LCA2008 Days Four and Five

Thursday

The morning started with Stormy Peters' keynote on Would you do it again for free?. It was a thought-provoking keynote on the motivation behind OSS development. Do open-source developers lose motivation once companies pay them? Once they leave that job, do they continue doing open-source development? Stormy suggests they don't lose heart, but suggests that hackers may move to another of their itch projects if they no longer are paid to hack on the original.

Then it was onto Clinton Roy to talk about Antlr. This was useful for the fact that using YACC and Bison can be black magic, so having another tool in the toolbox is great. Would have been better if we'd gone as far as generating a grammer and hooking it up to show the ease of building something real end-to-end. Tutorials are best if they kick-start usage.

After lunch was the usual high-quality Tridge talk on ctdb. Interesting to see the scaling of Samba instances being almost linear after this work, as well as various tridge hack discussions.

Next was Leslie Hawthorn on the Google Summer of Code program. While being an interesting talk in and of itself, what I really liked hearing about were the developers who were bootstrapped by the program continuing into open-source contribution afterwards. This is a wonderful outcome for OSS. Thanks Google!

Next was Michael Smith with Gstreamer followed by Dtrace by Peter Karlsson. This last one was more of a vendor talk, and would have been better if the presenter didn't spend 20 minutes trying to get video output working on a Solaris/MacBookPro combination, followed by failing to get OpenOffice to display his slides full-screen. That and the 15+ typos in the slides didn't reflect well on Sun or on Dtrace. The talk had so much potential, but the most useful part was the questions by SystemTap developers in the audience - hopefully both projects will benefit was better understanding of each other.

Dinner with Mikal and Steve tonight out at some random Italian place - quite nice food to go with the excellent company. Of course, after dinner on Lygon Street there is no better thing to do than to follow up with sorbet from Casa Del Gelato ice-creamery. Walking down Swanston found us Andrew Chalmers with friends Sam(??) and Amanda. Good hilarity followed, along with a kind offer from Sam and Amanda to bed down on their lounge room floor to save me from the backpackers. Some people are just genuinely kind - I'd never met them before, so their offer was very unexpected. Thank you for your offer and all the best for your trip Europe bound soon.

Friday

Wonderful keynote by Anthony Baxter on Python 3. Anthony presents well, and the confirmation of a supported 2.X series during the early days of 3.X deployment was appreciated. Glad to see that the Python project is taking the opportunity to clean up the inevitable cruft that accumulates in the language and reduce it so hopefully it stays small. Of course I'm biased with python being my favourite language right now, but I found this the best of the keynotes this year.

I went to Ralph Giles talk on Ogg Design Internals, followed by Paul McKenney on efforts to put concurrency into C and C++. Paul is a clever cookie, and did a great job of explaining how the C/C++ standards don't guarantee sequential execution as most people subconsciously assume, with the result being pain and suffering where you don't expect. Fortunately Paul has been there at the standards meetings representing our community's interests well.

That brings us to the afternoon with Keith Packard talking on fixing various X limitations. He started the talk with an Intel announcement surrounding their release of full developer documatation for the Intel 965. Wonderful news! Should mean that advancement on that platform should skyrocket even further, and hopefully spur the other graphics chip manufacturers into providing better support for the developers who generate end-user sales! But I'm preaching to the choir now...

Last talk of the conference for me was Eric de Castro Lopo on Library API Design. Another important topic, but I would have loved more depth.

Moving on, we went to lightening talks and conference close. We had the usual bag of both ok and excellent lightening talks - some of the Gaming Miniconf lightening talks could have been repeated to the larger audience to great effect. Bling bling bling. Notable was Pia announcing OLPC Australia, hopefully making a difference here with that excellent project.

Conference close was a time to reflect on all the cool stuff that the week contained, and gave the chance to wrap up all the loose ends. While Open Day follows on Saturday, the official conference close is Friday afternoon. We also got to meet Ben - and found out that the conference is going to Tasmania for the first time! Cool! Less than 400 sleeps to go now :)

Jeff Waugh went on to announce the transfogrimifation of the Rusty Wrench Award into (hopefully I get this right) the Australian Open Source Awards. It'll be great if the whole community get behind this and it becomes something truely representative of our community here in Australia. Website launch occuring soon.

Finally was the Google Conference Party. A traditional bbq stand-around and chew the fat s'more with conference attendees. A good time was had by all - thankyou Google and Leslie Hawthorn for organising it. Finally two old and tired previous conf organisers wandered back to Casa Del Gelato for that last sorbet of the week, contemplating with great expectations what Tasmania will do with linux.conf.au 2009 next year.

And for the absolutely final thought, a big say thank you to Donna and her team for putting on a very solid and fun conference! Thank you LCA2008 organising team! We know it's been hard-work, but we appreciate your sacrifices to serve the Australian Linux, Free and Open-Source Software community!

| 02 Feb 2008 | #