A couple of reasons, really.

1. I needed an interesting project to experiment with architecture and paradigms.
2. A strong personal portfolio is great leverage when negotiating professional pay.
3. I've seen the diku source, and all due respect to the godparents, but it's not a codebase I want to work in. I also want to create a new engine that makes it easier to create extensions that tap into more modern technological advances, like story telling AI.