Ubuntu Developer Summit Natty, Monday

I’m writing this blog post in a chair in the ‘Grand Caribe Convention Center’, at the end of the first day of the Ubuntu Developer Summit in preparation of the 11.04 Natty Narwhal release. It’s been a very interesting first day to say the least.

The 'Afsluitdijk' as seen from the passenger seat
The TomTom is for the time estimate, I swear!

I arrived in Orlando at Saturday evening, after a nine hours and 40 minutes flight from Amsterdam, which was preceded by a two hour car drive to Schiphol. My flight was delayed by one our, but I didn’t have the bad luck of the poor people arriving from Poland who had to circle for two hours in the air above Frankfurt because of heavy mist.

The road to the factory outlet shops nearby, next to a motorway.
On our way to cheap shopping!

Sunday morning we went with a couple of people to a factory outlet store group nearby. It was at walking distance, but we did have to risk our lives by crossing the busy motorway next to the hotel, and the 30° C temperature made things a bit more sweaty than it would have been in the 8° C there is at home right now. In the evening we took the Mears buses to ‘Downtown Disney’, for dinner. There were three buses, three drivers and three guides for the four of us who showed up. Most people where at the mandatory Canonical-only keynote, or had yet to arrive. If you’re at the Ubuntu Developer Summit, please be aware of the buses Canonical arranged to bring you to either ‘Downtown Disney’, ‘Universal City Walk’ or ‘Point Orlando’ in the evenings!

This morning opened with the kick-off keynote, started by Jono Bacon and finished by Mark Shuttleworth. Jono explained how the UDS works, Mark introduced the plans for Natty Narwhal. I could talk about the funny routine with the fly in the ice cream (morale: one bug can spoil it all) or the ‘cadence, design, quality’ focus line, but of course there is only one thing that everybody is talking about. That is this: for 11.04 Unity will replace GNOME Panel as Ubuntu’s shell for GNOME on the desktop as well as on notebooks.

I’m sure many people will be shocked by this, but it comes as no surprise to me. As ArsTechnica pointed out in their article covering the announcement of this morning, RedHat—which appears to exercising complete control over GNOME Shell development—and Novell are mostly focussed on the enterprise desktop, which is completely different from the end-user’s desktop Ubuntu is focussing on. Take that and the fact that it has been frustratingly hard for Ubuntu Community members and Canonical to contribute to GNOME, it is no surprise that eventually there is some divergence.

Exciting times are ahead, and it will not be easy to make Unity as good as it needs to be, but I have full confidence that we will have delivered a great product when the next LTS release is there!

The sessions I attended today were the Desktop Team Roundtable, the Ayatana Roundtable, App Review Board Review, Ubuntu Development Advocacy and Quickly Widgets.

The Desktop Roundtable made it clear that this will be a very interesting UDS for the Desktop Team. The fallback in case Unity is not supported by the hardware needs to be planned, and there is GObject Introspection, and the default application selection that will be discussed.

On the Ayatana front we will see a merge of the GtkMenu processing code of AppMenu and Application Indicators. This should help a lot to solve bugs that are currently plaguing exotic menu items and mundane submenus in Application Indicators.

During the Quickly Widgets session Rick Spencer gave a nice overview of the great number of widgets that are already available from this neat little project he has been running in his spare time. We probably won’t see Quickly Widgets moving to a full-blown library in the near future, but there will be improvements and exciting new widgets. Rick has been working on a webcam widget, and we briefly discussed video/audio-player widgets as well.

Today’s plenaries were interesting as well. The nicest was Rick Spencer’s plenary on ‘gifting’, in which he reflected on contributing, giving something to someone else, without gaining direct benefit from it for yourself. We should keep in mind that everything resolves around the user: we give to the regular user, not so much to other developers.

The most interesting two plenary sessions were the last two by Matthew Paul Thomas and Evan Dandrea, on the applications available for Ubuntu. Statistics were shown of the number of applications available for Ubuntu, which had grown to about 3000 (*.desktop files) in six years. To show the contrast with other applications, statistics of Android and iOS were shown. Android now has 100,000 applications available in its App Store, iOS 300,000. Both operating systems are younger than Ubuntu’s six years. Clearly, something’s missing.

