Vincent Untz has just announced the release of GNOME 2.31.6, which is the first development snapshot of GNOME 2.32 following the announcement last week that GNOME 3.0 is delayed until March 2011. With the 2.32 release not being the 3.0 milestone, the 2.32 release will continue to use GTK+2 rather than the still-in-development GTK+3...
Whats the point of doing yet another 2.x release this fall? It seems way too late in the development cycle to just switch back to gtk2 after all the prior testing was on gtk3.
Actually, that's not the case - most of the 2.31 series has been developed and tested under 2.0, and even in the previous development release (2.31.5), half the packages still hadn't moved to 3.0 and the rest had only just done so. So while it's certainly a late decision, I don't think it's as big a risk as you suggest.
As to why, 2.32 might not offer huge visible improvements to the user, but it does give them a chance to get some more of the low-level changes into a stable release prior to 3.0. Ideally, code targeting 2.32 (i.e avoiding deprecated APIs) should work unmodified on 3.0 (as some apps - e.g sound juicer or evince - already do). Wouldn't you have more confidence in a 3.0 release knowing that most of the code has continuity, rather than being largely re-written (as KDE did)?
Wouldn't you have more confidence in a 3.0 release knowing that most of the code has continuity, rather than being largely re-written (as KDE did)?
Not really, seeing how nobody gains anything from it? It doesn't help your API-design seeing how you can't really refresh that much. A redesign of that magnitude is going to be broken or pushed back either way(which happened to surprise only the craziest fanboys, but as a KDE-user I'm going with "Told ya!"), so you don't gain that much stability. As for code reuse, well, the point of a redesign is to "do it right"(or better, seeing how you can always make anything better), so you should keep that to a minimum anyway.
Also: Is your point really that the last stable release of a series that it probably maintained by third party(Red Hat) should make low level changes? Cause that is borderline insane.
Originally Posted by Delgarde
Actually, that's not the case - most of the 2.31 series has been developed and tested under 2.0, and even in the previous development release (2.31.5), half the packages still hadn't moved to 3.0 and the rest had only just done so.
That's not really something to brag about - GNOME just delayed their big release target to march, so saying that most software hasn't even started testing correctly... not a good sign.
I finaly had the chance to take the gnome-shell for a spin on Fedora and I was mayorly/majorly disapointed. As many people here would know I am not exactly a fan of Gnome, but I realy thought that the new shell would be a game changer and put a face on Gnome, rather than being 'the two panel Ubuntu interface'.
First issue: memory game.
Normaly computers are all static. Semantic is the new desktop revolution. KDE4 has semantics, but it's more of an 'add-on' rather than designed around the concept. But not having the programs, settings an places in one place (the ultimate user-friendly thing ever) is terrible for anyone who's past the age of 12. Zero organization. So how does the grandpa group gonna remember their way around?
Second issue: pre-timeslicing age
Having multiple windows in multiple workspaces is a massive productivity invention. Sometimes different sections in a single workspace (manual grouping of windows in one workspace where windows cover each other) can be realy usefull. This is totaly a living hell in Gnome-shell+Mutter; the mouse has to move all the way to activities everytime you want to raise a covered window. Now the idea, ofcourse, is to group everything by means of workspace with Gnome3's design. But what if you have to do a lot of mind numbing work that involves switching windows rapidly? Let's say between word processing, finance and the browser and these apps do not fit in a single workspace without seriously overlapping? At least two of them require more than half of the screen space, so this massive productivity killer.
Please hold back the release of the shell and go back to the drawing board, because in my book this GUI is a dissaster and totaly kills anything that comes even close to user-friendlyness.
Keeping code continuity where possible is a Really Good Thing, actually. Having a nice bridge release between 2.30 and 3.00 will make it easier on everyone to migrate - KDE 4.0 changed far, far too much at once.
(OB-microsoft: They're making that mistake with Windows Phone 7... squared, actually!)