I agree that Ubuntu never was a perfectly stable distro. You usually had to wait 2 or three months after a release before having most visible bugs fixed.But yeah, if bogot is going to suggest that all of these other distros are alpha; than yes, a comparison is fair-game and in this case; Ubuntu loses big time.
But things are improving in this regard. For example 13.04 was usable from day one for me, which was actually the first time since I use Ubuntu.
Development releases are also much more reliable than before. Now, when you download a daily, you can be reasonably sure that it will boot and that you will have a graphical session. That was not the case before.
Upgrades are also smoother than before (way less breakages).
In fact it became a bit boring to test an Ubuntu development release since everything is usually fine with it. There are bugs but the system is usable most of the time.
Of course there is still work to do, but it's definitely moving in the right direction.
Now, 13.10 will most likely be a special case because of Mir (still not switched on by default, so we don't know yet how much issues it will bring). But it makes sense to integrate it early to be more confident in its reliability in the next LTS.
"intermediate" releases are for advanced users anyway.
UOA/empathy is also extremely buggy in 13.04 to the point where its pretty much not usable: https://bugs.launchpad.net/ubuntu/+s...y/+bug/1170832, https://bugs.launchpad.net/ubuntu/+s...s/+bug/1069882, https://bugs.launchpad.net/ubuntu/+s...y/+bug/1168582 (I don't have any of these problems with upstream empathy/GOA, only ubuntu's crappy buggy UOA). And this one: https://bugs.launchpad.net/ubuntu/+s...n/+bug/1159411
Really annoying problems with ubuntu's patched nautilus 3.6 (I don't have issues like this at all with the same version of nautilus in any other distro): https://bugs.launchpad.net/ubuntu/+s...s/+bug/1173966
Really annoying issues with the version of glib in 13.04, that causes bugs in software such as pidgin: https://bugs.launchpad.net/ubuntu/+s...n/+bug/1108056 (don't have this problem with pidgin in other distros).
And some very annoying visual polish issues that also effect usability: b0rked libreoffice global menu: https://bugs.launchpad.net/indicator...n/+bug/1153350, unreadable text in nautilus gear menu caused by 13.04's gtk theme: https://bugs.launchpad.net/ubuntu/+s...s/+bug/1159449
13.04 is just as buggy as previous releases were. If ubuntu every hopes to achieve their goal of reaching mainstream, they need some serious improvements in QA. Its not OK for releases to be this buggy if you are aiming at the mainstream market.
Ubuntu has become synonymous with the word "buggy" for me
Last edited by bwat47; 08-25-2013 at 12:52 PM.
As XMir should not, if developed right, have any noticeable impact for end users, it makes sense to activate it by default as soon as it is ready (on a technical point of view I mean).
No bug is going to be reported and fixed if no one use the software. So better start such migration early and use multiple steps.
That's the whole point of "release early release often" in fact.
Well, just the fact that they are supported for 9 months vs. 5 years for LTS should be enough.Citation needed.
AFAIK, nearly all computers sold with Ubuntu pre-installed still come with Precise nowadays.
If it's still not enough for you, I know it was stated multiple time on ubuntu-devel during the "let's ditch intermediate releases and call what's between LTSs a rolling release" debate. You can look for it if you want.
Last edited by Malizor; 08-25-2013 at 01:57 PM.
I find it really amusing that Qt doesn't get the same flack by having a CLA and actually selling a proprietary version while Canonical doesn't do this and as far as we can tell never will.
Also, IRL, things do have an impact, and there is a concept in software maintenance called bug surface, which is why you don't just add pointless layers.
Yet again, XMir is just an X server running on top of Mir, and testing it for apps is relatively trivial. And still, being a tester should be optional (since Ubuntu doesn't pay us to use Ubuntu, it's voluntary testing, and voluntary testing should be, well, voluntary), and that's the main drama: they are forcing novice users to become testers, if not for between-LTS releases (since you consider them glorified betas), they are still putting that on a LTS release.
No, it isn't enough. Give an actual reason why 9 months support implies a more advanced user than 5 years support. Also, with citation, I meant citation. Quote any Ubuntu source for that.Well, just the fact that they are supported for 9 months vs. 5 years for LTS should be enough.
Irrelevant. A lot of people gets Ubuntu installed by a friend. All of the novice users I know, use it after a friend installed it. In several of such cases, I'm that friend.AFAIK, nearly all computers sold with Ubuntu pre-installed still come with Precise nowadays.
I know of such a debate, but that has nothing to do with being "for advanced users" or not, but with maintenance. It would become for more advanced users, but take a hint why they didn't make them rolling releases.If it's still not enough for you, I know it was stated multiple time on ubuntu-devel during the "let's ditch intermediate releases and call what's between LTSs a rolling release" debate. You can look for it if you want.
Second, when there is a single developer, you can't blame him for doing whatever he wants: it's his sole work, nobody helped him, and he did gave his code. In the case of the CLA, the CLA exists because it's expected that several people will contribute, and they probably contribute because they want to help a free software project (with the exception of, in this case, Canonical paid developers). But then, you have more rights than all of the other developers, which is not only unfair to the user, but to the developers themselves.
On the Qt vs. Canonical situation, I don't like such a CLA at all, either.
On the other hand, the fact Canonical didn't make that closed source version yet, doesn't mean they won't, and chances are they have such intention: if it were to protect the project from being illegally used in closed source applications, they won't ask the right to sublicense, but the copyright (since they need to be the copyright holders for suing). They explicitly ask for the right to relicense and not the copyright, and guess what, they probably do that because they want to use that right to relicense.
And to readdress the point of the single developer, you can fix it by having multiple developers, which happens to be the case in most of the important projects.