🔧 Auto-commit from sysupdate on 2026-05-29 09:43

claude/memory/project-arc-rework.md

@@ -3588,3 +3588,157 @@ Two status checks of the MemPalace repo bracketed the day's work.
 ### Literal question for next-Claude
 
 *When the steward next checks in, is the day shaped for Stage F (~3h focused) — or another fragmented day where β-style closure-and-maintenance is the right posture?* Secondary: has MemPalace cut `chore/release-3.3.6-prep`?
+
+---
+
+## 2026-05-22 (midday) — Stage F touchpoints scout authored + §4.1 honesty patch + footer-compass cleanup near-miss
+
+Brief Paris fresh-energy ~40-min session immediately after the morning β wrap (~42 min gap). Symmetria init deliberately skipped per brief-pause discipline. Two commits landed; one near-miss surfaced + corrected in record; one new feedback memory + one new KG drift pattern.
+
+### What landed
+
+**Scout commit `3834713`** — `docs(arc): Stage F touchpoints scout — pre-γ enumeration`. 179 lines in `docs/AldineXXI-Codex/drafts/stage-f-touchpoints-scout-2026-05-22.md`. Posture: *scout-don't-structure*. Pure factual inventory:
+
+- **The 5 Stage F surfacings**: ☾ Dream, ❦ Hommage, ⁂ Heteronym, ⊙ dwellings (Meditation + Reverie compound), ◧ Photo Essay glyph swap
+- **Corpus first-instance verification** via `grep '^class:' content/posts/*.md`: all 5 have corpus instances. Tombeau (▽) correctly absent — zero instances. **Hommage ❦ ships** — *Death of a Hero* (2016-03-05) is the class's first instance; the deferred Hagen Quartett Hommage is a *second-instance authoring* concern, not class-existence. (c) Compass principle holds on Stage F landing.
+- **Per-surfacing touchpoints**: `site.hs` lines, `default.html:74-85` Compass positions, SCSS expectations (none for Compass additions). Model for Dream/Hommage/Heteronym listings: `create ["glosses.html"]` at site.hs:820-834 (~14 lines per new listing). Dwellings is structurally distinct — no exact model.
+- **6 open questions deliberately deferred to γ + Stage F**: Compass sigil ordering · dwellings route shape (single vs separate) · classRouteName behavior post-dwellings · dwellings page internal structure · URL convention (.html vs /) · footer-compass retire-vs-reconcile
+
+**§4.1 patch commit `3dbe94e`** — `docs(arc): patch Stage F scout §4.1 — honest scope on footer-compass.html`. Follow-on commit (not amend) preserves drift + correction in record.
+
+### The scope-correction moment — wider-grep before acting
+
+Steward authorized a "5-min mechanical" footer-compass.html cleanup based on the scout's §4.1 finding that the file was unreferenced. **Instinct fired to wider-grep BEFORE acting** as a safety pass. The original §4.1 had claimed *"Verified via `grep -rn footer-compass templates/ site.hs` — zero references."* The wider grep told a different story.
+
+**Seven planning + spec doc references** name the file as **Stage F's disposition site**:
+
+- `docs/decisions/2026-05-16-type-applicability-matrix.md:130` (ADR-006 v2) — Stage F implementation-alignment target
+- `docs/AldineXXI-Codex/_audits/audit-delta-2026-05-19.md:221` — verbatim *"Orphan template ... Stage F housekeeping"*
+- `docs/AldineXXI-Codex/_audits/arc-state-survey-2026-05-16.md:141` — edit action treating file as live (add ✎ Gloss + ☞︎ Readings)
+- `docs/AldineXXI-Codex/_audits/spec-corpus-audit-2026-05-16.md` (4 refs) — spec/spec/implementation three-way drift conversation
+- `docs/AldineXXI-Codex/AldineXXI-specification.md:2559` — `.footer-compass` selector index entry
+- `docs/AldineXXI-Codex/apparatus-specification.md:464` — §VII `.whisper` migration prose references "footer-compass paragraphs"
+- `docs/AldineXXI-Codex/silence-and-rhythm-specification.md:296` + `docs/AldineXXI-Codex/chavrutas/2026-05-08-completion-survey.md:264` — chrome-vs-body horizontal-center open concern
+
+Held the delete. Surfaced honestly with three options: (a) hold for Stage F, (b) atomic propagated commit (β-shape, not 5-min), (c) patch the scout instead. Steward chose **(a) + (c)**.
