There's been a lot of discussion on the Linux kernel mailing-list lately about how the kernel development system / hierarchy / mechanism works (It's the longest single thread I've ever seen.). A lot of people are rooting for some changes, citing things like the proliferation of unofficial "developer" trees and the non-responsiveness of kernel maintainers.
I'm no journalist, and I'm sure Kernel Traffic
will cover this much better, but here are a few mails I thought were interesting (too bad the web-archives thread so poorly):
The thread started out with a mail with a subject of A modest proposal -- We need a patch penguin
, which called for someone to step up and create a unified development tree much like the -ac tree Alan Cox maintained for a year after 2.4 was released. Linus replied
to a reply to that message, defending the current system and starting an interesting thread, which seemed to imply some relationship between the organization of the kernel source and the organization of the kernel maintainer hierarchy.
Later, someone brought up
the idea of creating a "patchbot" which handle some of the more menial tasks of kernel maintaining, and would expose more of the work of the kernel developers to the public (so to leverage the whole open-source "bazaar" thing better, I'd guess.)
were started to create just that, but the most informative-seeming message
about the subject said it had been tried before.Linus also started a "tangential" sub-thread
in which he expressed some interest in using Bitkeeper
for managing the kernel source tree. This turned into a technical discussion on the shortcomings of BK, during which Linus went so far as to say he would certainly use it, provided certain problems with it were fixed.