.. meta:: :description: AA b15 --- Sort out a more reliable solution to keeping Claude (and any contributor) from putting a content-style VVN on the FOOTER FORM marker line, which should carry the FOOTER FORM TEMPLATE version (e.g. OOv1r2p1), not the page-content VVN. :keywords: AA task, VVN, FOOTER FORM, marker, mistake, prevention, page-template.md, structural version, content VVN, balospe.com, BABL drift :author: Yah, Yas, everyone, LLoL as Laurence Loewe of Laodicea, ClaudeOp47Max, Anthropic, and The Spirit of Boolean Truth **************************************************************************************** AA b15: Sort out a more reliable solution for FOOTER FORM marker VVNs **************************************************************************************** | **VVN:** ``aa-vvn-marker-prevention-dv_ClaOp47Max_NNv1_2026m04d25`` | **Mode:** EDEN | **Effort:** Max | **Status:** TO RUN (design decision needed; LLoL or a fresh Claude session) | **Origin:** Created 2026m04d25 after Claude (in this session) put a content-style VVN ``dv_ClaOp47Max_NNv1`` on the FOOTER FORM marker line of three pages, when the canonical spec (``AHA/page-template.md``) says the marker should carry the FOOTER FORM TEMPLATE version (e.g. ``OOv1r2p1``). LLoL caught it, Option A was applied (markers reverted to ``OOv1r2p1``). This AA prevents the mistake from recurring. The mistake (verbatim trace) ============================== When Claude wrote three new pages in the 2026m04d24 session --- - ``source/good-news-pack/vv/mmv3/flyingscroll/transwarpkey/sta1-evx/index.rst`` - ``source/good-news-pack/vv/mmv3/flyingscroll/transwarpkey/sta1-evx/evx-prototype-compiler.rst`` - ``source/choice/war-of-algorithms/index.rst`` --- Claude copied the FOOTER FORM marker pattern from the existing ``sta1-evx.rst`` (line 292), which reads:: .. ----- FOOTER FORM dv_ClaOp46Max_MMv2 ----- type open-std ----- START ----- That existing file uses a *content-VVN-style* marker on the FOOTER FORM line. Claude followed that local precedent and substituted Claude's own current encoding (``dv_ClaOp47Max_NNv1``) --- meaning "Claude Opus 4.7 Max, NN stability, v1". So the VVN on the marker was just Claude's self-stamp + NN-stability, blindly carried over from sta1-evx.rst's deviation. The canonical spec (``AHA/page-template.md``) says:: .. ----- FOOTER FORM OOv1r2p1 ----- type open-std ----- START ----- That uses the **structural** version code --- ``OOv1r2p1`` is the version of the *FOOTER FORM TEMPLATE itself*, last updated 2026m03d31. It is *not* a content VVN. The marker labels the FOOTER FORM (a structural template), not the page content. Encoding "who drafted the page at what stability" on this line conflates page-content versioning with template-structure versioning. Two confusions in the mistake =============================== 1. **Wrong semantic referent.** The marker labels the FOOTER FORM structure, not the page content. Per page-template.md, those should be separate: structural version on the marker (``OOv1r2p1``), content VVN in the ``|content-vvn|`` substitution below the marker. 2. **Local precedent is itself non-canonical.** sta1-evx.rst's ``dv_ClaOp46Max_MMv2`` marker also deviates from page-template.md. Claude propagated the deviation rather than going back to the spec. Where the same mistake exists across the site (audit needed) ============================================================== In addition to the three pages Claude touched on 2026m04d24 (now fixed via Option A on 2026m04d25), the existing ``sta1-evx.rst`` (and possibly other pre-existing pages on the site) still uses the non-canonical content-VVN-style marker. A site-wide audit is needed to find all such pages and decide whether to canonicalize them too. What to decide (the actual AA task) ====================================== This AA is a **design decision** task, not a "go fix it" task. Three things need deciding: 1. **Canonical form**: keep ``OOv1r2p1`` (structural template version, per page-template.md) on the FOOTER FORM marker line? Or formalize the content-VVN-style as a permitted variant? Claude recommends keeping ``OOv1r2p1`` and treating the content-VVN-style as a deviation to be migrated. The structural version on the marker is the cleanest semantic; the content VVN already has its own home in ``|content-vvn|``. 