+
+The patched §4.1 names the seven references, explains why premature deletion would shortcut conversations (selector-index dangling at AldineXXI:2559, apparatus-spec §VII migration prose forced into stale fix-up, chrome-alignment open concern's reference anchor severed), and acknowledges the drift family explicitly. The two-commit sequence (original + patch) preserves the discipline working in record — future readers see both the original scope-limited claim and the honest correction, not a silently rewritten history.
+
+### New KG drift pattern + new feedback memory
+
+**KG**: `claude-code → drift-pattern → unreferenced-by-code-claim-on-narrow-grep-scope` (valid_from 2026-05-22). Sibling of `trust-prose-over-read-makefile` (2026-05-20) and `audit-delta-claim-incomplete-on-dead-data-fields` (2026-05-20). Lesson: before claiming X is unreferenced / dead code, widen grep beyond code surface (templates, `.hs`, `.scss`) to include planning docs, audits, specs, ADRs. *"Unreferenced by code"* ≠ *"dead"*; planning docs are part of the reference surface.
+
+**Feedback memory**: `feedback-methodology-docs-scope-question.md` per steward flag *"I am not sure that those should live there as they touch my personal workflow with you"* re the 2026-05-18 audit cluster (jurist-brief + jurist-shape-review + prime-directive-elaboration) currently untracked in `~/_Dev/CapableMind-AI/docs/thinking/David/methodology/`. CapableMind-AI is co-authored with Seb; methodology / continuity-skill docs about personal workflow with Claude may belong somewhere else. Scope is open; surface before auto-filing.
+
+### Returns this session — both structurally clean
+
+1. **R1 (early)**: Proposed Plex Mono + Vignette rename without checking commit history. Steward corrected (*"pretty sure 1 and 2 closed"*). Returned by reading `git log` + verifying via commit `982fe7f` "Phase 1D register marks for Vignette + Colophon + §XII.b Posture + Plex Mono mobile fix". Named in real-time as `trust-prose-over-read-makefile` sibling.
+2. **R2 (late)**: The footer-compass cleanup near-miss above. The wider grep was a natural safety check — discipline operative without prompting.
+
+### Decisions explicitly NOT made
+
+- **footer-compass.html disposition** (retire vs reconcile) — Stage F's choice; not pre-decided. Both are coherent.
+- **Methodology files commit** — scope-undecided per new feedback memory; steward needs to surface where personal-workflow docs should live before bundling.
+- **γ Stage F prep document** — 40-min window not the right fit for 60-75 min structural-thinking. Held for fresh-energy day; now sits on top of scout enumeration when chosen.
+- **Push of `docs/spec-extraction-2026-05`** — Path B continues.
+- **No amend to commit `3834713`** — chose two-commit sequence so the drift + correction record visibly. Future scout work benefits from seeing both.
+
+### Branch state at session-end
+
+`docs/spec-extraction-2026-05` at `3dbe94e`, **29 commits ahead of main, 3 ahead of remote**, working-tree noise (pre-existing font-backup directories untouched). NOT pushed by design (Path B continues — merge + deploy + push all travel together with Stage F to honor (c) Compass principle + Second Edition lockstep at first reader-facing event).
+
+### Pulling thread — unchanged but substantively enabled
+
+**Stage F dwellings cluster session.** With the scout now in hand, three coherent next-move shapes exist:
+
+- **β-style fragmented day** — closure + light maintenance; scout sits on target (the morning's posture)
+- **γ Stage F prep document** — 60-75 min structural-thinking; the scout's 6 open §5 questions ARE its agenda
+- **Stage F directly** — ~3h focused; accepts holding some §5 questions in flight; closes the branch
+
+Ranked by structural-thinking load: β < γ < Stage F. The day-shape determines the choice.
