The availability heuristic: the last disaster defines the architecture

Whatever comes to mind easily feels more likely and more important. After a security incident, every decision is suddenly viewed through the lens of security. After a performance problem, caching appears everywhere. After a failed framework migration, every new dependency feels dangerous. After a successful project, its architecture becomes the supposed recipe for every project that follows.

Experience consists largely of remembered individual cases. That makes it valuable, but not automatically representative. A vivid story is not the same thing as statistics. So I try to distinguish:

• Have I recognised a recurring pattern here?
• Or do I simply remember a previous success or failure particularly vividly?