prototypelisted

# Prototype ## The Core Idea Prototypes are the thinking stage, not the shipping stage. Before you commit to a design—before you write the real code— you ask the hard question by building and running something small. The single clearest question shapes everything: what you build, how long you spend, where you put your energy. Two very different questions need very different artifacts. Mixing them guarantees confusion and wasted effort. --- ## Step 1 — Identify Your Question What are you actually trying to learn? | If you're asking... | Reason | Build this | Output | |---|---|---|---| | **Can this state model actually handle all the cases?**<br>Can we represent X? Does this data structure break at edge cases? | You can't reason about complexity on paper. Complex state machines need to be walked through by hand. | Interactive terminal app<br>(state explorer) | You push buttons, watch state change, discover what doesn't work | | **How should users experience this?**<br>Which layout works? What interaction feels right? Should this be a modal or a page? | You can't choose between designs by describing them. Visual comparison is 10× faster than debate. | Multiple UI variants<br>(3-5 designs, switchable) | You click to switch between designs, see trade-offs instantly | | **Both at once** | ⚠️ **Stop here** | Split into two prototypes | One per question, one per artifact type | **Ambiguous?** Default to the question type that matches your surrounding code: - Backend module →