“I believe the 2.6 kernel is slowly getting buggier said the rather respectable looking man with short hair, spectacles and absolutely no beard whatsoever. What on earth could Andrew Morton know about Linux then? Quite a lot as it happens. Morton is the ‘lead maintainer’ of the Linux production kernel, and when he gets worried, we should all get worried. But who said anything about being worried? Oh, yes, I forgot to finish that quote “It seems we’re adding bugs at a higher rate than we’re fixing them.
Speaking at the LinuxTag Conference in Wiesbaden, Germany last Friday, Morton’s words sent a shockwave around the Linux developer community. Not necessarily because they disagree, but rather more because he said it at all. Perhaps the fact that he went on to say that if the statistically evidence proves his suspicions, then he will consider halting kernel development, at least on a temporary basis until enough issues have been resolved to play catch up. This would almost certainly involve a bug fix only kernel cycle, requiring developers to reallocate their time so that more is spent bug squishing. But with an increasing number of these developers being employed by IT companies with little financial interest in the legacy hardware that tends to cause most of the problems, how do you motivate them to participate happily, if at all?
What must rattle the cages of Linux developers everywhere is the timing of such a statement. Just as it is starting to gather some real momentum beyond the hobby user camp, finding a footing in the lucrative world of the business enterprise. All that the Linux evangelist needs is a spanner the size of this one to be thrown into their works. Of course, the open source diehards will argue that it’s just proof that the process works, it is being open. The same problems effect any long standing development lifecycle, but in the secretive world of proprietary development such things are kept locked down and woe betide anyone who is found responsible for leaking such commercially damaging information to the wider world. But those diehards are wrong, very wrong indeed. If anything it is proof that the oft vaunted claim that open source development responds more quickly to bug fixing is fatally flawed, particularly at the top end of the product food chain.
While Linux developers retain the tunnel vision that leads them to focus on new features at the expense of fixing old ones, the problem isn’t going to go away any time soon. I don’t for a moment predict the death of the Linux movement, but I do hope that Morton wields a big stick and hits the development process hard, and fast.
