Ubuntu’s Feature Friction

One common thread among dissatisfied Ubuntu community members1 recently is that they’re concerned with the speed and haste with which new untested features are being dropped into the main Ubuntu release. To find out how people feel about feature introduction:

How do you feel about feature introduction?

  • I like them, but they're a bit under develoepd. (42%, 196 Votes)
  • I'm happy with the new features. (27%, 127 Votes)
  • There are far too many bugs for a mainstream release. (18%, 84 Votes)
  • Introduced without Community Consensus (10%, 48 Votes)
  • Introduced without Warning (3%, 11 Votes)

Total Voters: 466

Loading ... Loading ...

Do You Test Ubuntu throught development?

View Results

Loading ... Loading ...

What is your Primary Concern?

View Results

Loading ... Loading ...

Please respond only if you use Ubuntu or have used Ubuntu within the past two years for an extended period.

1 Internally to the Ubuntu community, not from other communities, see Aseigo’s excellent post on community identity.

Tags: , ,

16 Responses to “Ubuntu’s Feature Friction”

  1. Isaac says:

    IMHO, the only way for new features to be extensively tested is through incorporating them into a release, and the best releases to put them in are the non-LTS releases – maverick and natty. So what is all the fuss about? If we need a new photo management app, then let’s try Shotwell. It’s development is going a lot faster than F-Spot. And if we are going to switch to Unity (yay!) then lets try that too.

    If you need stability, go with a LTS. If you want to help make the next LTS even better, then test and help build some great standard releases to test new features and begin supporting new applications and libraries.

  2. Hey Martin,

    That first question is very problematic. In a multiple choice survey question, one key rule is to make sure that the answers are mutually exclusive. You should also ensure that you only ask only one question at a time. For instance, I agree with both “I like them, but they’re a bit under developed” and “Introduced without Community Consensus.” There seem to be two questions mixed in there, one about the quality of the features and one about the process by which they were introduced. If the question isn’t sound, your results will not be either. I would find it hard to draw significant conclusions from this survey.

  3. Shane Fagan says:

    I think the problem is not that the features arent tested its that the known bugs arent getting fixed in time for release. The me menu in 10.10 is a great example of that with the text not going away when the user clicks on he box, its a known bug and will be fixed for 11.04 but just didnt get fixed in the timeframe of 10.10. So most of the bugs in the releases are known and just need a little bit of time to fix.

    Hardware compatibility is the biggest problem at the moment and thats because we need to have a great out of the box experience.

  4. ethana2 says:

    I think releases are generally at the level where full-scale testing by the Ubuntu community as it is might be considered appropriate, but every release needs a .1 after like a month in order to have a CD people can use that’s competitively stable with current installations and competing mainstream desktop platforms.

    I’m always on alphas and betas and all that, because I’m a new feature addict, but I don’t *DARE* put anyone else on a current release until it’s been out for like three weeks.

  5. ethana2 says:

    ..In other words, we need to account for and deal with the fact that MS and Apple have more people in their beta testing programs than we have in our entire. community.

  6. qhartman says:

    I agree with the point above about the first question being poorly stated. Too many things I agree with in there.

    My biggest problem is that there is so much effort going into “new and cool” when there are hoards of known, long-standing bugs in foundational apps that are being kept. For example, Evolution has tons of “papercut” type bugs, some of which have been a thorn in my side for years. I’ve done as much as I’m able to help fix them, but still they persist.

    I’m excited about the new stuff, but I sure wish I could stop worrying about / getting bitten by all the little broken rough edges.

  7. Randall says:

    IMHO we need *much* stronger and better messaging around who should be running pre-release software, “regular” Ubuntu releases, and Ubuntu LTS releases, and on which hardware (certified please.)

    We have a user community issue in that many/most Ubuntu users came from platforms where there was no opportunity to be meaningfully involved in anything other than fully released and supported versions on (de-facto) certified systems.

    It’s time to snap out of it :)

  8. The thing is once Ubuntu uses something new by default it gets lots of attention and bugs get fixed. Most of the time there’s a PPA you can use to get those fixes sooner, but you can always wait 6 months.

  9. mdupont says:

    The main issue IMHO is not introducing new – and potentially buggy – features, it’s introducing such features en masse in LTS releases. It seems that LTS releases are treated like regular ones (eg. all this new stuff changes in 10.04).

    The model I’d expect is the following : introducing radical changes and bleeding edge features in the release after a LTS ones, stabilize them and introduce less intrusive changes in the releases after, in order to have mostly well tested and solid features when the next LTS release goes out.

  10. NosajDraw says:

    Can’t make your polls work, so whilst the first one says I’ve voted, it still offers me voting options, whereas the 2nd too stop me from voting saying that my vote is being processed…

    So apparently, I have voted, not voted and am in the process of voting all at the time time.

    Schrödinger’s cat ain’t got nothing on me!

  11. David says:

    @Isaac: LTS releases have typically seen the introduction of new features not present in previous releases. Rather than use the non-LTS releases as testbeds for getting things rock solid for an LTS release, Canonical has a bad tendency of introducing new and untested “features” in LTS releases, then fixing them in the subsequent non-LTS releases. That’s one of the reasons why I’ve held back from upgrading past Karmic.

    @Shane Fagan: Totally agree. The thing that drives me nuts is that bug fixes are typically not included in updates, and so I have to wait 6 months for the next release before I get the bugfix. But instead of getting the bugfix, I get a new release that includes other bugs that won’t get fixed for another 6 months, and so the cycle continues.

    @ethana2: I like the .1 updates for each release idea!

  12. “What is your Primary Concern?”

    Not listed. My primary concern is accessibility. Features are added without testing for those of us that have accessibility needs. 1 pixel handles for resizing when my resolution can be 1900 wide or more, just doesn’t cut it. New colors, great for some, for those of us with vision difficulties, not so great.

    So add buggy features all you want, just make sure that the rest of us forgotten can shake our fist at the lack of development rather than the fact that we can’t even verify its lack of usefulness. :D

  13. MarkC says:

    I agree that the first question isn’t a good one for radio buttons – there are three options there that I would have liked to have chosen!

    One other problem with sticking with LTS releases is that it holds back all the other software as well. I use non-LTS because I want more recent versions of Firefox and Inkscape than the LTS versions usually provide. But in doing so I have no choice but to also get the half-baked new features.

    With the current cycle of an LTS every four releases, how about this:

    LTS: Stable, well-tested code, no fancy new features.
    LTS+1: Similar to LTS, but with the new features available as an option for people who want to test them.
    LTS+2: The more stable and complete new features get turned on by default; less stable ones remain optional
    LTS+3: All the new features that are expected to make it to the next LTS are on by default to provide more widespread testing.
    LTS-NEXT: Stable, well-tested code, no fancy new features…

  14. mark_b says:

    As I understand it the development of Ubuntu is more to do with the core system as opposed to application development?

    If this is the case, then I would propose that in between each LTS, a ‘blank’ or core Ubuntu is made available for people to test the applications and perhaps through popularity, the best behaved apps make it through to the next LTS? Even updating the core based on feedback every 6 months.

    This way the distinction between LST and dev releases are made very clear and new users gets a distro that works out of the box and the rest of us get t test/play/muck around with configurations etc to our hearts content :)

    my 2p’s worth

  15. Marlowe says:

    Perhaps you should use Debian.

  16. doctormo says:

    @Marlowe – Not everyone here is looking after number one, lots of people are advocates are compassionate social suppliers of IT services to their local community. Do you propose I install Debian on my mums machine even if she’s 3k miles away?