Much like we write tests to assert that the code is really working the way it should, Nathaniel Talbott thinks we should be able to write experiments to provide us with facts that assert the usage of a feature of software is really valid. This is a novel concept known as experiment driven design, and he discusses it briefly in this interview over at InfoQ. It’s a common occurrence that customer representatives in a software project sometimes state with absolute certainty the importance of a feature when they really have no facts to back those statements up. EDD is meant to give customer reps and developers a common set of tools to measure the importance of usage of features before taking the time to fully build them. Assaf Arkin over at labnotes.org created a prototype framework for creating EDD experiments. You can check it out here. This implementation is currently in ruby, but I can see the .NET community adopting it (as they do with so many other open source frameworks) in the near future. However you can start using the framework to validate business assertions today even if the implementation of your project is in .NET, the experiments will just be in ruby.