Question-shaped throwaway experiment

These examples and illustrative results are independently authored teaching materials, not measured model results.

Use case

Choose whether undo restores only your last edit or an entire object snapshot. Teaching state has title/flag: A, own edit to B, another action changes flag, then undo. An in-memory demo exposes semantics without a real database.

Mechanism

State question/options at the top and choose an artifact that reveals them. Use scratch state, simple actions and visible before/after fields for normal/conflicting cases. Compare treatment of the other action and record observations/decision. Mark throwaway; production design/checks follow separately.

Bad example

Build polished production UI/database without defining undo, inspect animations only, hide flag changes and claim semantics confirmed.

Good example

Create a teaching in-memory comparison of local undo versus snapshot restore. Show title/flag/action log after A→B, external flag change and undo. Let readers inspect whether others' flags are affected, then capture reasons/untested cases. Answer semantics only without real data or appearance as production-readiness evidence.

Why the change matters

The question shapes the prototype; visible state makes conflicts inspectable. Scratch implementation avoids unrelated deployment effort and a captured verdict survives disposal.

Observable expectation

Snapshot restore may also restore the flag; local undo changes title only. Show both without assuming one universally correct. Inspect the conflict path and observation-linked decision.

Hidden flags, normal-only paths or feels-good verdicts do not answer the question. No prototype is executed here.

Limits

A prototype establishes neither concurrency/durability nor production safety. Real semantics may differ. Source artifact formats are workflow choices; use the smallest useful authorized experiment.