Skip to main content
A hotspot is a file where complexity and change frequency overlap: complicated code that the team also touches all the time. That intersection is the highest-leverage refactoring target — CodeScene reports that top hotspots typically occupy ~5% of a codebase yet absorb ~18% of development effort and ~23% of fixed defects. Complexity alone over-flags stable legacy code nobody touches; frequency alone flags churning-but-trivial files; the product flags what actually costs money.

​
What mehen emits

history.hotspot is one of the default mehen diff columns.

​
Better than the classic form

The open hotspot definition from the literature uses LOC as its complexity proxy because that’s what a history-mining tool can compute without parsing code. mehen has real parsers, so it upgrades the proxy to cognitive complexity summed over the file — a measure of how hard the code actually is to understand, not merely how long it is. The complexity input is family-aware: SQL files use sql.cognitive_complexity and Markdown files use their own cognitive-load measure, so hotspots aren’t degenerate outside the shared source-code suite. Note that CodeScene’s prioritization and ranking layer on top of this signal is proprietary and probabilistic; mehen deliberately implements only the open, reproducible complexity-times-frequency overlap.

​
How to read it

​
References

​
See also