12.04 Ubuntu Developer Summit Proceedings

From 31st Oct 2011 – 4th Nov 2011 the Ubuntu Developer Summit took place in Orlando, Florida, USA. Attracting 800 attendees from 42 different countries, this mix of Canonical employees, volunteers, upstreams, vendors, and partners engaged in 420 sessions across 9 tracks.

These sessions were used to discuss, design, and plan the Ubuntu 12.04 LTS edition that will be released in April 2012. This post provides a summary of many of the outcomes and decisions finalized at the event.

The Keynote

Mark Shuttleworth, the founder of Canonical and Ubuntu and head of Product Strategy at Canonical, provided the opening keynote, first talking about where have been focusing on Ubuntu:

  • Engaging and embracing developers is important (for Free Software) to bring applications to the Ubuntu platform.
  • developer.ubuntu.com makes it easy for developers and ISVs to target the Ubuntu platform and it’s large number if users by providing tools, a quick start guide, documentation and a software distribution mechanism through the Apps Portal and the updated Software Center client software.
  • Ubuntu One is the leading Client Services platform and the now available Windows client attracts more users to the Ubuntu platform.
  • User Testing is an important driver for Ubuntu to improve usability and make Ubuntu look and feel great (LightDM, AppSwitcher, Music Lens and monospace were called out specifically).
  • OpenStack became the cloud infrastructure of choice in Ubuntu and administrators can expect a tight integration into the Ubuntu Desktop.
  • Ubuntu is set to become the preferred cloud platform, as best guest but also as best host.
  • ARM gains more traction in traditional PC areas, e.g. Data Centers and Ubuntu contributes to the enablement of ARM.
  • LTS releases are now supported for a period of 5 years, which caters more to the needs of Corporate Customers.
  • Services in the Cloud will become more important and Ubuntu’s wider cloud offering and specifically JuJu defines Cloud Services as so called Charms, which distill expertise into reusable units.
  • The demand of Corporate Users is rising and Ubuntu will deliver a reference Corporate Desktop for these large scale deployments.
  • Power Users will be a focus for the upcoming release, where the same amount of effort that was spent on the general usability of the Ubuntu Desktop will be geared towards the needs of Power Users.

Mark then moved on introduce what he referred toas The Ubuntu Vision:

  • The Ubuntu mission of Linux for Human Beings cannot end at the Desktop but needs to take into account the devices that will be used by Human Beings in the years to come.
  • Ubuntu will power tablets, phones, TVs and smart screens from the car to the office kitchen by LTS 14.04.
  • Unity will be the coherent user interface for all these devices.
  • Ubuntu powered personal computing platforms will connect seamlessly to Desktop, Server and Cloud.
  • Ubuntu’s personal Cloud – Ubuntu One – will deliver, share and synchronize data and content between Ubuntu powered devices and other devices that are supported by Ubuntu One.

See the video of the keynote.

Track Summaries

Throughout the 400+ sessions there was a huge amount of content, projects, proposals and plans discussed. Here is a summary of each of the core outcomes in each of the tracks. If you want a finer-grained level of detail, you can read all the blueprints and work items.

Desktop

  • LTS release with continued focus on quality, performance and stability over new features.
  • Focused plans on specific areas including Multi-monitor support, boot speed, text-free boot and power consumption.
  • Priority focus on stable platform.
  • Discussed which versions of software to ship and support in LTS including X.org and GNOME.
  • Details were discussed around system-config-printer, using Unity-Greeter as lock screen, Unity configuration (design TBD) and Unity-Greeter enhancements.

Server and Cloud

  • This is an LTS, so we will concentrate on improving stabilization and reliability, prioritizing production ready features, and building out a robust test infrastructure.
  • We will integrate the next OpenStack release, Essex, and work with the project itself around improving automated test efforts.
  • Streamline and simplify the Ubuntu Cloud install experience.
  • Improve Orchestra to deliver a better experience for provisioning, deploying, hosting, managing, and orchestrating enterprise data center infrastructure services.
  • Build on our ARM Server rollout in 11.10, with a focus on ensuring we have the necessary support for anticipated hardware and targeted workloads.
  • Continue to develop Juju, making it ready for production-ready deployments, by adding support for features such as co-located services, as well as closing open bugs blocking deployments.
  • Build up the collection of Charms for Juju as well the community of Charmers.

