Debugging

Trying to quantify or measure debugging seems incredibly hard.
- First attempt, relying on information theory: https://twitter.com/kunalbhalla/status/1497610926803107853
- This works intuitively, but doesn't seem meaningfully automatable.
Papers I was pointed to
- Test suite diagnosability metric: https://ieeexplore.ieee.org/document/7985702
- Amy J. Ko
  - https://faculty.washington.edu/ajko/papers/Ko2004Whyline.pdf
Designing the Whyline
- Lifecycle
  - Hypothesis
  - Observation
  - Restructuring data
  - Exploring restructured data
  - Diagnosis
  - Repairing