> It's not really at all clear that this was where the mistake lay. I never thought I would be advocating more management, but here goes...
Debian, as another poster pointed out, has grown from ~50 to ~2000 developers. And those developers, being "geeks" rather than "suits", respond to problems by working harder on their jobs (as I believe Collin pointed out), rather than by strategic re-alignments. Other engineer-centered organizations (HP and Boeing are famous examples) have run into this problem before. Now as Manoj has pointed out many times, Debian is not a corporation and has no way to force developers to work on a certain project. But Debian as an organization does have some pretense to becoming a widely adopted, enterprise-class distribution -- I think this was made clear by the debate during the last DPL elections and by the candidate who won -- and doing that requires a degree of responsiveness to the needs of its real (as opposed to ideal) users. I haven't experienced first-hand the styles of the previous and current RMs, but at least looking from the outside, it would seem that the current RM has been much more effective (aside from failing to take into account the security infrastructure problems) than the previous one in driving woody to release. This would seem to indicate that management can make a difference. What about increasing the "management overhead" for Debian? Instead of just having DPL and RM, the RM could be given some minions, there could be PMs for different areas, etc. They would make tactical and strategic decisions regarding releases, and advertise needs as they arise. There, the idea is out there; wail away at it. By the way, the frustrating thing for me personally about the release process hasn't been the long delay in making woody stable. I have never even used stable. The frustrating thing for me has been how the freezes and concentration of effort on getting woody stable has kept unstable from moving forward with the bleeding edge. For me, a faster stable cycle could actually be bad, since it could mean more freezes. This may well be a minority position, but, in the interest of Debian getting to know the needs of its customers (a phrase calculated to annoy Manoj :-), what are the percentage users of potato, woody, and sid? I assume this could be estimated from average daily activity for each archive. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
