Day three of OSCON kicked off with several keynotes. Highlights for me included Jay Parikh of Facebook presenting on the growth of the Open Compute Project. I learned about this project when it was much younger and fully populated by Facebook engineers, so it’s was great to learn that they’ve since grown considerably and now have several companies involved and have become close to having a completely open stack for the datacenter. Carin Meier had perhaps the most entertaining keynote of all with a live demonstration of a flying robot and a hacked roomba, I uploaded a video of the flying robot doing a flip here: Flying robot #oscon. Finally, I enjoyed Mark Shuttleworth’s keynote Redefining What’s Possible. Ubuntu has been heavily focused on convergence with the recent focus on tablets and phones, and with the announcement of the Ubuntu Edge campaign Mark was able to articulate the goal of Canonical to disrupt the mobile market with this crowd-funded phone option.

After keynotes, it was off to sessions! First one I went to was More Instantly Better Vim where Damian Conway presented an array of awesome hacks for getting vim to work the way you want it to. Tips ranged from managing vim swap files more effectively to tips for handling column width. I highly recommend downloading his slides and scripts here: Instantly_Better_Vim_2013.tar.gz – even if it’s just for SWTC.vim, demonstrated here:

I then attended Jon Roberts’ Using Open Source in the Classroom Every Single Day session. My involvement with Partimus made this one particularly interesting to me. Roberts works for an alternative high school and has pretty much worked autonomously to get Linux and Open Source into his classrooms, going as far as keeping his Linux machines off the school network so not to raise the hackles of the school’s IT staff. I think what was most interesting about his talk was the realization that he (and other teachers I’ve encountered who do this), are doing precisely the same thing that people in companies were doing 10 years ago to bring Open Source into their environment. Businesses have largely caught up and now appreciate the value of Open Source, I hope schools get there soon too. During his talk he also demonstrated the use of several KDE Edu tools that he says have been improving quickly these past couple years, great to hear! Finally, it was cool to learn that he came up with a non-boring way to talk teenagers about computer history: by giving the history of video games instead.

After lunch I went to Rikki Endsley’s talk on How to Recruit, Hire, and Retain a Diverse Team where she focused on how outreach is done (are you going where the candidates hang out?), language used in job advertisements (“rock star” and “ninja” were among the non-collaborative, off-putting terms for many folks) and perks offered (beer and foosball do not appeal to everyone). I was already pretty familiar with these strategies for improving outreach, but unfortunately these key points from Rikki’s talk haven’t made it very far in the industry, we got to walk out of her talk to a job board looking for ninjas and rock stars. She also gave a bunch of great tips for improving retention, including appropriate onboarding procedures, working to reduce burnout and motivation check-ins with employees.

From there I joined my colleagues in Tom Fifield’s Planning your OpenStack Cloud where he gave an overview of considerations when you’re looking to deploy an OpenStack-based cloud, including storage, networking and more. I then went to Google’s session on Automated Testing For Accessibility. Unfortunately they assumed the audience was familiar with accessibility tools going in, so I was jotting down notes of “things to look up later” while trying to keep up with the presentation itself, including ARIA (Accessible Rich Internet Applications) and WCAG 2.0 (Web Content Accessibility Guidelines). It was cool to learn about the Chrome Accessibility Developer Tools (and extension) and they had folks from the Capybara-accessible project join them to talk about how they’ve been using those developer tools inside their automated integration testing, cool! I have to admit all my sites are pretty inaccessible so I may have to give these tools a spin in the near future to have my sites at least become marginally better.

Last talk of the day was the one where I probably learned most to help me in my day to day work: Demystifying SELinux: WTF is it saying? by Dave Quigley. Up until about 4 weeks ago I had never worked with SELinux at all, and boy was I in for a world of surprise and pain when I found myself on a CentOS box faced with SELinux issues. I was able to hack my way through the problems in several hours and get my results pushed into puppet so things worked, but it was a trial and I was never fully certain I was using the right commands properly. This talk was an excellent overview of the basics that I sorely needed and I walked out feeling about ten times more comfortable with SELinux than when I walked in. This slide was particularly demystifying:

This and other slides linked here. It was also super useful to learn of the “-Z” flag in ls, ps and netstat commands to actually learn what was going on, and the great ausearch tool for logs. I was happy to learn that I had been using restorecon, semanage and setsebool properly. Finally, there are a lot of applications these days that come with an selinux specific man page reachable by: man servicename_selinux (ie `man httpd_selinux`), sweet!

By this point in the day my notes were getting messy and I was pretty tired. The plan had been to go out to one of the OSCON parties but with my own talk coming up Thursday and the realization that I had just wrapped up my 5th day of conferencing I made the wise decision to stay in to finish prepping for my talk and get some rest instead.