Assistant: codex Assistant-Model: gpt-6-astra Assistant-Session: 01a0cffe-206f-7323-89d3-b20ca4cf4bf9
44 lines
2.3 KiB
Markdown
44 lines
2.3 KiB
Markdown
# Schema and source preparation review
|
|
|
|
Date: 2026-09-24 · Reviewer: codex (self-review)
|
|
|
|
## Schema decisions
|
|
|
|
The working 0.1 contract resolves the initial review's representation issues:
|
|
|
|
- Stable FEP-NNNN pattern identities independent of mutable family/scale labels.
|
|
- Editorial status separated from evidence maturity; seed remains a catalog label.
|
|
- Explicit scale lists and registered kebab-case force IDs.
|
|
- Conditions in prerequisites; pattern dependencies use typed requires edges.
|
|
- Typed relation targets with justification and local evidence references.
|
|
- Source IDs, review paths and claim labels form a traceable citation chain.
|
|
|
|
The isolated synthetic examples exercise two pattern IDs, a relation endpoint,
|
|
force lookup and a source/review/claim reference. No example becomes a corpus
|
|
pattern. `make schema-test` checks agreement and resolution. A comprehensive
|
|
validator and migrations for future versions remain in FEP-WP-0004.
|
|
|
|
Outcome for FEP-WP-0002-T01: complete. This settles serialization for the first
|
|
cycle; it does not validate the nine-responsibility theory or any candidate pattern.
|
|
|
|
## Primary-source preparation
|
|
|
|
SRC-0001 is the author's DEV publication, linked from Ryan Carniato's profile.
|
|
Its displayed date is Sep 22; the API metadata request was denied, so the year
|
|
has not been independently verified. The packet pins the accessed web edition.
|
|
SRC-0002 is the authors' Ink & Switch essay, dated April 2019; it supplies a
|
|
conference citation and DOI. We use the web edition, not an assumed identical PDF.
|
|
|
|
Both packets retain short original passages, exact section/sentence locators,
|
|
clear editorial boundaries, source URLs and SHA-256 digests. They isolate authority
|
|
placement rather than attempting full article extraction. No existing research stub
|
|
was used as a substitute for a primary source. Each workspace has a four-call plan;
|
|
neither has been executed or independently reviewed.
|
|
|
|
## Next evidence work
|
|
|
|
Run the two bounded investigations, inspect claim fidelity and negative results,
|
|
then extend source review to mutation, persistence, reconciliation and recovery.
|
|
Do not complete FEP-WP-0002-T02 on the basis of these excerpts. The full case reviews,
|
|
independent known uses, six-to-ten-pattern synthesis and VC-01/VC-02 remain open.
|
|
No provider/model was chosen and no paid generation was started in this preparation.
|