Framework first, or how I buried a project

Since things had gone well in Delphi, I figured I'd do it the same way, write a framework first and the system would then practically build itself. And that's where I failed. The framework kept growing, the system didn't, and I kept adding things nobody needed yet. In the end I had a framework that could do almost anything, except cost research studies.

Meanwhile my manager changed. We agreed to bury the project and do it properly, in Ruby and Merb. It worked out.

The lesson even has a name: YAGNI, you aren't gonna need it, so don't write something nobody needs yet.