Foundations

  • We will push the envelope on python versions, and try to get at least one application from the desktop CD fully ported to python3 for precise.
  • We will work with the desktop team to improve the boot experience and ensure that the desktop is always free of text error messages on boot, shutdown, and suspend/resume.
  • We’ll work with the kernel team to investigate making boot even faster, by dropping or minimizing the initramfs in the common case.
  • We will do a better job of collecting information from all our users about problems, instead of just early-adopters, by implementing an actual crash database.
  • And we’ll have a huge focus on things we need for a 5-year LTS: cleaning up our DNS handling, cleaning up friendly recovery, robust IPv6 support, continued polish on upstart, and usable CD images every day to help keep the velocity of development high throughout the cycle.
  • We might switch to 64-bit CDs by default depending on the outcome of investigations this cycle.

Hardware

  • Use the v3.2 kernel.
  • Merge the amd64 generic and server flavors.
  • Investigate dropping i386 non-pae flavor.
  • Investigate dropping 32bit non-smp powerpc flavor.
  • Investigate providing an armhf flavor.
  • Help community to support a -lowlatency kernel.
  • Ubuntu Kernel Delta Review.
  • Proposed disabling and then dropping aufs.
  • Ubuntu Kernel Config Review (just some of the bigger highlights).
  • Move ext2 to be enabled as a module, rather than builtin.
  • Build in SATA_AHCI.
  • Build in XEN_BLKDEV_FRONTEND and XEN_NETDEV_FRONTEND.
  • Investigate enabling DYNAMIC_DEBUG.
  • Ubuntu Kernel Power Management.
  • Investigate power management improvements for the kernel.
  • LTS backport kernels.
  • Will be shipped by default on the 12.04.2 point release.
  • LTS backports kernels only supported for 18mo life cycle.

Community

  • Great discussions around the Ubuntu Vision and Televisions, Tablets, and Phones. Many ideas were shared and fleshed out. Mailing lists will be set up soon to continue those discussions.
  • We successfully ran our first Ubuntu Leadership Mini Sprint and reviewed many of the points raised in a community survey with a series of outcomes.
  • We had some great discussions around continuing to build the Ubuntu app developer programme, the next iteration of developer.ubuntu.com, and how to continue to improve Qt integration in Ubuntu.
  • We discussed how to better recognize community contributions and put some plans in place for encouraging the frequency in how people thank each other as well a syndicated set of posts to encourage this culture.
  • Daniel Holbach is creating a developer advisory board to help refine and grow developer and packaging participation in Ubuntu.
  • We had some great discussions for better design team collaboration. We created the concept of experience teams, the Ayatana mailing list will be closed and other lists set up to better facilitate discussion.
  • We reviewed our community governance and agreed to propose that governors cannot sit on more than one council at a time.
  • The Application Review Board has been expanded and a set of bottlenecks were reviewed. The group also proposed time limits on responses and when automated packaging can be used.
  • The loco.ubuntu.com team agreed to implement Locale based teams and discussed some other minor refinements.

Design

  • We conducted some design sessions during UDS to help community members with design and to promote design thinking. This effort continues informally.
  • We communicated benchmark user testing results.
  • We brainstormed ways to involve the community in design and research with a commitment from the design team to provide guidance and support

Security

  • As an LTS, primary focus is on making the release supportable in terms of maintenance and security updates
  • Focus on testing and quality of team owned projects such as AppArmor, ecryptfs, ufw and general OS hardening.
  • Support Server and Cloud LXC containers work via AppArmor
  • Support via Apparmor the Application Review Board, Foundations and developer.ubuntu.com goals surrounding application confinement

ARM

  • LTS release for the first time on ARM
  • Bringing up ARMhf (hard float) Archive as that will result in between a 2% – 10% better thoughput on both Desktop and Server. ARMel archive will be support for at least 12.04 but not as LTS.
  • Focus is on ARM Server, with the release of 11.10 ARM server on Panda as proof of concept, will test and stabilize ARM server on real hardware.
  • Drop initrds if possible: while initiated by the ARM team, the dropping of initrds (or rather the use of a minimal in-kernel initrd) will happen for all arches. Full initrds will only be used if actually needed which should speed up booting a lot.
  • The ARM bootloader handling tool “flash-kernel” will finally be refactored, this will gain us a hardware database which we can re-use in other tools as a nice sideeffect
  • ARM QA (server and client) will see a lot more automation, even on actual bare metal, not only in virtual machines.
  • Ubuntu ARM and Linaro will work closer together. Many sore points and duplicated work between the two teams have been/will be solved.
  • Virt. will ship with LXC, KVM will be researched and improved by ARM and will be incorporated by team into Ubuntu as its ready.

