On Wed, 2009-05-13 at 10:15 -0400, James Hubbard wrote: > I think being explicit in target user expectations is a good idea. It > seems like bike shedding, but it comes up enough that it's probably a > good idea. Yeah, it really isn't. I'm coming up against this *all the time*. Simple example - we're reviewing important issues for release. Imagine a bug which affects, say, five pretty popular models of Intel graphics card. They completely fail to work on boot to X - you just get a black screen. But it can easily be worked around by adding 'nomodeset' as a kernel parameter. Is that a really critical issue? It depends! If we're a distro for Aunt Flo, yes, it is. We can't go telling Aunt Flo to stick 'nomodeset' on the kernel command line. Or, at least, we should try really hard not to. If we're not a distro for Aunt Flo, it's not very important at all. If you're not Aunt Flo you should bloody well be able to figure that out yourself, or at least you should be comfortable reading the release notes or poking someone on IRC, and not at all fazed by the answer "oh, yeah, just throw 'nomodeset' in as a kernel parameter". That's just one example, it comes up all the time. And, where this thread came in - it's a very important factor in the 'what should updates policy be' question. It's important in the 'what kind of questions should the installer ask' issue. It's important all over the place. I find it tricky to work on QA or packaging stuff without knowing who our target audience is. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org http://www.happyassassin.net -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list