For each file: - creation date - last modified date - modification frequency Then derive: - “ancient but volatile” - “young and heavily modified” These correlate strongly with defect density.
For each file:
Then derive:
These correlate strongly with defect density.