When it comes to metaprogramming, even the most technical folks can find their eyes rolling back into their heads.  I’ll admit, with all of the fun and new technologies out there, a discussion about metaprogramming is, in the name of brutal honesty, just outright boring…  but, so is being addicted to brute force.

During a recent ETL implementation using T-SQL, changes in architecture and scope began to threaten the project’s delivery dates. I desperately wanted to begin development but feared throw away work resulting from requirement changes coming out of ongoing design and analysis discussions. How do you develop code without a finalized set of requirements and prevent re-writing each piece of code when requirements change?