dandi8

A while ago I wrote an extensible dummy data generator for Java.

I needed to fake some scientific data for a project at work and wasn’t satisfied with how closed for modification existing data generation solutions were, so I decided to tackle writing a library on my own.

It was my first major contribution to open source and had some architectural challenges which were fun to solve, not to mention the learning experience :)

IIRC from the original report, the claim here is that even “gaming” these metrics leads to the desired result, as you can’t game these metrics without actually improving your processes. I tend to agree.