A heuristic for deciding whether to repair or replace legacy software based on its historical design quality. If a system was well-designed originally, it likely contains encapsulated learning worth preserving and should be repaired. If it was poorly designed from the start, replacement is more attractive. The key insight is to build temporal models — understanding how a system evolved and why — rather than only structural snapshots of current state. This historical thinking reveals whether the system's purpose and constraints have shifted, and informs smarter decisions about technical debt repayment.
2 Impressions