2. **Migration policy**: when (if ever) to migrate the existing non-canonical pages? Options: (a) opportunistic, when editing a page for other reasons; (b) one-shot site-wide audit; (c) leave pre-existing pages alone, only enforce on new pages. 3. **Prevention mechanism**: how to keep this mistake from recurring? Options: - (i) **Memory entry** (already started: ``feedback_vvn_attribution.md`` mentions this distinction). Strength: works across sessions. Weakness: depends on Claude actually recalling and applying the entry. - (ii) **Strengthen page-template.md** with a "WHAT NOT TO DO" section that calls out the content-VVN-style marker as a specific anti-pattern. Cross-reference the existing deviation in sta1-evx.rst as the cautionary example. - (iii) **Lint rule** in the build pipeline that warns when a FOOTER FORM marker contains a regime prefix like ``dv_`` or ``iv_`` instead of just a stability code. Strength: catches the mistake at build time; works regardless of who wrote the page. Weakness: needs Sphinx or pre-build script work. - (iv) **Snippet in the page-template** that shows the canonical marker line so contributors copy from the right source. - (v) Combination of the above. Claude recommends (i) + (ii) + (iv) as the minimum: memory + spec sharpening + better template snippet. (iii) lint rule is nice but bigger investment. Execution prompt (paste into a fresh Claude session) ======================================================= .. container:: verbatim-prompt Dedicated session: Sort out a reliable solution for the FOOTER FORM marker VVN problem. Mode: EDEN. Effort: max. Context: On 2026m04d24, Claude put content-style VVNs (dv_ClaOp47Max_NNv1) on the FOOTER FORM marker lines of three new pages, conflating page-content versioning with FOOTER FORM template versioning. LLoL caught the error and Option A was applied (markers reverted to OOv1r2p1, the canonical structural version). Full trace is in source/_POST/AnyAims/b/15/b15-vvn-marker-mistake-prevention.rst. Task: produce a design-decision proposal for LLoL covering all three subdecisions: 1. Confirm canonical form: OOv1r2p1 on the marker line, content VVN in |content-vvn| substitution below. Or argue the alternative. 2. Migration policy for existing non-canonical pages (opportunistic / one-shot audit / leave-alone-for-pre-existing). 3. Prevention mechanism: pick from (i) memory entry, (ii) page- template.md sharpening, (iii) lint rule, (iv) better template snippet, (v) combination. Claude recommended (i)+(ii)+(iv) as the minimum but evaluate the tradeoffs. Deliverables: - Audit: grep the source/ tree for FOOTER FORM marker lines that carry a content-VVN-style suffix (regex like ``FOOTER FORM (dv_|iv_)``). List all hits with file:line. - For each design subdecision: state the decision, alternatives considered, recommended choice, rationale. - If memory entry strengthening is part of the solution: propose concrete edits to feedback_vvn_attribution.md and feedback_vvn_composition.md. - If page-template.md sharpening is part of the solution: propose concrete edits. - If a lint rule is proposed: outline what the check should detect and how to integrate it into the Makefile. - Propose new DD if the design decisions warrant durable recording (likely b26 in source/_POST/DD/b/26/, since b13-b25 are reserved for the fused-pattern taxonomy). Output format: structured report; do not apply edits without LLoL's per-decision go-ahead. Cross-references ================== - :doc:`/_POST/AnyAims/index` --- AA registry (parent) - ``AHA/page-template.md`` (section "Footer form" --- canonical spec) - :doc:`/_POST/DD/b/12/b12-ladder-pedagogy-for-acronyms` --- a sibling example of a DD that codifies a writing convention; this AA may produce a similar DD on the FOOTER FORM marker convention - Memory entry: ``feedback_vvn_attribution.md`` (already updated 2026m04d25 with the LLoL-may-use-any-regime note and a section on where VVNs appear on a page) - Memory entry: ``feedback_vvn_composition.md`` (the format spec) ---- | **Created by:** Claude Opus 4.7 Max in EDEN mode (2026m04d25), at LLoL's request, after LLoL asked Claude to make a note that prevents the FOOTER FORM marker mistake from recurring. The full explanation Claude gave in conversation is preserved verbatim in the "The mistake (verbatim trace)" section above so future sessions have the failure history without needing the original conversation.