Le 20/06/2022 à 07:12, David C. Rankin a écrit :
Rolling releases have the Achilles's heel of having to coordinate among various upstream packages to make a desktop work. It's trivial to keep your vanilla distro running in a rolling manner without anything that taints the kernel, but mix nvidia (or other proprietary drivers), and kernel updates and you have a real recipe for downtime until whatever new incompatibility can be resolved -- which sometimes can take weeks. (look at Virtualbox and the 5.18 kernel now, Larry had a patch a month ago, be we still have no 6.1.36 release from Oracle)
Thank you for your eloquent answer, David. I'd say I like my Linux distribution as I like my motorcycle: boring. Of course it's fun to fool around on a Honda VTR 1000 Firestorm (I know, I did it). But when I'm planning a trip around Europe, I'd rather use my battered old BMW 750 which doesn't look very sexy but which takes me to the end of the world. The problem has been stated in an earlier post, and you can even read it in various distribution release notes. All the "exciting" new features, all the "technology previews". When I read "exciting", I understand "half-baked implementation of something that will potentially ruin my weekend because I will have to find a fix". There's a curious phenomenon going on in the Free and Open Source world. Whenever something reaches a state nearing perfection, the developers/maintainers toss it out and start again from scratch. Remember the transition from GNOME 2.32 to 3.0? Remember the upgrade from the latest KDE in the 3.x series to 4.0? Red Hat's implementation of KDE was one of the cleanest I've ever seen. So clean that they decided to ditch it completely and go for GNOME (meh). I get it: you have to change things to go forward. But sometimes - as in the cases stated in the last paragraph - progress comes in the form of a huge regression, and the user ends up with a series of components that have to be potty-trained again from scratch. That's precisely the "raison d'être" of a release. Provide a platform that sports no surprises, with security updates over the whole support cycle. Part of my job consists in teaching computer students here at the École des Mines d'Alès, and go figure, most of them have a hard time to grasp the concept of Long Term Support and low-risk security updates. They all run Arch, BTW. :o) My initial rant was fueled by the fear that OpenSUSE Leap in its current form might just go away or be replaced by something so exciting it's barely usable. Yes, I first read about it on Distrowatch and then asked about it in the forum. Maybe it's just an unfounded rumour after all, go figure.
But through it all, another very good 15.4 release was made (aside from a mirror hiccup or two) and many bugs were authored, many fixed, and the SLE codebase is the better for it. It's hard to see a desire to change that process by eliminating the community release model -- whatever they call it next.
Yes, OpenSUSE Leap is one of the nicest things that ever happened to the Linux workstation. So one word to the OpenSUSE maintainers. It JustWorks(tm). That's where the *real* excitement is. Cheers, Niki -- Microlinux - Solutions informatiques durables 7, place de l'église - 30730 Montpezat Site : https://www.microlinux.fr Blog : https://blog.microlinux.fr Mail : info@microlinux.fr Tél. : 04 66 63 10 32 Mob. : 06 51 80 12 12