prototyping

Match fidelity to the question being asked: paper sketches answer structural questions, wireframes answer layout questions, clickable prototypes answer interaction questions, and coded prototypes answer performance questions. Always pick the cheapest fidelity that resolves your specific question.

Throwaway prototypes are built to answer one specific question and then discarded, such as a performance benchmark. Evolutionary prototypes incrementally grow into the final product, like an MVP that matures through continuous user feedback.

Skip prototyping for trivial, well-understood decisions like standard CRUD form layouts, and for safety-critical final stages where formal verification matters more. Also avoid endless prototyping that delays shipping without answering new questions.

Throwaway prototypes often hit the scaling-and-integration problem: something that works in isolation fails under real load or integration. Concept prototypes suffer the artificial reality problem, where polished mockups look functional but hide unproven assumptions.

State explicitly what the prototype proves and does not prove before showing it. Communicate the fidelity level clearly so stakeholders understand a polished mockup demonstrates structure, not production readiness.