There is a bottle-neck: Debian and we, the distribution, package most, if not all of the software available on Ubuntu. It takes long for all applications to get updated, some get neglected, a lot of software is never packaged and developers cannot present software to the users as they want it, it has to go via a proxy: the distribution. This does not scale, hence it would be much better (at least according to my personal opinion) to make the developers themselves responsible for packaging and delivering the application to the users. The figure Evan Dandrea showed in his presentation was a circle—instead of DEVELOPER->UBUNTU->USER—with Ubuntu being an upstream for the developers, providing the platform and feedback from the users. The users give Ubuntu the feedback, and Ubuntu provides the channel (Software Centre) for the software to arrive at the user’s doorstep, but the developers themselves would do the packaging.

The hall of the conference centre at the end of the dayOf course this would require a lot of automation, but it is something that the App Review Board is very busy with at the moment. We’ve got Mago to do automated testing and there is work going on to write a Python framework for automated  testing. Ubuntu is heading in the right direction!

See you all tomorrow!

Join the conversation

9 Comments

  1. Great post about the UDS. Like many people, I don't think I'm going to like the Unity-interface (have to use it first, but from what I've seen; I don't like it).
    Those Quickly Widgets might be interesting, we'd have to see first =D
    And lastly: you mean to say that Ubuntu is heading towards a Rolling Release model? A bit like Arch linux. Well, you've got my blessing xD

    1. Ubuntu itself is most certainly not moving to a rolling release model, that would be very bad for our adoption among businesses and end-users. What we want is to allow third-party application developers to get their stuff to the users, no matter where the release cycle is.

      The 6 months cycle may work great for the base platform, the operating system, but it does not make it easy for people developing small or medium-sized applications to reach out to their users and make sure they get the newest version (or any version at all), even when the OS in in stable release already.

  2. The original speakers, actually, one of which is a high-profile Canonical employee, and the other is a Canonical employee as well. Canonical is committed to making it easier for people to offer their applications (for sale) in the Software Centre, so it is most certainly acknowledged already.

    Lets see what the next cycle will bring! Awareness is the first step, the next step is to actually change things!

  3. 'it has been frustratingly hard for Ubuntu Community members and Canonical to contribute to GNOME' – can you explain what you mean by that? Examples would be useful…

    Thanks!

    1. When people are told when forwarding a patch upstream to not mention it is needed for Ubuntu, because then it then risks being refused. When the design of software is already fixed and your design team is expected to do nothing but implementing the existing design. When instead of reusing your library and standard implementation, it is decided to rewrite an implementation that does the same. People often use the fact that most GNOME maintainers are RedHat employees to prove that RedHat does so much for GNOME and Canonical so little. I say, because most GNOME maintainers are from RedHat, that company has a large say about the direction of the GNOME project, and Canonical has not.

      Of course, this is also a collision of egos. The Canonical and the GNOME/RedHat designers all have very distinctive ideas of where the desktop should go to. Canonical would like to see the innovation take place in the distribution, Ubuntu, and GNOME/RedHat would like to see the innovation take place in GNOME. You cannot deny that causes friction, and that and the fact that there are two parties with their own ideas make it hard to get to contribute to one implementation.

      1. 'When people are told when forwarding a patch upstream to not mention it is needed for Ubuntu, because then it then risks being refused.'

        Told by who? Someone in GNOME or from in Ubuntu? If a patch was ever refused because it was from Ubuntu it would be totally out of order. That said, I'm pretty surprised by this. I've spent a lot of time in GNOME Bugzilla (though you've maybe spent more?!) and I've never seen anything like it. Please send me the IDs of any bugs where this has happened.

        'When the design of software is already fixed and your design team is expected to do nothing but implementing the existing design.'

        That story is different from the one that I have heard. I've personally witnessed the Shell designers asking for help, and they've certainly been flexible about some parts of the designs. (They're also pretty nice people, btw!) I also haven't seen a major effort by any Canonical designers to influence the shape of GNOME Shell (which I would have loved). I don't see it as a problem with GNOME frustrating Canonical, more a lack of effort on Canonical's part.

        'When instead of reusing your library and standard implementation, it is decided to rewrite an implementation that does the same.'

        I presume you are talking about libappindicator? The main reason that libappindicator was rejected was because it doesn't integrate with GNOME Shell. Was that wrong?

        I agree with your second paragraph – this is a conflict about which project should be the centre of innovation. I wouldn't say it's just about egos though. It's also about values. If things happen in GNOME, others can participate and share. I'm a GNOME volunteer – I work alongside people from lots of different companies and projects. That's what I really like about it, and that's what I value.

        Thanks for talking. 🙂

  4. Thanks for posting all the happenings at UDS. Some of us are unable to attend, so it is appreciated. I enjoy your personal opinion on developments, and can see you have some really valid points. I look forward to your other postings. Enjoy Orlando!

Leave a comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.