Allen White

TechEd Final Day - Friday

Well, we've finally reached the end of Tech Ed 2006. Last night's party at Fenway Park was great (except when the woman who got hurt during the Train concert, of course. Hopefully she's ok.) I called my Dad while standing in left field in front of the Green Monster - yep, had to do that. I also walked across the top of the Green Monster, and my wife took a picture of me in the visitor's dugout (where else would a Cleveland Indians fan be?)

This morning I focused on chalk-talk sessions in the TLC area, the first one by Durga Gudapati on SQL Everywhere. This looks like an interesting product, and a way to build small, very portable applications that can capture key transactional data. I can see some interesting possibilites with that product.

After that was a session on SQL Server Express, by Chad Hower. It's a fully functional version of SQL Server that has some memory, processor and database size limitations, but very capable for client-type applications. Chad had an interesting idea, though, in that it can be used for apps that can run entirely within the SQL Server engine, so that you send the SQL Server Express runtime and a prepopulated database to a client system and the application, by being built as a SQL CLR application, requires no other setup. Again, some interesting possiblities to think about.

This afternoon I attended Stephen Forte's presentation (DAT340) on Database Design Patterns. He talked about design patterns for OLTP databases, SCD (Slowly Changing Dimension) databases and DW (Data Warehouse) databases, and how each one of these types requires a different approach to the relational model. He firmly stated that these functions should never be mixed in a single database, and that one of the biggest mistakes application builders make is by doing reporting or analysis against an OLTP database, for example. I'm still a newbie when it comes to the BI space, but I see his point, and will be looking more closely at this in the near future.

Finally, the session I'd been waiting for all week, DAT342 - Programming SMO, presented by Richard Hundhausen. Before the presentation I had time to talk with Richard about a couple of the things I'd run across in SMO that I felt were lacking and, after doing a quick test and finding the same results I'd found, agreed. He's since put me in touch with the people at Microsoft who are responsible for SMO, and I'll be sharing my "pain points", as Microsoft is calling them now, with them, and we'll hopefully make some progress with that great product. Richard's presentation was well done, and he showed us an application he's written he called a "concordance" application, which evaluates columns within a database, shows their datatype and the tables they're present within, and red flags every instance where the same column is defined differently within a database. Very cool use of SMO.

This was a great conference, and I learned a lot, got a lot of cool ideas, and got to touch base with key people in the SQL Server world.

Allen