Testing Software Architecture
summary
Proposes treating architecture as a third test type — architectural tests — that measure local and global complexity, not just features or performance. Demonstrates using static analysis (NDepend) to compute cyclomatic complexity, coupling and public method count per build and to plot trends in CI so growing interfaces flag trouble early. Gives concrete contrasts (a commission module with low local complexity vs an ad/finance module that becomes opaque) and shows this kept a product viable in production for seven years, avoiding a big ball of mud.
AI-written summary · about
chapters (10)
- 0:00Introduction and welcome
- 1:50From legacy code to architecture lessons
- 7:31What is software architecture
- 15:57Modularity and system complexity
- 19:53Evolutionary architecture approach
- 23:04Testing architecture with feedback
- 29:08Automating architectural tests
- 34:45Evolution drivers for change
- 40:43Balancing complexity and evolution
- 42:07Book recommendations and closing
related talks
40mTesting Software Architecture
43mTesting Software Architectures
51mTesting Software Architecture
58mTesting Software Architecture
01h 15mHow to measure your architecture
01h 34mThe Intersection of Architecture and Implementation
20mSupporting your architecture with code coverage and static analysis rules
58m4 Key Trends Towards Better Software Architecture