Post a Comment
I don't know how you think that's going to be prevented. When you look at other desktops like KDE, but mainly Windows and OS X, Gnome 3 has a very, very long way to go before it gets up to those standards. A new shell isn't really going to cut it. That's simply going to be a painful process of release early and release often, as is normal in the open source world, until it stabilises.
Personally, I think March 2011 is pretty optimistic and if they manage to release something finished then then it probably won't be good enough.
You do remember KDE4 being released while they tried to tell everyone it was beta? Nobody cares if it's not ready yet, it was big new and shiny.
Something happened to me as a Linux user when i switched completely. When software gets a new release i want it. I'll sit on my PC all night if it doesn't work because for some reason I have to have it.
That's why delay is good, keep it away from the ravenous compilers who also complain the most when something isn't that great.
You do have a point, more than a few distros played a pretty big part in the KDE 4 situation pushing the new shiny code out as default before it was ready. To that end, we can only hope they'll be more responsible this time.
But... the KDE team also made the brilliant move of trying to redefine decades of accepted standards of what a .0 release was.
I'm not saying keep it away from the early adopters who are used to having things break and enjoy it, but for the love of $DEITY don't make it default so that Johnny New User's first impression of Linux is a pretty but half working system
But... the KDE team also made the brilliant move of trying to redefine decades of accepted standards of what a .0 release was.
I'm not saying keep it away from the early adopters who are used to having things break and enjoy it, but for the love of $DEITY don't make it default so that Johnny New User's first impression of Linux is a pretty but half working system
KDE4.0 wasn't even default on Kubuntu let alone any other Linux distro. It was a seperate downloadable ISO for people who wanted to beta test.
The problem was people were impatient - went out of their way to run beta software and then complained when everything didn't work properly.
So they really only have themselves to blame, not KDE, Kubuntu, OpenSuse nor anyone else that offered up KDE4.0 packages in parallel to KDE3.5.x
Edited 2010-08-02 17:09 UTC
That is open source I'm afraid. Developers get to decide when they want to release something otherwise you have something like Debian - it drifts on for years. If it isn't good enough then distributions shouldn't ship it, or ship it as a development option at install.
That's exactly what KDE did do - and it was totally, utterly and completely ignored. That's a problem for the distributions, and one that they have generally failed utterly and miserably at. They ship anything that is >= *.0 and then whine when the bug reports come in.
Edited 2010-07-30 17:23 UTC
That's exactly what the KDE team did and, though KDE4.0 was labelled as a ".0", it was very clearly distributed as BETA.
The problem wasn't KDE being released too early, it was impatient users ignoring the warnings and expecting release quality from beta software.
Heck, even the KDE distros weren't offering KDE4.0 ISOs as the download - users had to specifically choose KDE4.0 and read through the warnings before downloading.
People are already calling it a disaster and they did the same with KDE4. In my view, GNOME Shell is very stable and if it does crash, it drops back to GNOME 2 gracefully.
Novell have implemented it very nicely into openSUSE 11.3 with it's own session. I'm currently working on a GNOME Shell Sonar theme (The default GTK theme in openSUSE), which is nice because it uses CCS3 and SVG elements.
I think GNOME Shell is rather innovative but it takes some getting used to because of the different workflow.
Edited 2010-07-29 19:22 UTC
But I think the decision to release a 2.32 is a mistake, instead they should have released a bugfix release of 2.30 and focused on fixing a great public Beta of 3.0 in September, so both users and developer could help test/port/evaluate the new desktop and its new API's before it's release in March. One should remember that after Gnome 3 is released, it will be nearly impossible to change API's, so the more programs is ported to the new API's during the Beta stage, the more is the systems design can be tested and refined before the final 3.0
Irrelevant, because that porting is already happening, and has been for years. Most of the "new APIs" are things that have been gradually added to 2.x over the past few years, and much of the actual Gnome 3 porting effort is in getting applications to stop using old APIs that 3.x is dropping.
Well written 2.x apps and libraries can be ported to 3.x just by running a search/replace over their Makefile, no code changes required. Some of them are even doing that during the transition, able to pick one or the other at build time.
I now that, much of Gnome 3.0 is to remove legacy stuf (bonobo,gconf,exposing internal structure in gtk, theme engine...). But does for example dConf or new theme engine support the need of all application? How many application does access internal structures in gtk in their own legacy code, and does gtk+ 3.0 api let them do what these 2.0 application need it to do? The new theming engine with CSS support, how well is it tested?
I am well aware of this, but how many "bad" written 2.x apps exist? They will need to be ported/rewritten also.
Like KDE4?
This is ridiculous hypocrasy. There have been plenty of GNOME (as well as other open source software projects) releases that have been underdone. There have also been plenty of commercial software programs released that are definitely NOT underdone.
Disclaimer for trolls and people that can't help but get emotional over logical criticism: I am no GNOME or KDE advocate (I like both). I also support open source software. I DO NOT like the constant theme of releasing underdone software which seems to be particulrly common in the open source world (not that it doesn't happen in commercial software, just that it seems to happen less frequently).
So, hopefully GNOME fulfills its promise and a "finished" version is released!
Edited 2010-07-30 09:09 UTC
I can't wait until March 2011. Sorry. I should say, I can not wait. I am over 30 now and rather do something else than spend the rest of my life waiting for software to be released.
I avoided KDE4 because it was new and required a commitment and tinkering so I stuck with Gnome waiting for their next version upgrade. But March 2011 is too far so tonight, I install KDE 4.x and cross my fingers hoping that Gnome 3.0 won't make want to change again. 
Seems to me that the solution to this is so simple I don't know why it hasn't hit everyone over the head. I enjoy playing with alpha and beta software, and I'm currently running Firefox4-beta2 and Google-Chrome-dev (without issues, I may add). I also expect to install Ubuntu Maverick-alpha3 next week when it's released (on a Btrfs partition!).
Any good reason why I can't install both Gnome2 and Gnome3 and choose at login time which one I care to use? At least with Ubuntu, it's very easy to choose which desktop I want at login time, and most distros are similar (granted, some distros do make that difficult).
Seems to me like the best of both worlds. I can choose stable or unstable whenever I login, depending on my whim at the moment. Because Gnome3 is not currently available in the Ubuntu repository, it's not accessible to me right now (unless I want to compile from source, which I don't). Making the binaries available would get more people experimenting with it, and help speed development along.
But yes, Gnome3 should be clearly marked "beta" or even "alpha" so as to eliminate any confusion with the stable Gnome2. That whole KDE4.0 fiasco should not be repeated. So Ubuntu developers (if you're reading this) - go ahead and hit me with Gnome3.0-beta1 - I'm ready!
Edited 2010-07-31 03:54 UTC




