I had a good time at OSB this year. But I enjoyed it more in past years, and I’d say that this year was good, but not great.
In terms of focus and information, OSB is geared for projects first and technology second. Most of the talks are about the psychology, logistics, mechanics, or organization of open-source projects, often from the perspective of a project organizer or lead developer. If you use (but don’t organize, run, or contributed often to…) open-source projects, you have to internally re-map these talks to benefit from them. There are some technology talks, and of course technology is also brought up in the project talks. But no talk delves into technology to any great degree.
The other (big) open-source conference, OSCON, is (I think) inversely geared. It’s for technology first and projects second. Most of its talks are about effective technology use and migration; a minority are about project management and organization.
The Unconference day…
Pro tip: For pure shock value, nothing can beat sitting in a talk and having it slowly dawn on you that the speaker married someone you knew in High School, eons and eons ago. Jeepers, that was odd.
In the morning, Puppet. I’ve absorbed enough Puppetry via osmosis that I’ll probably use it instead of Chef at IP Street, when the time is right.
Not sure about the afternoon sessions. Unconference session quality is more uneven than planned sessions, as is typically the case. The tradeoff is getting more variety and finding diamonds-in-the-rough. But these sessions aren’t ringing my bells.
And I’m out.
Open-source processes for security vulnerabilities: The speaker works for ISC, and the talk focused on a different software level than what I usually work on. But I still pulled some pearls from it, such as security bug classification, and security vulnerability terminology.
Privacy and Security: The speaker works for Mozilla Labs, to which I had once applied to for a job, and which eventually tossed application into a black hole. Heh, I won’t hold that against him. ZOMG, I just realized the speaker is married to a someone I knew in High School! A good summary of current challenges in user privacy controls within modern browsers.
Yesterday’s “Text Lacks Empathy” talk got me thinking about the times when I’ve put others on the spot.
Sometimes empathy is overrated. Sometimes it’s a waste of time, and sometimes it’s wrong. There’s a time and place to pin someone’s ears back against their head and clean their clock. It can be very productive to employ linguistic judo. And sometimes very cathartic.
In 2006, I had a billing problem with Countryside Pet Supply. They weren’t answering my e-mails, so I escalated the issue by sending a message to multiple countrysidepet.com addresses. A company employee did a reply+all to my mail. He asked another employee to deal with this, and referred to me as a “west coast jerk.” He didn’t realize he had done a reply+all. He thought he was communicating only within his company.
Not sure I’ll listen to today’s keynote. Update: Nope, I didn’t go to the keynote.
How not to Release Software: My warp engines are finally on-line. A good talk. It was oriented, not surprisingly given the conference at which it was being given, toward web sites/servers/applications. Holy crap, there are two Mozilla projects still maintained in CVS repositories!
At IP Street, we commit a few of the “sins” mentioned. Every system has tradeoffs, and a small start-up has all kinds of judgement calls of the form, “What’s the right amount of process for us to now have?” Here’s a hint: If you worry about implementing CMM levels, you’ll fail.
<Your Favorite Programming Language> Loses: Emacs used as a slide show! Some good ideas, some tedium, and an unsatisfying S:N ratio. This could have been compressed into half the time, or, this could have covered a lot more ground at the same detail, or, this could have drilled down deeper into each topic. The issues the speaker deemed important made me feel like I was taking a time warp back to 1994.
Yadda yadda yadda. Live post. You know the drill.
Nginx: I’ve been an Apache httpd user since FSM knows when. I’ve never been persuaded by artificial benchmarks. I don’t worry about small front-end inefficiencies when the application spends gobs of back-end time talking to other servers. It’s more important to use a front-end technology that is familiar and rock-solid, and httpd & mod_wsgi fit that bill. Still…I want to stay abreast of server zeitgeists.
This talk was thin. It had a lot of this and that, but no smoking httpd+mod_wsgi gun.
I’d like to see a no-nonsense informed comparison of Nginx vs. Apache httpd. This talk isn’t it. An inauspicious start was a six-year old quote about a lighttpd memory leak, which their bug database says wasn’t a bug at all.
I’ll do a live-post each day with bits about the talks and goings-ons.
I met Amye for Happy Hour at the hotel where I’m staying. Because I was in a “Mad Men” state of mind, I had two Old Fashioneds. I am the Don Draper of my world.
1530: Closing ceremony, conference feedback, etc. Lots of chatting. Spirits are high. Lots of commentary. I’m sure the Open Source Bridge staff will be sifting through them (and the online survey forms) for a few weeks.
1422: couchdbkit or couchdb-python?
1304: Pasta and meatballs that were quasi-Italian. Then back to CouchDB.
1205: I’ve successfully built and installed CouchDB. I rule.
1545: Keeping Agile at the Heart of the Internet,Larissa Shapiro. This turns out to be a Scrum overview. Not what I expected.
1430: Data Warehousing 101, Josh Berkus. “Big data” is relative to contemporary typical commercial data sizes. Archiving: WORN data. Usually needed for regulatory compliance, and has very liberal response time requirements. Data mining: You don’t know what’s in there, but you want to find out. Semi-structured, lots of data, data is produced as a side effect of other business processes. Analytics/BI/DSS/OLAP: Use large amounts of data to produce interesting and informative visualizations.
Dimension synonyms: Facets, taxonomy, view, secondary index. ETL: Extract, Transform, Load. This is the entire process of turning external raw data into useful database data. E.g., processing apache logs and storing the result into a web analytics db. Clean up, normalize attributes, de-duplicate data, etc.
1330: Technical Debt, Elizabeth Naramore. Code or practices that will hinder your future progress. Too much cripples the team. This is another talk where you’re very happy to not recognize your code or practices in anything that’s up on the screen. Hmm, Sonar.
1900: Guided Tour of IRC, Peter Fein. This promises to be a treat, listening to he of Telecomix renown. Hmm, should I use Colloquy or Snak on OS X? We venture onto the Telecomix IRC server. Then we venture onto AnonNet. Ooooo!
1545: Testing Antipatterns, Matt Robinson. Not running tests often enough. Not fixing broken tests before committing new code. Tests are too slow. Inaccurate mocking. Etc. This was one of those, “I’m glad I don’t see myself in these slides,” talks.
1430: The Big Data Exploratorium: Data Mining, from Patents to Memes, Devin Chalmers. (I had too many carbs at lunch, hoo boy.) A ton of great ideas here for IP Street, if only we had the resources to look into them.
1630: Rather than hitting the night’s parties and socials, I’m going back to my room to chill. Have to pace myself. :-)
1545: Massively Scaling Django for a Global Audience with Playdoh, Frederic Wenzel. The speaker works at Mozilla, which has more than 140 live websites. Lots of sites, lots of load, lots of scaling. About 35 web developers divided into seven teams to support this. This doesn’t include QA or IT. “Mozilla loves Django.” Playdoh = Django plus libraries plus docs. Playdoh is a grab-bag of low-level libraries.
Mozilla add-ons web site = 5 MM page views/day. Firefox input = 30 K page views/day. The smallest server setup: 1 load balancer, 2 web servers, 2 MySQL, 2 memcached servers. Performance, optimization, localization, security, and deployment tools. The Playdoh library suite looks very cool, I’ll keep it in mind for IP Street.
1430: A dozen databases in 45 minutes, Eric Redmond. In trying to make this topic more approachable, the speaker gives short shrift to ACID, BASE, the CAP theorem, and a host of other topics. This is a bad survey talk. The speaker seems very smart, so, maybe he’s just ill-prepared for his talk today.