The Linux Kernel's Fuzzy Future (InformationWeek)

Posted Dec 9, 2004 17:59 UTC (Thu) by sepreece (subscriber, #19270) [Link]

We've been told, by people with enough OSS credibility that they should know, that the OSS community generally doesn't want input on what they should be building, it just wants code. That is, "Don't tell us what you want, build it and give it to us and, if it looks to us to be good enough and if we think enough other people might want it, maybe we'll put it in the mainstream."

Those of us who build products (or systems) typically aren't interested in being OS houses, too. We generally want to spend our resources on building our domain functionality and want to be able to treat the OS as a platform that we don't have to worry about.

For us, building consumer products on a non-X86 architecture, this means that we get our Linux from a distributor that does architecture and feature tailoring for us, and if we need specific features we may work with a third-party to develop them, under whatever license is appropriate.

This isn't a complaint, it's just a description of [our perception of] the way OSS development works. And it's fair to say that it is a barrier to Linux adoption for at least some potential adopters. I think that was the article's point - that the lack of plans for extending Linux in specific directions and the lack of a stabilization approach in the new development model are a problem.

There are benefits that we balance that problem against, and, so far, we believe that the benefits outweigh the problems (or the cost of working around the problem) for us, but that doesn't mean that it isn't a problem.