Very excited about going to New York April 24th and speaking at SQLSaturday. My girlfirend and I will go down on Thursday and watch some Broadway shows and visit a museum or 2.
The talk is '3rd Key Normal Form: That's crazy talk!!!' - a fellow employee has this on his whiteboard in his cube. In all my time as a developer and DBA, there is always some denormalization, or really no fully normalization to begin with.
The session is going to show some of the experiences I have encountered over the years. One is in the Laboratory Information System Management arena. My first job out of college was to help normalize data from a file system (written in Profressional Basic - DOS 3.3). We used FoxPro DOS to create a report writer for our clients. Many years later MS Access came along, and so did Visual Basic. I had some great ideas that the developers (Chemistists by training) did not want to implement.
The second example is a FEMA program for funding distribution and Performance Indicator tracking. This was a great example of compound primary keys geeting larger and larger. Two lessons here: Not every table is part of a hierachy and Parent-child can benefit from ID (identity field) primary keys.
If I have time, I think I want to get the audience into a 4-6th normal form discussion. Paul Nielson will be there, and it would be great to visit with him about database normalization.