design-an-interface

Generating multiple interface designs forces you to explore divergent API shapes before committing. It compares competing designs on simplicity, efficiency, and trade-offs, preventing you from committing to the wrong module shape early when requirements are fuzzy.

The best way to compare different software design options is to generate multiple interface designs in parallel and evaluate them against structured criteria like simplicity, implementation efficiency, and ease of correct use before making a final synthesis recommendation.

When requirements are fuzzy, gather caller operations and constraints, then generate 3+ competing interface designs with structured outputs including signatures, usage examples, internals, and trade-offs to clarify the best module shape.

Yes, you can evaluate trade-offs across competing paradigms by generating multiple radically different API shapes simultaneously. Each design includes a structured comparison of simplicity, general-purpose versus specialized fit, and implementation depth.

Designing an interface twice means creating multiple radically different API shapes to compare alternatives rather than iterating on a single idea. Parallel sub-agent runs generate these competing designs simultaneously to explore divergent module shapes efficiently.