The applicability trap
summary
Component content systems that use transclusion and partial transclusion hide a poison pill: heavy conditional overhead can exceed any savings from deduplication. The concrete test is arithmetic — if a single conditionalized task’s byte size is larger than the sum of separate variant tasks you are in the trap. s1000d applicability and the super plane D example show lost change history, failed audit under 14 CFR, and costly cleanup when a variant is actually a new product.
AI-written summary · about
related talks
45mAbstraction Considered Harmful (and other opinions about writing maintainable code)
57mVariant analysis
56mMy Journey to the Dark Side: Under-Approximation, Incorrectness, Proof
01h 01mProgram Obesity: A Code Health Epidemic
01h 03mModern C++ Development: Limitations and Problems in std::function and Similar Constructs
21mIntegrating Dependency Building with Document Checking in Coq
42mSuggesting Valid Substitutions For Typed Holes
15mSyntactic Completions with Material Obligations