There’s a contest going on at PASS looking for answers to “Best Thing I Learned at PASS”. I don’t want to enter since I’m going and I don’t need to free admission to PASS, the hotel, etc, but it is a great question to answer. VP of PASS, Bill Graziano, wrote a blog for his answer, Thomas LaRock wrote one, and Andy Warren wrote a short note that he doesn’t have a “Best Thing”.
The contrast got me to stop and think for a few minutes.
The last few PASS Summits, Seattle '08, Denver '07, Seattle '06, Seattle '05, Dallas '04, have all been about business for me, not technology. I've been more concerned with meeting people, saying hi, having conversations that would inspire me, than really "learning" anything about SQL Server. Before the sale of the site, I spent most of my time meeting vendors, or working with people at PASS to arrange things for the next year than attending sessions. It was also one of the few times that I would get together with Andy and Brian to talk business.
At other conferences, and I'm not sure why, I've attended more sessions, almost "reporting" on them to people that couldn't come. I did that at Tech Ed a few years, with some long blog posts about information that was coming out about SQL Server 2005 or 2008. The same with the Business of Software and the Microsoft BI Conference, but PASS hasn't been like that.
When I think back to earlier PASS Summits, I tried to learn more, and I know that I attended lots of sessions. There are a couple that stand out, and which changed the way I looked at things, though not practically as I hadn't had the need to implement the knowledge.
Years ago there was a real low-level hardware guy that worked for one of the vendors. I can't remember which one, and I can't remember his name, but Andy and I ducked into a performance session on disk drives. He was showing the actual calculations of latency and seek time that could occur when you let a disk drive get more than half full as well as the issues with different RAID levels. It was fascinating, and while some of that knowledge has been rendered obsolete by SANs and large caches, it's still good to know. Disk drives fill from the outward in, so those outer tracks contain more data, require less head movement to get data read, etc.
There was also information about price/performance for various RAID array sizes based on the number of disks. Apparently beyond 7 or 8 disks in an array (R5), you lose value. Since no one recommends Raid 5 anymore for databases, this doesn't matter, but it was a very interesting session, one that I really enjoyed.
The other one was an identity debate that Brian Knight and I had one year. We argued back and forth, almost panel like, about the pros and cons of identity values. It was great fun and I think people enjoyed it. What stands out is a few people from Microsoft's IT group were in there and they told us that at the time (SQL 2000 timeframe), they avoided identities. They used GUIDs because above around 400 or 4000 inserts/sec they had issues actually allocating the identity values, something that didn't happen with NEWID() for some reason. I never had a system to test it out or worry about it, but that was fascinating to hear them talk about it.
At the first PASS conference, in 1999, I saw Kalen Delaney speak in a session on SQL Server 7. I’d been digging into SQL Server for about 7 or 8 years at that point, had started writing, and wasn’t sure how much I knew. Listening to Kalen’s presentation gave me a lot of confidence. I realized that here was a very well known author, standing room only in her presentation, and I knew pretty much everything that she talked about.
I think that was part of the reason that I started to write more and put more effort into sharing my own knowledge and experiences with SQL Server. That eventually led to SQLServerCentral.
So you might say the best thing I learned at PASS was something about myself.
I don't know what, if anything, I'll learn this year, but I definitely look forward to attending and meeting more people.