I'm back, having made the annual pilgramage to Ottawa for this year's Linux Symposium. I went to present about Fedora's MirrorManager application I've been writing over the past 18 months, to hear about new developments across Linux, and of course, the popular hallway track. The conference organizers did a great job this year adding the late-night speakers cruise on the river. The Whiskey BOF was truly enjoyable if not prophetic, held at a nearby government bomb shelter, built over 3 years to withstand a 5 metaton bomb blast, but was rendered obsolete a year later with the advent of a 50 metaton bomb. Sometimes I feel like I'm on that technology trajectory with the fast pace of development around me!
Fedora had a big presence this year, with at least 10 presentations, tutorials, or Birds of a Feather sessions. My favorite was Chris Tyler's talk on the college course he developed which involves college seniors in Open Source and Free Software projects. A distinct change from the "waterfall" method I was taught in college (which even then felt pretty dated and bureaucratic), students learn how to interact with a community, find a relevant and achievable way to contribute to their project in one semester, and establish themselves as community members and participants.
I shot video of 5 of the sessions. Soon to hit Fedora TV's RSS feed are:
- Getting the Bits Out: Fedora's MirrorManager (me)
- Secondary Arches, enabling Fedora to run everywhere (Dennis Gilmore)
- A Model for Sustainable Student Involvement in Community Open Source (Chris Tyler)
- Creating your own (Fedora) Distribution (Jesse Keating)
- RHEL and Fedora kickstart file generation on the fly (Scott Moser)
I look forward to the Firefox 3.1 release that includes native OGG video support so these are easily watchable everywhere and without encumberance.
Mark Shuttleworth gave an interesting keynote speech, advocating for the coordination and synchronization of the release schedules of more parts of the Linux ecosystem. I do believe there is benefit in having a regular cadence to releases (such as the regular 6 month schedule for consumer / desktop distributions). I think some alignment of Enterprise-class distributions could benefit the larger ecosystem as well, at a much slower cadence (~2-3 years). Third party applications, those not included with the Linux distributions themselves, often look to leverage their Linux development investments so as to be useful on several distributions. The Linux Standard Base has long advocated another method to provide this stability (feature-based instead of time-based), but being asynchronous to any distribution's release schedule, has not achieved all it has hoped. At the same time, it would be difficult for computer manufacturers, such as Dell, to complete development, testing, and release readiness for multiple Enterprise-class Linux operating systems simultaneously. I will be interested to participate in the ongoing conversation about Mark's proposal.
Next year we'll all be exploring Montreal, I look forward to seeing you there.