Posted Mar 11, 2010 20:47 UTC (Thu) by rriggs (subscriber, #11598)
In reply to: Who is Fedora for? by iabervon
Parent article: Who is Fedora for?
Personally, I prefer a Fedora to be short-term stable. What's the average lifespan of a Fedora release? From Alpha to EOL, we talking less than 18 months? That's not a long time to ask packagers and maintainers to maintain a package. At most you have 3 versions to maintain, and that's with a very aggressive release cycle.
I have the following options:
CentOS/RHEL: long-term stable for Servers and such.
Fedora: short-term stable for development, playing with fairly new features. Some stuff is broken (e.g. ATI OpenCL support). I run this on my desktop.
Rawhide: unstable for experimenting with bleeding edge features on machines that can be partially functional most of the time and non-functional for short periods. This I usually reserve for VMs.
I use each of these options. They all have their place. It does suck to have my email to get screwed up because of an significant upgrade to Thunderbird mid-cycle.
I really don't see a need for a midpoint between Rawhide and a short-term stable Fedora. Anyone who thinks they want this is really asking for "stable bleeding edge", and that is just oxymoronic.
I need something that I am comfortable running on as a primary OS on my desktop. With that I am able to provide valuable feedback to Red Hat and the Fedora teams. If they get too wild with Fedora, I will no longer be able to use it as a primary platform. I'll play with it from time to time. But I will no longer have the incentive to provide detailed problem reports.
Posted Mar 11, 2010 21:22 UTC (Thu) by iabervon (subscriber, #722)
[Link]
I was actually suggesting that the current Fedora is not actually short-term stable (because it accepts changes that aren't security/stability fixes and breaks your Thunderbird); the current policy is actually more like something between short-term stable and rawhide, where changes that are considered production-quality but are different in ways that may matter to users are accepted. My idea was that Fedora Rolling would be managed like Fedora N is now, and Fedora N would be more conservative, where switching to KDE 4.4 within a cycle wouldn't even be considered.