On Thu, May 6, 2010 at 8:38 AM, Ryan Oram <r...@infinityos.net> wrote: [...] > > The idea of developers being better maintainers is a bit of economic > theory. My goal is to make the Linux distribution more scalable. If > developers concentrate on their packages and distributions concentrate > on the core operating system, this make for a much more efficient > system there is much less duplicated work. The cost of adding more > software to a distribution under this system would rapidly approach > zero, as the distribution would just run a minimal check and do > minimal testing. > > Ryan
Hi, this may sound attractive but it really feels like a half-backed argument. First, as some already mentioned, most upstream have little or no knowledge on packaging. More importantly, not all upstream are ubuntu-centric. The linux world is very wide and while ubuntu is for sure a very visible distribution, it is not the only one by a long shot. Furthermore some upstreams are not even linux-centric... Even if all upstreams were ubuntu-centric, this kind of approach has pro and cons, you can't just pretend the pro outweight the cons without a detailed study of the userbase, which is quite large... Packaging is hard and maintaining a consistent distribution is a huge task. I for one would love to have newer versions of some software when I need newest features or specific bug fixes, but throwing new versions of working stuff all over the distribution would eventually lead to more broken stuff. PPAs are great for this, it is a clear improvement for the previous alternatives (stick with whatever is in stable or run unstable) even if it may be a definitive solution. I do think a proper way to achieve this may be worth discussion, maybe by extending PPAs or making it easier to have nightly builds of the latest upstream version for existing packages, possibly in a semi-official and integrated way (like offering test channels for some applications in the application manager) but I'm pretty sure this must be used with caution: it can cause fragmentation and combination-dependant bugs that are hard to track (but if done properly, ubuntu-bug would also be able to collect this kind of information). BTW, I think something in this line has already been experimented (and failed) in the debian world, wasn't it the purpose of Ian Murdock's "Progeny Componentized Linux"? Best regards. -- Aurélien Naldi -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss