On Mon, 2006-04-24 at 21:23 +0930, Tim wrote: > Tim: > >> One of the dislikes I have with Fedora *is* the release schedule. > >> There'll be a release around a certain date, ready or not, sensible or > >> not. A new release comes up around the time the last one has many of > >> the bugs ironed out, yet the new release is so radically different that > >> you can't take advantage of the information gleaned over the last one. > >> It won't be a fixed version of the prior release, it'll be a different > >> version. It's case of throwing the baby out with the bathwater. > > Paul Howarth: > > Taking it as read that Fedora is a distribution that tries to keep up > > with upstream releases, how long would you suggest the interval > > between releases be? The longer you leave it, the more different it's > > going to be. > > When it's ready. When it works. When something is a significant > improvement over a prior release to justify a whole new OS. > > There's zero value in bringing a product out on a certain date > regardless of its operating condition, and from some points of view, > there's *negative* value in doing so. > > There's a significant advantage in having a long-lived OS, which allows > programmers to build for a known goal. Some program development is > quite long-term, and to have the underlying system changed on you, > several times, might well mean that you just can't be stuffed developing > for it. It'd be different if there was a usable system standard that > worked across all distros, but they're just about all different and > require custom implementations. I much preferred how my old Amiga > worked, I had the same OS on it for many years, it did its job well, and > didn't need replacing. I updated *applications* when I felt like it, > and that was all I needed to do. > > Remember how older, Red Hat Linux, releases used to have longer lifes, > with sub-versions before radical changes (e.g. 7.0, 7.1, 7.2, 7.3)? > Where OS faults (supposedly) got fixed, before moving onto a new one. That's fine if that's what you want; distributions like this exist - CentOS for example. These distributions do, by their very nature, lag behind the "leading edge". *Someone" has to be at the leading edge in order for problems to get resolved, and there's a fine line to be drawn between being too close to that point that the distribution is highly unstable and difficult to use (perhaps this might apply to rawhide) and being too far behind the leading edge that people are wanting widely available new bits of software that aren't deemed stable enough yet to be included. Given the popularity of Fedora and the roughly equal numbers of people clamouring for new releases of KDE/Gnome etc. and complaining about instability, my gut feel is that the balance is about right. Paul.
