Software Design X-Rays: Fix Technical Debt with Behavioral Code Analysis

Adam Tornhill

If you like the notes, go ahead and buy the book!

What do we want to achieve?
- Find Hot-Spots - places in the code with high “interest rate”,
- Hot-Spots are code fragments worth improving - fixing them means saved development time in the future
- the goal is to prioritize technical debt
How to find Hot-Spots?
- By Change Frequency (number of changes over time)
  - code changed too often signals possible problems with design
- By Complexity (number of LoC, indentations)
  - too large modules are hard/impossible to reason about (regardless of its change frequency)
- Combining above metrics, visualize it
- Apply Organization/Team/Project context
Hot Spot
- has low cohesion - contains several unrelated parts, lacks modularity
Change Coupling
- files (lines, components) which are changed within the same commit

Interesting insights