?!
I am still waiting to start the Strelka test,just becouse my SCCT tournament conditions are:
''All engines must be own original work (its not allowed clones)''
Hurry up guys,the poll expires in 3 Days !!!
Best Regards,
Sedat Canbaz
Strelka is not a clone. It is a carefully derived work of Fruit that was also carefully designed to be like Rybka. Even the 'stolen' tables have been removed and replaced with functionals. It is a Rybka work alike and no more a clone than Rybka is. If two unrelated students go to the same school and learn the identical proceedures and then apply those proceedures in their work, they do not become clones of each other.
I really don't think we know if it is a clone or not, because computer chess has no formal definition of exactly what constitutes a clone and everyone has a different opinion. Whenevery I ask someone to clarify what it means to them, I get back no answer or something very ambiguous.
To me, a clone means:
You took someone elses's program, made a few small tweaks (such as altering string tables or a few small tweaks of the eval) and claimed that you wrote it. This is the common sense of the term, and most of the programs on Ron's list are of this sort. To me, this is a serious crime -- the equivalent of plagiarism.
To me, there is a separate category where you take a bunch of code from another program and, without permission, paste it into your program with very minor tweaks (just enough to get it to work in your program). I am not sure I would call this a clone program, but I would call it stealing.
To me, not a clone means:
You wrote a chess program based on the ideas of other programmers articles, books and programs but did not use their code. Or, if you did use their code you used it with permission.
There is gray area here also. There are some who have taken a very simple program and completely reworked it so that it is no longer recognizable. I know of several instances of this. It the current form, the program does not play like the original at all. What to call it? I'm not sure.
Reverse engineering a commercial chess program is also a grey area to me. I don't particularly like it myself. But if the people who do it are within the law, then who am I to complain?
And an aside:
If GPL software really does work the way that people claim (you enter into a contract not to learn what it does and copy the idea) then I think that GPL is really the worst imaginable software license. If (in the future) it should become clear that it really does work like that then I will no longer work on GPL projects. To me, it would be morally repugnant, similar to software patents. I would (of course) obey any laws relating to GPL software.