Michael Sherwin wrote:
The GPL may be more binding here in this country than in others. Fabien, the author of Fruit released his program under a license which is supossedly a legal document with something to it (or is it all bluff?), that expressly forbids someone to distribute any derived work whether it be in part or in full. So, once again I say that if Vasik (who also lives in the United States) used Fruit as a guide (as a template) to create a work alike to Fruit, but, faster and then merly added a few things then Rybka is a derived work and the GPL does only allow this, if the license is agreed to and followed. The only real question is whether it is binding under law or not. Period.
I believe the binding exists, however I don't know if this GPL license has been fully tested in the court of law. Most people prefer to settle without going through the entire process.
As far as derivative works are concerned then the copyright license applies only to the pieces of code that were under that license and that constitute a licensable item. For example, a complete implementation of an algorithm is subject to the license, where 'for(int i=0; i<10; ++i) cout << i;' is not. Furthermore, what doesn't get covered by the GPL is the algorithms and ideas (etc.) and in this sense if a programmer implemented the GPLd algorithms on his own then his work is not a subject to the GPL.
Some aspects of programming are not copyrightable (such as 'ideas', etc.) and to those the GPL doesn't apply.
Consider this (while keeping in mind that all those examples are often insufficient) -- you can copyright "Romeo and Juliet" but you cannot copyright the story (idea) behind it.
In case of Rybka it is immaterial how he got to the final (released) version, since as a private user he can do anything he wants with the GPLd code, however the GPL comes into play when he released the software to others. If the version that was released to the public contained GPLd code in it (pieces of code that are copyrightable, not just 'ideas') then Rybka should have been GPLd. If however, there was nothing from the original GPLd code included in the released version then Rybka is not subject to the GPL.
What Vasik used as a template to get going is irrelevant, the only thing that matters is what he actually released to the public. One can always request (if so inclined) the source code from Vasik for the original Rybka with compiler instructions such that after the source is compiled it produces the same exact binary that was distributed to the public and then analyze this code for similarity with the GPLd code. If there is no similarity then the code is not a subject to the GPL license.
The GPL is not a bluff, BTW. There are very few places where one (company) can rewrite the original code from scratch -- it is money/time inefficient and quite often the resulting product is not as good as the original and the future developments that continue to occur on the GPLd side are then lost (!) to those who have rewritten it all. The GPL is very real and very good license which in the long run makes more business sense than succumbing to proprietary 'hell'. But the GPL will become hell if it implied (or even contemplated) the restrictions that you are talking about.
Regards.