Half way through the talks, sitting here feeling full and fat after my jumbo chicken burrito from Taco Del Mar (only place I could walk to without getting drenched en route).

As I mentioned before, I originally hoped to blog all through this continuously, but the aforementioned Murphy's Law screwed that all to hell.  No sooner had I hit the "submit" button on that first post after suggesting that things would go to pot that someone sat down next to me and spilled his coffee on me.  Missed the laptop, thankfully, but got me and my seat.  Had to relocate to an area that's turned out to be impossible to type from.  Good news, though, is that the hot coffee warmed me up after the cold, wet walk in... 

First up was a pretty good talk by Josh Berkus about the features in the upcoming 8.3 version.  She's in beta now with a hopefully release date of December 1st-ish.  That's assuming, of course, that some of us get off our lazy arses, download the current beta, and start hammering on it.  Go ahead.  Do that now.  I'll post more details about some of the features Josh talked about in a later article once I've put my notes together.

Second up was David Wheeler, giving a quick once over on Rails for PostgreSQL people.  My exposure to Ruby and Rails is all second-hand at this point (I worked some place that started using Ruby on Rails, but never had first-hand experience with it), so I'm not really the right person to give a detailed account of his talk.  In all honesty, the only thing I can really remember clearly right now was the line he gave when he was trying to take control of his session again during an ad hoc Q&A that began about stuffing methods: 

"The point is you can reopen the Rails classes and you can fuck 'em up."  I loved it.  Good form, Dave.

Next came Robert Hodges, CTO of Continuent.  Robert managed to walk the line between just giving his company's schpiel (they create middleware for clustering PostgreSQL and MySQL farms) and giving an informative talk on the hassles of scaling databases.  If his data is true (and considering his delivery I don't have any reason to think he'd intentionally mislead us), it looks like their uni/cluster product does a great job.  You don't get exactly linear improvement as you add nodes (in other words, adding a second server doesn't exactly double performance), but it looks like a good solution for companies that've outgrown their single server databases.  Again, more on this later.

Last before lunch was Neil Conway and his presentation on PostgreSQL query execution.  I really hate to admit this, but I didn't pay nearly enough attention to this one.  Right before lunch, I was cold and tired, and had just received a text message from my wife that our bloody roof (one we just had replaced entirely last year) was leaking again.  Suffice to say, I was a little distracted.  Sorry, Neil.  Good news, though, is that Neil's presentation should be thrown online in the near future, so we'll all be able to catch what we missed.

Oh, and I almost just had ANOTHER cup of coffee spilled on me.  Not my day...