In a fascinating TV-type interview with ZDNet Australia, Linus Torvalds (your choice on who speaks for Linux, if anyone does) says he's in no hurry to update the Linux kernel from its present Version, 2.6, which has been out for about 2 1/2 years.
A major upgrade is a major headache for everyone, from designers and programmers to sysadmins and users. Proprietary vendors, like Microsoft, have a market incentive to do this regularly. The more upgrades you force, the more money you can make.
Open source does not work this way. Open source upgrades happen either when they have to, as in the case of a security hole, or when dictated by a wealth of new features.
The idea of "more trouble than it's worth" is stood on its head by this process. The source of the trouble shifts, from the balance sheet to the user. The idea of worth also shifts, from the producer of software to its consumer.
Imagine if, in the proprietary era, Microsoft launched a major Windows upgrade while its main rival stood still. That would be advantage Microsoft. Under the present system it may well be advantage Linux.
But what does this say about the pace of change, and the need for innovation? I don't think it says anything. It's just moved up the stack.