Practice of organizing tests into nested descriptive contexts with prose that reads as specification. Introduced to the Ruby world via RSpec. Ladd argues the real value is the reinforcement loop: agonizing over clear, concise technical language in context strings exposes when you don't actually understand what you're building, and the vocabulary you produce gives you a head start when documenting the software. Lost much of its value when RSpec defaulted to green-dot output — once nobody reads the descriptions, they stop being specification — which Ladd correlates with the emergent 'describe ClassName / describe "#method_name"' style that replaced real context specification. Test Bench prints context/test prose and data values on every run to preserve this feedback.