My Week at OSCON

At over 3000 attendees, O’Reilly’s Open Source Conference (OSCON) is the nation’s largest conference focusing on the open source ecosystem and yet it still manages to feel like a community event. Sure, there’s plenty of cold, hard tech on display–much of which currently centers on the latest and greatest in cloud computing–but what really gives the conference its energy is the commitment to and celebration of open technologies. And, of course, the community — which is really just another open source technology.

Being committed to open source means sharing ideas and products, but it also means recognizing that the fruits of our collective labor are only as compelling as the producers are diverse. While there is a lot of work to do before the open source community comes to more closely reflect society at large–or even the tech community at large–there seems to be growing recognition that diversity matters. In fact, if not the theme of this year’s OSCON (hardware may have taken those honors), diversity and inclusion certainly played a close second.

A number of talks and keynote addresses highlighted these topics, including an inspiring address by ElasticSearch’s Leslie Hawthorne called, “Checking Your Privilege: A How-To for Hard Things”. Also, while women made up just 8-9% of speakers at OSCON from 2007-2009 (according to the Geek Feminism Wiki), a(n admittedly unscientific) scan of this year’s schedule showed that rate jump to roughly 20%. Not perfect, but a nice bump. I noticed more women-focused booths in the Expo hall too, including ChickTech (a Portland-based mentorship community for girls 18 and under), Girl Develop It (an education network for women and friends), and, of course, Pyladies — an international organization for women Python programmers at all levels that I am involved with as the co-organizer of the Portland chapter.

This is the second year that I’ve run the Pyladies booth at OSCON, and returning this year gave me the chance to reflect on what we’ve accomplished as an organization over the last year: we’ve watched members successfully transition into technical careers and found start-ups; we partnered with PyCon organizers in outreach efforts that produced the highest percentage of talks by women (33%) in conference history, and raised $40,000 in grants to help women attend; and we nearly doubled Pyladies locations from 26 to ~45 by helping groups get started in Brazil, Argentina, Peru, Bangalore, Barcelona, Houston, Taiwan, and other cities around the world. It’s been a great year for us, and I was once again more than happy to be able to share successes, strategies, and resources with other technical women from all over the tech map.

My conference experience wasn’t all ‘work’, however; I was also lucky enough to catch a few of the talks. I learned about some of the possibilities inherent in using promises with generators per EcmaScript 6 (there’s a good article on native promises on HTML5 Rocks) in “Syncing Async”, and I picked up some nifty new command line tricks in “Advanced Bash Scripting.” I got a peek at the real-time events architecture the team at Uber developed using Node and Kafka, and I collected some solid teaching tips from two Oregon State University Open Source Lab professors in “DevOps for College Students.”

Some of my favorite talks this year, as last, were given by the Netflix team. The first, “Netflix API: Top 10 Lessons Learned”, stressed the importance of taking a pragmatic, rather than dogmatic, approach to development. This has lead their API team, among other things, to deal with the proliferation of devices and platforms by shuttling their resource-based API (where a request might look like: /movies/movie_category/:id) in favor of an experienced-based API (request looks like: /ps3/homescreen or /ipad/homescreen, etc). The second Netflix talk, “Big Data Pipeline and Analytics Platform Using NetflixOSS and Other Open Source Libraries”, detailed the evolution of data analysis from bare-bones past (using ‘awk’ and R) through to the development of a robust, real-time data pipeline and analytics engine that employs a suite of open sourced technologies (Hadoop, Druid, ElasticSearch, Kafka, and Netflix’s own OSS projects, including Suro). Most interesting here was the discussion around solving the need to handle ad-hoc queries of n complexity (ie, granular data on-demand) by using Druid, a cube-like database that the team rated ‘great for exploring highly dimensional data.’ Even though I’m not a data nerd per se, I found the talk accessible and left feeling like I learned a lot about many of the tools available for trapping and analyzing data at scale.

All in all, another great OSCON experience! If you’re enthusiastic about open source and have the chance to go in the future, I highly recommend it. And if you do, don't forget to swing by the Pyladies booth and say "hi"!