Other

  • We reviewed the release processes used in Oneiric and brainstormed ways we could make the overhead of managing the release (meetings, reports, etc. ) more efficient for Ubuntu and the recognized derivatives in Precise. Clarification on what was meant to be a recognized derivative flavor, and what it meant to be LTS was discussed.
  • The release schedule for 12.04 (Precise Pangolin) LTS was reviewed and some adjustments were planned based on the feedback.
  • We looked at ways of improving the defect analysis processes, metrics that would be useful, and infrastructure to support doing the prioritized defect analysis.
  • The underlying bug life cycle was revisited with the Launchpad team. The developer survey results on bug states and usage were reviewed, and steps towards articulating and unifying the bug state transitions in use with different Ubuntu teams were discussed.
  • The documentation team and news team members started brainstorming some improvements they’d like to see in Precise.
  • Ubuntu Studio team will be aiming to incorporate a low-latency kernel by default to support audio work.
  • Edubuntu team is targeting to become an LTS release in 12.04, and will be applying to the Technical Board, and working on tasks to support this and future roadmaps.
  • Kubuntu team had discussions on improving support for muon, filesharing, accessibility, CJK languages, and documentation. A new Kubuntu Active CD based on Plasma Active is being planned.
  • Lubuntu team will be focusing on improving the user experience in this release, as well as furthering some of the technology transitions in progress.

Be sure to keep up to date with us on Google+, Facebook, Planet Ubuntu, and Canonical Voices.

Article contributed by Jono Bacon on Wednesday, November 16, 2011.

Bookmark the permalink.

9 Responses to “12.04 Ubuntu Developer Summit Proceedings”

  1. Ronald Says:

    Please ! resolve the problem that affects the majority of laptops/netbooks users about the overheat, temperature, fans.

  2. James Says:

    I think I remember reading an article just a week or two ago that talked about this. I think the regression is upstream in the kernel, and there are patches that should take care of it. I _think_ it was something along the lines of, rather than adhering to the proper spec, the kernel devs decided to copy Microsoft broken implementation of acpi which is, unfortunately, what hardware manufacturers develop for.

  3. Jay Ginete Says:

    Of the things I want for Christmas is to make Ubuntu work with my laptop’s optimus video card.

  4. Doug Says:

    Pleased to hear of a focus on power users. The importance of power users cannot be overestimated, as it is still power users who are pushing typical end users to try Ubuntu. Think about us, and our desire to have customization of our desktops, small icons, etc., and not get hung up on a ui that is sub-optimal for users who can keep track of dozens of windows without lenses, etc.

    Keep up the good work. I’m sticking with you through this rough period with unity. Please make it worthwhile.

  5. R.Mena Says:

    I think that the support for new laptops with hybrid video cards is important, in general every people have problems with video, although solvable, but this kind of laptop cant work with both cards, being this a disadvantage. I think that ENVIDIA provide restrictive drivers, but and AMD(ATI)?

  6. Md ashickur Rahman Says:

    Please solve the issue of WiMax modem. We Bangladeshi are facing a lot of trouble to use wimax dongle modem.

  7. nlefin Says:

    I think the most omportant issue are laptops with hybrid video cards and the HDMI port on ATI/Intel hybrid cards.

  8. Tom Huber Says:

    i use ubuntu now for about 4 years, consider myself a power user in a corporate environment. But i still wonder, why there is so little effort on IMPROVING exinsting office suites, in particular libreoffice; Tried lotus sympohony, imho it is not a real alternative to LO; When i say improving i mean GUI, ms office compatibiltiy (docx, xlsx, pptx,..) AND improved functionalities! Make it even BETTER than MS office, this shouldn’t be too difficult! So this is my wish for the future.

  9. Patchiu Says:

    This is my mock up to improve the dash of Unity:

    http://gnome-look.org/content/show.php?content=146829

    I hope my dash design is taken into account by Canonical for Ubuntu 12.04

Leave a Reply