Monday, September 21, 2009

Atlanta Linux Fest 2009 and Crowdsourcing

A couple months ago, upon reading the tentative speaker schedule for Atlanta Linux Fest 2009, I was surprised at the Canonical omnipresence. There was personal trepidation regarding how such a presence may be interpreted, and sure enough, I wasn't alone. Certainly in my talk, I kept the vast majority of the physical speech focused on current FOSS and mentioned Ubuntu only in the confines of what we've done poorly and how we're resolving those shortcomings.

For instance, one of the most difficult things to "get right" in a default (clean) install is configuring various audio settings. A great analogous post by Ed Felten really made me reconsider the command line as the preferable interface for troubleshooting.

Many parts of Ubuntu are contributed by volunteers (thank you!) like myself. There seems to be a considerable amount of bad blood in the Linux community regarding what Canonical "doesn't do". We all need to do a better job of thanking our fellow volunteers and remembering that corporate overlords do not drive everything.

That said, I enjoyed the brief moments spent at this year's ALF (had to leave early to return to work), particularly the thoughtfulness of all the organizers (please thank them!) and attendees.

See you this weekend at Ohio LinuxFest!


  1. I think the amount of bad blood aimed at Canonical is directly related to the amount of public attention Shuttleworth have sought out. Even now Shuttleworth is planning on pushing his 2 year cadence agenda at as a keynote at Plumbers even after the negative response from the Debian community concerning freeze syncing. I would not be surprised if this upcoming keynote stirred up more bad blood over the issue of Canonical's lack of involvement in kernel plumbing.

    Without making a judgement as to the concept, the fact that Shuttleworth as a Canonical rep is pushing the agenda brings with it a certain weakness. I would humbly submit that a lot of the bad blood that is out there could be avoided if Shuttleworth would allow a prominent upstream developer to take point on the cadence discussion.. Canonical employee or Ubuntu volunteer as long as they had standing as an active upstream contributor in a plumbing layer project not controlled by Canonical.

    So...about thanking volunteers... in your opinion is there an optimal way to know who is working strictly as a volunteer and who's work is influenced by their employer without asking them point blank?

    The impact of volunteer effort is a pretty hard thing to track. Its even harder to assign that volunteer effort to one project versus another as volunteers are free to cross project boundaries in a way that employees can't cross corporate boundaries.

    To your knowledge has there been any attempt to track and summarize volunteer versus corporate contributions inside different areas of Ubuntu release development? Do you have an accurate image of how much of the Ubuntu release effort depends on volunteer effort and how much is corporate affiliated manpower?

    For example.. out of the set of Ubuntu core developers how many are affiliated with Canonical or another corporate partner like Dell versus strictly volunteer? How much of the core development work is being handled by these corporate affiliates? Similarly with MOTU how much of the activity inside the scope of MOTU is being done by corporate affiliated people?


  2. @Jef No, I can't think of a good way to discover whether someone is (perhaps unduly) influenced by an employer other than asking point-blank.

    In the past, people have asked for metrics regarding volunteers in the ~ubuntu-core-dev Launchpad group. All Canonical employees have an e-mail address.

    With respect to the release effort itself, for Ubuntu, it seems to be coordinated largely by Canonical staff. For Xubuntu, Edubuntu, Ubuntu Studio, and Kubuntu, it's coordinated largely by volunteers (although in some cases, Canonical employees spend off-the-clock time for these remixes). Clearly the greater distribution (consisting of Free/Open Source Software from Debian and many, many other upstreams as a result) benefits from both corporate-sponsored and volunteer efforts.