Why code flexibility should sometimes override domain model constraints in software design

Summary

The article argues against the popular software design principle "make invalid states unrepresentable," which advocates tightly binding code to domain models. The author contends that code should be more flexible than the domain model, allowing for evolution and adaptation. The principle typically involves enforcing single sources of truth in database schemas and using type systems to prevent illegal states. The author presents a controversial counter-perspective that rigid domain binding can lead to brittle, hard-to-change systems, and advocates for looser coupling between code and domain representations.