The last two releases of Ubuntu (9.04 and 9.10) seem to generate a lot of complaints of bugs. The biggest problem is hardware because if your hardware doesn’t work, you can’t test out the software, and the hardware bugs are the hardest to find because nearly every computer on the planet is different!
I agree and I have found bad bugs in both releases. In 9.04, my Intel video driver was slow and leaked memory like a sieve. In 9.10, that problem has been fixed, but now sound volume maxes out at 26%, my mouse doesn’t work when coming out of sleep, and when I open the laptop lid, the backlight doesn’t always come on. (The last two bugs appear to have been fixed post-ship.) Of course, I’ve found bugs in every release.
The good news is that this is all very natural, and even to be expected given the deep changes that are being made to the stack. And for every user that has problems, there are users who have a better situation with the latest release. Bugs cannot be fixed until they are found, and they cannot be found until users are running the code. Ubuntu’s large user base means that it will find bugs not found by the upstream developers, which are mostly teams of just handfuls of people. And the bugs that Ubuntu users run into are nearly always bugs in the upstream code, so it isn’t entirely Ubuntu’s fault.
Ubuntu has several choices. It can be conservative about what versions of software it runs, especially hardware-related components, so as to let the smaller distros find and fix the bugs. The downside of this is that Ubuntu is not carrying its own weight, and in general new software has new features that users want. What is the point of making a release every 6 months if it contains 6-month old software? A corollary of this is Mark Shuttleworth’s “cadence” suggestion: by having multiple distributions ship on the same day, they would presumably choose to use the same versions of software and share the load in finding bugs.
Some say that Ubuntu should ship less frequently, but a better solution is for Ubuntu to put more resources on the fundamentals. With every release, Ubuntu seems to have some shiny new trend it is talking about: cloud computing, new notifications, etc. and I worry they seem to become easily distracted and forget to keep fortifying their investments in the basics: Step 1 of Linux World Domination is World Installation.
Another reason that this problem is happening is that Ubuntu has chosen to be a separate team from Debian. Many of the bugs were found before the release of Ubuntu, but there just wasn’t enough people / time to track them down, and work with the upstreams to find a fix. And because Debian is a separate team, they are not engaged in this battle. (I have been complaining about the mistake of Ubuntu being a separate organization from Debian for 3.5 years so I won’t go into this any more.)
Linux is making good progress but has a ways to go still. Ubuntu currently has 74823 bugs. Focus on the bugs! You might not believe it, but at Microsoft we had it beaten into our heads to fix bugs: a bug meant an unhappy customer, and a bug that affected just 1% of users meant that there were millions of unhappy customers. Software that doesn’t work is not worth anything, and the bug list is the most important metric an organization could possibly be focused on. It is a problem that people in the Linux community talk much more about boot time than bug count.
Follow me @keithccurtis on Twitter.