I'm writing this at 39K feet on a United 757 flying home to Denver. This week was SQLTeach
in Vancouver followed by the PASS Board meeting in Redmond. SQLTeach is a smaller (400-500 attendees fits in a hotel) developer/SQL conference organized by Jean-Rene that rotates across Canada (Montreal - Vancouver - Toronto). This is my third SQL Teach and it's a always a blast - it's good to see face-to-face friends with whom I usually see only by email or the blogosphere - Adam Machanic, Peter DeBetta, Jean-Rene, and to make new Canadian friends.
My week started with my Smart Database Design pre-con on Monday. I love presenting this topic because we get into so many good conversations about database design and best practices. It's my contention that the database schema design is the primary factor in the success of the database and Smart Database Design walks through the database layers to explain how to design a database that performs great with equal emphasis on extensibility, data integrity, and usability. The crowd in Vancouver was engaged and intelligent, so thanks to all who attended. Bill Vaughn also presented a pre-con on Monday and it's always a pleasure to say hello to Bill.
Tuesday was a pure treat for me. Stephen Forte presented his data architecture topic: Database Design Patterns. By patterns he doesn't mean one-to-many normalization, he's referring to three uber-patters: Multi-user highly concurrent transactional pattern (OLTP), slowly changing dimension pattern (OLAP) and data warehouse pattern. Stephen's a great presenter; he has a wealth of experience to draw from and a wise perspective on our industry. Stephen has a Masters in History and told the how the term patterns came to be. Following the Great Patriotic War (WWII for non-Russians) when the architects went to rebuild Europe they wanted to plan the cities using urban planning and patterns for the new rebuilt cities the citizens didn't want a different type of city so those patterns migrated to America and the Latin American cities (which is why so many large cities have the same drab concrete look). That's the source of the usage of the word patterns as it applies to engineering.
Interestingly, Stephen relates data architecture to urban planning. According to Stephen, he seldom actually designs the data model. He designs the concept of what types of patterns will be stored on what type of technology and the overall flow of the application. I tend to think that the data architect should be more intimately involved with the data model, but I respect Stephen's point of view as well.
Wednesday I presented two topics - Top 10 Worst Database Design and Development Blunders. OK, I had 15. For each blunder, I told the story and as a group we discussed what they were thinking, which database attribute they risked (usability, data integrity, performance, extensibility, availability, security), and what they should have done instead. Using this story-based method of discussing best practices turned out to be even more fun that I had hoped. My goal was that by telling the stories, everyone has an anchor to remember to evaluate design and development choices by their impact on the database attributes. I think everyone in the room had an opinion and nearly everyone had a chance to share it.
Unfortunately for me, Stephen Forte was presenting XQuery Deep Dive across the hall. While I had a great time, half of me wanted to be listening to his presentation.
I heard a great story from a Canadian after the session. It seems that a Canadian province recently implemented a new parking ticket database that included a replicated database in every town throughout the province. The poor chap with the license plate "ABC-123" ended up with 600 parking tickets. The courts wouldn't believe him that he couldn't possible have all these tickets. "If it's in the database it has to be right". (Assuming a James Bond imitation), "Hello, my name is Data, Test Data."
While the 10 Worst Blunders session packed the room, my Nordic 2.0 (O/R dbms in T-SQL) launch drew a smaller crowd. Even so, it was good to walk through the new technology with friends.
A late flight Wednesday to Sea-Tac and cab ride to Redmond brought me to the PASS Board meeting. (The board meets 4 times a year and Microsoft or CA often host the meeting to cut costs.) This was Kevin Kline's last Board meeting as president and Joe Webb's final Board meeting. Both have put in an incredible amount of work for the community and we owe them a huge Thank You. To untangle PASS from our previous management company and move PASS to the new management company (C&C) tripled the amount of work done by the Board, especially Kevin, Joe and Wayne for the past couple of years, but the result was definitely worth it in the long run. Having seen Kevin and Joe pull this rabbit out a hat while maintaining their cool, I have the highest possible respect for them both.
As Wayne Synder takes up the reign as the new PASS President, I like his enthusiasm, his drive for transparency, and his desire to see everyone with PASS succeed. I remember when I first saw Wayne and he was bellowing from the stage how much he loved PASS week. It also seems that the level of cooperation between Microsoft and PASS is growing exponentially and the early plans for putting the WOW in the 2008 PASS Summit will make it the most exciting, blast of a conference yet. If you love SQL Server you won't want to miss the Summit in 2008.
The PASS Board is a working board with each board member assigned to a portfolio. I'm moving from the Chapters portfolio to Special Projects, and while I'm disappointed to leave chapters, the good Dr. Greg Low from SQL Down Under is stepping up to serve chapters. Greg's a great guy and I'm sure he'll bring good things to the chapter community. But I'm not really leaving chapters, because Wayne's first special project for me is to package up the Colorado PASSCamp idea into a PASSCamp in a box, to give chapter leader the templates, checklists, and steps to pull off a great local PASSCamp event (PASSCamp is a code camp type event optimized for SQL Server).
We're about to start our decent into Denver and they're calling for turbulence, so I'll wrap this up and post it later tonight; a great week, but I'm glad to be going home.