Measure and evaluate a renderer-exclusive slim distribution

[woksin]

Problem

Selecting MUI or PrimeReact changes declared slots but does not remove the built-in React Aria implementation from the @cratis/components installation. Components 4 accepts this to preserve a zero-configuration fallback and avoid a premature package split.

A future renderer-exclusive distribution could reduce install cost, but only if measured consumer demand and package overhead justify the permanent release/identity burden.

Direction

Measure before designing:

• install size and cold-install time;
• production bundle closure for representative Foundation/Advanced profiles;
• duplicated React Aria usage in applications that directly use the same packages;
• SSR/runtime memory and startup;
• independent release-cadence pressure;
• migration and provider-identity costs of an additional distribution.

If justified, add a new slim package behind unchanged component subpath contracts. Do not reshape @cratis/components or remove its fallback in place.

Acceptance criteria

• At least two representative, anonymized capability profiles provide measurements and demand.
• A package split meets the documented split criteria in Documentation/ui-foundation.md.
• Provider/context identity, CSS, types, and adapter compatibility are proven across full and slim distributions.
• Existing @cratis/components imports and zero-config behavior remain source-compatible.
• If measurements do not justify the split, close with the evidence rather than implementing one.