Hi Adrian, On Mon, Apr 04, 2005 at 08:54:11PM +0200, Adrian Bunk wrote: > [the usual discussion about release procedures]
I'd like to point out that these discussions lead nowhere just as many others. Changing Debian is currently: - discussing the problems - discussing approaches to solve them - flame war Maybe at some time something is decided but nobody is able or has the time to do it. What I think is that we lack the possibility to do any changes because it is by far too much work. The whole Debian infrastructure can only be recreated because there are a few key people knowing how everything works and we even have backups. I mean - how is a ftp-master server installed? Somebody throws a Debian release on a machine, installs the archive maintenance scripts from CVS and configures everything. Or a backup is restored. Anyway - reproducing the master server is hard because only few people have the knowledge and they don't have time. How is a buildd setup? There are more developers knowing how to do that but it is still knowledge that is not distributed. My point: We need the possibility to recreate the current Debian infrastructure for proposing changes to the current situation. So changing the way Debian runs should be more like: - find a problem (easy :)) - create a team to work on it (not as easy ;-) - implement a proposed solution and test how it performs - have a vote if it should be adopted as official So basically to suggest a new release process we could have an inofficial Debian release (Adrian's Debian :)) which is much faster by using a different approach and after seeing how it works we could adopt it for the project. Greetings Torsten
signature.asc
Description: Digital signature