Measure, don't guess! It was the bottom line in the awesome talk by Kirk Pepperdine and Heinz Kabutz on Java Performance and Tuning, or 'Debunking Speculative Tuning' as they tend to refer to it themselves. Kirk and Heinz did a great university session in which they compressed a lot of the contents of their 4-day performance and tuning masterclass. On top of that they even came up with a specially prepared Java EE application that featured some complex performance problems. The aim of the university session was to guide the audience through the different aspects of performance and tuning by means of measuring and hypothesis. As the session progressed we came closer to the cause of the problem and at the end we got it all figured out. I guess this is what the ideal university session should be. Perfect? Well almost…Kirk and Heinz depended on the JavaPolis wifi in order for the audience to download the application and tooling required for the hands-on part. Unfortunatelly the JavaPolis wifi network is total crap. I'm sorry, but I don't know a better way to describe it. It is simply impossible to get a decent connection and when you have one, there is no internet connectivity whatsoever, and after about 10 seconds the connection is already gone because the el cheapo access point died in action when being DDOS-ed by over a thousand other frustrated JavaPolis attendees that also would like to have a wireless connection. Alright, enough about this. While Heinz did the talking Kirk was all over the place with some borrowed USB sticks and finally everybody got their software.
After this exciting session it was time for lunch break and a strawl around the pavilion booths that opened up this morning. As I expected the booths take up a lot of space and it was now far more crowded than yesterday. Not in a bad way though. Let's see what tomorrow brings when the conference part of JavaPolis starts.
While I was eating my lunch I had to decide on the afternoon university that I would go and see. My first intention was to go to the Web 2.0 talk by Greg Murray, but as I read the abstract again I anticipated another jMaki and mashup rant. Been there, done that. I've seen pretty much all of Greg's talks on JavaOne, so I already got my share of it. Also, the positive feedback from my coworkers that went to the first part of the Swing talk this morning persuaded me to pick the Swing university part 2 as my afternoon session. Although it started off with some basic ground it was a good session with some golden oldies, some stuff I'd already seen at JavaOne and some new stuff regarding the roadmap to Java 7. Good to see there is plenty of exciting stuff happening in the field of Swing. So in the end I was quite happy with my pick. This actually goes for all the university sessions that I attended in these first two days of JavaPolis. As far as the university part of JavaPolis is concerned, nothing but compliments to the JavaPolis team for organizing great content, great speakers and a very nice atmosphere to experience it all. Too bad the wifi sucks.