Refactoring Databases: Evolutionary Database Design

by Scott W. Ambler

When I give evolutionary database design talks at conferences I like to start out by asking the question, "Who can go back to their office tomorrow, rename a column in the Customer table in their production database, and successfully deploy it into production by the end of the day without breaking anything?" Many people in the audience laugh at the idea, although one or two people will stick up their hands and claim they can do it. Invariably, they either work in a very small company with a handful of applications, they have fully encapsulated access to their database, and/or they take an agile approach to development. The rest of the audience, typically 98%-99%, believes that its data management (DM) groups can't fulfill this rather trivial request. Something is clearly wrong in these organizations, and I think it's time we start to address the problem.