Preferred form for whom?
Posted Mar 11, 2011 18:14 UTC (Fri) by rahvin
In reply to: Preferred form for whom?
Parent article: Red Hat and the GPL
unless those reasons are explicitly spelled out in the license, no, you can't possibly be agreeing to them. what you do agree to is what's in the license, nothing more, nothing less.
If the licensor hasn't made his definition of preferred form (if it's different than the GPL definition) available to the licensee at the time the license is executed the extrinsic evidence available to the court will be considered. The best of that evidence is the statements by the FSF and RMS on what the GPL means and why they wrote it, after all they are essentially the legal team that crafted the license. To legally demonstrate any other interpretation you would have to prove that the licensee was aware of some alternative interpretation at the time of license AND prove that the GPL definition is ambiguous. Whatever your preference is (and most people will have different preferences) is not necessarily what the licensee interprets when he draws a license.
The problem I have with this discussion is that people are pulling a single sentence out of the license and trying to reinterpret that sentence however they want without regard to the agreement or it's history. There are numerous places within the GPL where the preferred form is defined implicitly although not directly.
3. B. Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or,
The reference to preferred form, in my opinion, is a direct reference back to the paper copy exploit section and the binary source exploit section. Although the language could be a little clearer I don't think there is any doubt that the preferred form is machine readable source code on a medium typically used for software interchange. How that source is arranged as long as it's not obfuscated is beyond the license agreement. It would be absolutely silly to assume that whatever the licensee interprets to be their preferred form is the form the licensor has to comply with.
There is absolutely no doubt that the vanilla with patches and comments method is far more useful. But the license only requires that you be provided the source code in a manner that is editable and machine readable.
As far as providing quotes of RMS's speeches I have better things to do with my time. I've seen him discuss this and the methodology they went through when they created the agreement (basically trying to figure out how a naughty licensee will try to exploit the license to prevent free use of the source). I'm sure a paralegal can dig up plenty of references to RMS's and the FSF's intent when the GPL was drafted. Maybe at some point he'll weigh in on the issue but I doubt he will. If you want to pursue this line though you need to consider that there are probably more projects using the unified tarball than there are projects that are providing vanilla, patches and comments.
I think RedHat's move will complicate development of the kernel. But I also believe it's a necessary move if RedHat has any intention of responding to Oracle's attempts to destroy them. And I also believe RedHat could address the concerns by giving the concerned developers access to the information that's now missing. And I believe that if they fail to address this issue it could hurt them more than it helps. But I don't believe this is a GPL violation and I think it's a dramatic situational reinterpretation of the GPL and is something that could scare a lot of companies away from GPL. Just imagine for a moment that anyone that uses the software can redefine "preferred form" however they would like and what that would do to the community.
to post comments)