Saturday, 12 February 2011

What now for MeeGo?

So Nokia has dramatically reduced commitment to MeeGo and has cited, amongst other thinks, MeeGo's inability to deliver a focussed baseline with sufficient speed. I happen to agree with this failure (and given Nokia was a significant part of MeeGo's management I don't think there's a blame issue - more a how do we fix it issue)

Assumptions and observations:

  • MeeGo is intended to provide a viable but focussed baseline upon which vendors can build compliant products; not to be an expansive and 'complete' linux distribution.
  • MeeGo has limited dedicated resourcs and focusing them on a reduced MeeGo core will improve quality.
  • MeeGo's main customers are not end-users - they are device vendors : they should be the focus of our core engineering team's design, delivery and QA effort.
  • MeeGo core does not appreciate the difficulties a vendor has in tracking MeeGo;
  • A visibly secure development model is important to the perceived integrity of MeeGo - so visibly restricting write access to the core is important.

Proposal:

  • MeeGo Core is confirmed as not being a linux distribution
  • An open MeeGo project (openMeeGo?) is created on the community infrastructure to provide a reference MeeGo distribution
  • Packages not *essential* to the delivery of a compliant MeeGo Core are moved into the community OBS (emacs, vi etc - maybe even the reference UXes) where they are available for use by development teams and end users.
  • "openMeeGo" acts as a reference vendor and provides a forum for reviewing and improving the processes MeeGo uses to communicate releases
  • MeeGo community (which includes core developers) has a significantly lower barrier to entry.

3 comments:

  1. It has been a long time that people emphasize the value of "MeeGo Core" but I always doubt that because every disto has such kind of core and nothing proves MeeGo's superior and MeeGo core is NOT designed to be superior from high level, for example, rpm. (See distrowatch, you will find how well deb based distros do)

    I always view MeeGo's UI and App framework is the key (well, it could be part of the "core"). But it is yet to be proven and MeeGo brand issue already drove away some cooperation.

    Last thing, don't forget Intel has the full control of MeeGo, so it is JUST another Moblin now.

    ReplyDelete
  2. I think this proposal will be a non-starter with Intel and the Linux Foundation. But perhaps you ought to take this up with the Steering committee, half of which is now irrelevant. :-)

    ReplyDelete