architecture-design-contest

Architecture design-space exploration identifies multiple competing designs for a feature and compares their trade-offs to guide final selection. It is needed when you want alternatives before coding or require an RFC-level review to inform architectural direction.

To compare architecture trade-offs, frame the design problem with goals and constraints, then run parallel explorations against your codebase. This synthesizes diverse designs into an explicit comparison highlighting risks, trade-offs, and a final recommended path.

Yes, design-space exploration requires grounding in the current codebase and documentation. Multiple subagents explore the existing code in parallel to ensure the generated architecture options are viable within your specific technical environment.

The best way to evaluate competing architecture options is to run 2 or more parallel architecture explorations and synthesize the results. This delivers an explicit, opinionated recommendation that highlights trade-offs, risks, and a migration plan if needed.

This method provides a clear, opinionated recommendation, not just a list of alternatives. It frames the design problem with decision criteria, compares trade-offs across parallel explorations, and synthesizes a final recommended direction to guide implementation.

You should avoid parallel architecture exploration when your design problem lacks clear goals, constraints, or decision criteria, as the synthesis relies on framing the problem succinctly to generate viable competing designs and a justified recommendation.