Oscon 2007 Wednesday Wrap-up

apparently nothing eventful happened until Wednesday (or I can’t legally talk about what happened… you decide). Anyway, here’s yesterdays recap… First action item of the day was Berkus’s performance whack-a-mole talk. Normally a three hour tutorial, he tried to cover as much on tools, strategy, and examples as he could, though he had to skip a bit of stuff. The topic itself is an important one, I might try to crib his work and come up with a true 45 minute talk on the subject. after that there was nothing hot on the schedule, so I ducked into the exhibit hall to check out the pg booth and say hi to folks. My overall impression of the hall was it was low on schwag, and the t-shirt count is not looking too good, though there were at least some interesting things to look at (maybe shopzilla can help me pick an HDTV?). I did have a quick conversation with the SolidDB guys, and briefly discussed the troubles of an “outside” contributor to the mysql distribution process. after this I skipped the conference lunch, which I was told was a really high quality buffet, and went to chinatown for some veggie pizza (the bolier room iirc). yummy, and nice to see some more of portland, though I might stick around for todays buffet. when I got back, I swingged into [http://www.mysqlperformanceblog.com/ peter zietsiev’s] open source (by which we mean mysql of course) storage engine review. the talk focused on the status of mysql’s transactional storage engines (innodb, falcon, pbxt, and solid), on how they operate, how they perform, and what areas they try to target. There was a lot of good takeaway from this, but I think the most important thing was that none of the engines outside of innodb are really production ready at this point, mostly do to bugs and performance issues he ran into while trying to get them all running). istm this talk could be updagted and given again, if he does it I’ll volunteer to add postgresql into the talk if he wants the help. after lunch I went to the mysql->postgresql talk but kevin falcone. this talked focused on difference between mysql and postgresql and how to accomplish similar goals with the different tools. I thought the talk went pretty well, though he did mis-speak on a version number early on which i thought was going to send fetter into a rage, but luckily no trouble ensued. After the talk I chatted with Jay Pipes, and spun some ideas on making a general “how to do that over here” page, for people trying to work on the two systems; we’ll see where that goes. i took a break from db related stuff to swing into the maslows hierarchy for software talk. this was really well put together, done in a larry lessig style, and i’d recommend folks to go see it if it pops up on the conference circuit again. it’s a little touch to sum up the talk, but one key thing that jumped out (mostly because I had written an email earlier today about this) was the idea that people need to focus on how thier products does the users job better for the customers (rather than why the current solutions aren’t good enough). getting back to the technical, i swang into the oracle to mysql talk. this talk didn’t have a lot of people in it, i suppose due to a lack of oracle users maybe? my main take away was that it sure seems extremely painful to move from oracle to mysql, enough so we never could have moved [http://www.lethargy.org/~jesus/archives/84-SunOmniTI-PostgreSQL-Case-Study.html our TB ODS system] to it, and quite frankly I think anyone who does must be a glutton for punishment. I did find it funny that monty interjected a number of times who certain problems were either bugs that should be reported, or things fixed in cvs. it’s probably worth pointing out that the speaker noted that his company didnt choose mysql based on it’s suitability for the task; rather they tried oracle grid, realized they couldn’t afford it, and then moved to mysql because it had been used on a previous project. such is the way tech decisions are made. after the conference, we went to a couple of parties and dinner, of which I normally wouldn’t have much to say about, but I do have to make note that I got to play [http://en.wikipedia.org/wiki/Footbag hackey-sack] at the Sun party, so that’s worth something.