Linux forking is not likely, kernel maintainer says
Linux devotees need not worry about the Linux kernel ever forking into multiple, incompatible derivatives, Andrew Morton, lead maintainer of the 2.6 version of the kernel, said at the Open Source Business Conference here on Tuesday.
Forking of the kernel would result in a $100 million-per-year expense to maintain the forked version, thus making it undesirable for anyone to fork it, according to Morton, of Open Source Development Labs. Besides, forking would require a massive fallout amongst the kernel development team, and Morton said he has never seen any indication that that could ever happen.
"In my opinion, forking is impossible," in the Linux kernel, Morton said.
Morton acknowledged that during the SDForum open source conference, held in November in Santa Clara, Calif., he did say the kernel could be forked to accommodate sets of patches. But he clarified this statement on Tuesday.
"What I should have said in November was branching," of the kernel instead of forking, he said. Branching is a common practice in development in which a production "tree" of Linux would be branched off from the development tree, he said. Branching does not bear the negative connotations of forking, which is associated with fallout among developers resulting in the existence of multiple, incompatible versions.
Commenting on future enhancements to the Linux kernel, Morton said it could be difficult to add clustering support because there are so many different clustering technologies available, it may be hard to choose one.
"I'm pessimistic about it," he said. Morton added, though that clustering is available for Linux, but just not in the kernel.