Dev Log: When the design contradicts itself, stop typing

Short summary

The author revived a C# data generation library and discovered design contradictions between general matching rules and special-case exceptions for certain types. Rather than implement the quicker fix, they chose a UX-first approach: preventing plausible-but-wrong synthetic data (e.g., 'Toyota' in a printer's Make field) through conformance testing. The reflection illustrates a core design lesson—protecting users from confidently incorrect outputs matters more than maximizing inference success—and introduces a pattern to catch future catalog drift before it ships.

• •Discovered design contradiction between general rules and special-case exceptions in a data generation library
• •Prioritized UX over implementation speed: prevent plausible errors rather than maximize inference success
• •Implemented conformance tests to lock catalog quality and prevent future drift

Generated with AI, which can make mistakes.