; caching.spl — Caching theory with intentional conflict ; Reasoning about cache validity, backend selection, and trade-offs. ; Run: zetl reason conflicts -d .
; Cache is mtime-based (from architecture/Cache.md) (given mtime-based-cache) (given incremental-rebuild) (given disposable-cache)
; — Defeasible rules: cache coherence —
; Merkle hashing gives reliable change detection even when mtimes are stale (normally r-merkle-over-mtime (and mtime-based-cache incremental-rebuild) merkle-hashing-preferred)
; Two-tier invalidation is sound when both mtime and hash are available (normally r-two-tier-sound (and mtime-based-cache content-addressed-hashing) two-tier-invalidation-sound)
; — Intentional conflict: cache reasoning results —
; Cache the theory for fast startup (normally r-cache-theory mtime-based-cache cache-reasoning-results)
; But recomputing ensures results are always fresh (normally r-recompute-theory incremental-rebuild (not cache-reasoning-results))
; No (prefer …) between r-cache-theory and r-recompute-theory. ; zetl reason conflicts will surface this as an unresolved tension. ; This demonstrates how defeasible reasoning models real design debates.
; Cache is safe when disposable and vault is read-only (normally r-cache-safe (and disposable-cache read-only-vault-access) cache-safe-to-deploy)