Ubuntu: 1/3 of Installs, Upgrades Result in Fail?
Inevitably, there will be what we call ’showstopper’ bugs in distribution releases. When this happens, in the vast majority of cases, those responsible for the distro get fixes out immediately. There has been some complaints lately about why Ubuntu should take a closer look at their quick release schedule. Take a look at these articles: “Early adopters bloodied by Ubuntu’s Karmic Koala” and “‘Ubuntu Needs a Longer Release Schedule!’” Whoa, now. Not pickin’ on Ubuntu. Fedora had one in its 12 release in November, as well: “Unpriviledged users could install packages.” Not good. OpenSUSE isn’t immune either. In the release of 10.1, the package manager was horribly broken.
Seems like we could tolerate such ’showstopper’ bugs once in awhile. But to have this many complaints about excessively-problematic releases of one particular distribution release after release seems a little off to me:
- Is Ubuntu broken?
- Tech-no-media blogger asks, ‘Is Ubuntu broken?’
- How to Fix broken Ubuntu Feisty Fawn
- Idea #15033: Don’t release 9.04 with broken major-hardware support
- Hey Ubuntu, Stop Making Linux Look Bad
- Advice Against Upgrading Ubuntu 9.04 to 9.10
Seems like I’m not being objective here, doesn’t it? Maybe I’m trying to get people not to use Ubuntu, or give Ubuntu a pounding. Nope, merely observing numerical data.
What data?
First, let me explain something: An LTS release is generally accepted to mean “a fairly rock-solid release.” The nonLTS releases are interim releases for those who want bleeding edge. By definition, they should have more bugs in them.
Secondly, these numbers might be taken with a grain of salt, as they did come from a support forum. How many people log into a support forum to say that everything worked perfectly? Probably not as many people as had problems.
With that said, let’s look at the numbers:
| Release | type | Upgrades w/many unsolvable problems | Installs w/many unsolvable problems | Flawless upgrades | Flawless installs |
|---|---|---|---|---|---|
| Gutsy Gibbon | nonLTS | 35% | 33% | 22% | 22% |
| Hardy Heron | LTS | 30% | 33% | 21% | 25% |
| Intrepid Ibex | nonLTS | 44% | 45% | 18% | 25% |
| Jaunty Jackalope | nonLTS | 31% | 35% | 29% | 32% |
| Karmic Koala | nonLTS | 33% | 38% | 31% | 34% |
Make an objective judgment of that for yourself. The LTS release had at least a 30% failure rate. Nearly the same failure rate as the nonLTS releases. The flawless upgrades and installs were nearly always lower than the number of seriously problematic installs/upgrades. Well, weren’t they? Where did these numbers come from? this spreadsheet, to which ubuntuforums.org linked.
One or two problematic distribution releases is forgivable. But an apparent failure rate of 1/3 on every release?
Mr. Shuttleworth… bro… ease up a little. Implement some improved testing or quality assurance policies or something. You have an incredible user base. Every time you release, you are seemingly causing measurable aggravation for about a third of your users. Each person that experiences a failed Ubuntu install/upgrade incorrectly attributes the problems to Linux in general. Try and help the Linux community out — not by sheer numbers, but by quality, polished installs and upgrades. Spread the joy. With all your resources, surely you can make some progress here.
While I applaud Canonical’s success in helping Linux become widespread, they also seem to be giving Linux a black eye amongst would-be Linux users.
And please, if you are one of the third that has a marvelous experience with Ubuntu, don’t shoot the messenger. I’m merely sharing data.
