A new day at SDWest.

Well, today we managed to really attend sessions. There were some work things happening that really just prevented us from providing our full attention, but given that the topic of conversation was also the money source for our time, the only real problem I see is the expense of being in Santa Clara. That cost to work is much less than the cost of two days of work.

Anyway, that being said, yesterday I did attend Native XML Databases by Elliotte Rusty Harold. His presentation was pretty informative, but I really cared the most about the first half. Since, I knew nothing about XML databases and only a little more than that about relational databases, I got a lot out of it. I'll have to go back to work and re-evaluate some of the database choices I made now that I've learned a little more. Maybe, as I write more of this stuff, I'll even be able to communicate more of it back to you.

Today, I attended four whole sessions.

The first was building RESTful web services with Axis2/Java. I think I was doing fine with it at first, but I decided to sit too far from the front to read all of the slides and I can't find the presenter's slides anywhere. I'll have to update the post later with the new information.

The second was building a multi-threaded server in Java with NIO. Ron Hitchens has posted the slides and source code at the site for his book. He had a very informative presentation I thought. Ron talked about using selectors, which is something I've never taken the time to think about. I was going to spend the time but when I figured out that there was no Multicast datagram channel, I decided not to spend any more time on it. Although some of the audience sort of seemed to miss the point as far as what the topic of his talk was about. They were asking questions as if he was teaching them how to write a web server and they wanted him to tell them exactly what to write. Some the questions just seemed a little out of place. Of course, ironically, Paul Tyma was giving a presentation in the afternoon that was essentially just the opposite of Ron's.

The third presentation was by Brian Goetz and was titled Effective Java Concurrency. Based primarily on some of his Java memory model work and his book. He had lots of good information, but I suspect that if I finish reading his book, I'll get to that information again.
Some notes I took:
- Allocation is cheap - use it unless the objects are very large
- final is the new private
- Make objects immutable if at all possible
- Document as clearly as possible whether an object is thread safe or not
- Try to eliminate or shorten stretches of code that can only execute serially or basically, make synchronization code blocks as short as possible
Also, Brian, if you read this, sorry we were bad people and made so much noise at the end. We were just excited to move on to the puzzlers.

The fourth presentation was Java Puzzlers with Neal Gafter and Joshua Bloch. That was a very entertaining presentation and I think the best that can be said is to read Neal's blog. He posts puzzlers regularly.