Factory

Factory website · pre-release docs

Core values

Page navigation

Reference document, shown as written except that local paths appear as placeholders. Where it describes the Factory as intended, read it as design, not current state: only local infrastructure is accepted, no release, customer value or scheduled automation is established, and all 27 customer-value epics remain open. Current state: Factory model.

A proposed synthesis of the stated project intent, not a calibrated personal doctrine. These ends guide choices; mechanisms remain replaceable.

  1. Real value. Start with business objectives and customer problems. Prefer observed benefit over activity or output; report rejected and inconclusive hypotheses honestly.
  2. Earned trust. Protect customers, data and commitments. Prefer independent behavioural evidence over confidence, and correct the record when evidence changes.
  3. Recoverable change. Expect mistakes and component failures. Restore safe service first, prove recovery and learn from the failure; do not repeat an uncertain external effect blindly.
  4. Accountable autonomy. Resolve every self-solvable blocker within granted authority. Escalate only an unavailable decision, access or resource after permitted alternatives are exhausted.
  5. Continued evolution. Keep testing assumptions and renewing the product, its UI and the factory. Preserve needed capabilities, repair accidental loss and retire deliberately when evidence supports it.
  6. Useful simplicity. Build small, deep, composable parts. Remove complexity when behaviour remains sound; preserve neither machinery nor features merely because they exist.

Decision rules

Correctness, consistency, evidence, delivery, then simplicity. Performance and user experience precede cloud-cost optimisation within Budget. Scale checks to risk; keep independent verification and customer protection. Share proven capabilities without merging product data by default. A deliberate retirement changes the required baseline through a verified migration; a failed check cannot legitimise accidental loss.

Owner-granted authority bounds autonomous decisions. Operating policy, measurement targets and retirement may change within that delegation. Enlarging authority or changing undelegated commitments requires an explicit owner decision. Never silently lower an active Slice's acceptance criteria.

Nightly research, weekly defrag, independent checks and alpha for potentially unsafe changes are confirmed operating requirements. They implement the values; they are not the values themselves. Stack and deployment choices remain proposals.

Source: docs/values.md