Skip to main content
Ownership metrics describe who wrote a file. Number-of-authors is one of the most-validated defect signals in the literature, and Bird et al. showed that files with many minor contributors — people who wrote only a sliver of the code — have significantly more defects, while a strong top contributor is protective.

​
What mehen emits

​
Semantics

  • Authorship = added lines. Ownership and the minor-contributor classification count lines added per author across the walked history — matching PyDriller, the reference implementation. Deleting someone else’s code or renaming a file does not count as writing code, so a deletion-only toucher appears in history.authors but is never misclassified as a sub-5% “minor contributor” of code they didn’t write.
  • Identity is the author’s lower-cased email (falling back to the name for commits without one).
  • A history of pure renames and deletions has no added lines; ownership is then reported as 0 rather than dividing by zero.

​
How to read it

​
References

  • Bird, C., Nagappan, N., Murphy, B., Gall, H. & Devanbu, P. (2011). Don’t Touch My Code! Examining the Effects of Ownership on Software Quality. ESEC/FSE 2011.
  • PyDriller process metrics — contribution-based ownership and the 5% minor-contributor threshold.
  • code-maat authors, main-dev, entity-ownership (repo).

​
See also