025 — BT1 Canonization and Lane Separation

Date: 2026-03-22 22:19 UTC

Status: Draft

Domain: Research operations, repo governance, canonical publication

Thesis

When a research program crosses lane boundaries, canon must be separated by repository identity. Otherwise, receipts accumulate but authorship and retrieval collapse.

Event

A full second-run corpus was proven by witness surfaces and operator trace, but discoverability collapsed because outputs were distributed across runtime lanes, fallback pages, and mixed branch histories.

Correction

A dedicated repository (srida2) was created and declared BT1 canon.

This moves from:

• implicit canon ("it happened somewhere"),

to:

• explicit canon (named container, stable URL, seeded artifacts, declared scope).

Principles

1. One lane, one canon root

2. Witness ≠ canon (witness can prove existence; canon stores retrieval)

3. Exposure before expansion (if users cannot access it, it is not shipped)

4. Separation preserves judgment (do not blend placeholder/access scaffolds with authored papers)

Practical Standard

For each completed run:

• create or update canonical repo,
• publish run index + provenance ledger,
• verify public witness URL,
• gate new run until parity is visible.

Closing

Canonization is a product decision, not a clerical step.

BT1 became durable when lane separation became explicit.