GPLv3, Linux and GPLv2 Compatibility
April 14, 2007 1 Comment
A third possibility is that the Linux kernel developers will decide that it’s not worth the hassle and just accept the GPLv3. I suspect that this is what the FSF is hoping will happen. Depending on the changes in the next two drafts of the GPLv3, it still might. But, it’s not looking likely that the kernel developers will yield. Frankly, if I were in the kernel developers shoes, I wouldn’t either. The GPLv3 serves to further the goals of the FSF, but the current draft actually hinders Linus’ goals and the goals of Linux in general.
Another possibility, complete speculation on my part, is that the Linux Foundation becomes more than just a loose consortium of companies sponsoring Linux kernel development. It becomes the copyright holder for the Linux kernel, not taking copyright assignments from contributors like the FSF, but copyright licenses like Apache does, so the kernel developers still hold their copyright on the code. The Linux Foundation releases a license with basically the same terms as the GPLv2, but without the legal ambiguities, obscure language, and anachronisms. Like the GPL, this license is copyleft. Like the GPL, this license requires the release of modified versions under the same license. This license clearly defines the concepts of linking and modified works, making it easier for Linux distributors to be sure that their segmented distribution trees are in compliance. Over time, more and more projects currently released under the GPL adopt the Linux license, because it is more legally precise and more comprehensible to the average developer than either the GPLv2 or GPLv3. Eventually, Linux distributions switch over to the Linux license, leaving only a small branch of GPLv3 (or v4 or v5) code to be downloaded separately (if the user chooses to do so).
It had occurred to me that it might be nice if the FSF did a sort of updated draft of the GPLv2 that included very minor improvements while not introducing the major fundamental shifts of the GPLv3. It’s clear they wouldn’t do this now, as it would hinder the adoption of the GPLv3. I hadn’t thought of the possibility of someone else improving on the GPLv2. Seems unlikely, but maybe just the thought of it will smooth some things over in the GPLv3 process.