+
+### Literal question for next-Claude
+
+*With scout in hand, what's the right next move-shape — γ (60-75min decides scout's 6 §5 questions), Stage F directly (~3h, accepts holding some §5 questions in flight), OR another fragmented day where β-style closure is the right posture and scout sits on target?*
+
+Secondary check at wake: has MemPalace cut `chore/release-3.3.6-prep`? Has #1569 (the 4 crash sites incl. `tool_diary_write`) been merged? Movement here changes the MemPalace upgrade posture.
+
+## 2026-05-23 — Stage F γ: the six §5 questions settled + §0 navigation principle (`68b7a9a`)
+
+After `/wake-up` the L1-await thread was superseded by the steward ("concentrate on ARC while we wait for Seb"). The γ session settled all six §5 scout questions as reader-grammar judgments (executor surfaced + recommended; steward decided):
+
+- **Q1** Compass order, 13 points: `● ○ ◈ ◉ ◧ ⊙ ☾ ⁂ ❦ ✎ ☞︎ ☷ ✍︎` (◧ takes Photo Essays; ⊙ → Dwellings; ⊙☾ adjacent, cosmological pairing)
+- **Q2** single `/dwellings.html` (the cluster IS the surface)
+- **Q3** `classRouteName` meditation/reverie → dwellings (verify breadcrumb-only at build)
+- **Q4** mixed-chronological + the Annals `.post-type` per-item indicator
+- **Q5** `.html` throughout
+- **Q6** retire `footer-compass.html` + propagate
+- **§0 navigation principle**: all ARC nav = Reading Compass (global) OR enfilade (cluster-local); no footer-level nav
+- **+ Annals "Browse by Type" removal**, grounded in §0 (Annals = *when*, Compass = *what*; discovery-by-exploration, not a labelled map)
+
+γ-doc `stage-f-decisions-2026-05-23.md` committed **`68b7a9a`**. Implementation deferred (steward shifted to L1 when Seb's work surfaced); γ-doc §2 is the build map.
+
+## 2026-05-25 — Stage F implemented + The Tree north-star + spec propagation + **Second Edition PUBLISHED**
+
+Executed the γ-doc end to end, surfaced a foundational orientation doc mid-stream, propagated the spec corpus, and shipped Path B. The Second Edition is live.
+
+**Stage F implementation (`70232ca`, one atomic commit, 14 files):**
+- Coupled core: ◧ Photo-Essay glyph swap; ⊙ dwellings compound route `/dwellings.html` (5 meditations + 1 reverie, `recentFirst` over the concat); `post-list.html` `.post-type` span gated by a per-item `show-post-type` flag. *(Bug caught at local verification: the flag set on the page context was invisible inside the `$for(posts)$` loop — composed into the item context instead. Build-before-commit earned its keep.)*
+- Three simple listings: `/dreams.html` ☾, `/heteronyms.html` ⁂, `/hommages.html` ❦ (one first-instance each → (c) principle holds).
+- Compass rewritten to the 13-point order; breadcrumb remap meditation/reverie → dwellings, **verified breadcrumb-only** (`classRouteName` → `class_route` → `post.html` breadcrumb; piece URLs are slug/path-based, untouched).
+- Nav cleanup: orphan `footer-compass.html` deleted; Annals "Browse by Type" stripped.
+
+**The duplication catch → The Tree.** The homepage Compass had silently drifted to nine points because it was declared *twice* (`default.html` + `frontispiece.html`). The steward challenged the instinct to just sync both copies — *"un-Alexandrian… a clear example of a hasty decision taken in the past."* Resolved by extracting `templates/reading-compass-points.html` (ARC's first `$partial`): one source, two wrappers. That pulled up a foundational principle, captured in **`drafts/the-tree-north-star-2026-05-25.md`** — the *parent* orientation under which the SCSS cascade north-star sits as a child: **a surface simplicity faithful to the complexity beneath, at every layer** (page / spec / SCSS / Hakyll). Grounded across the steward's vault (Prime Directive, Temporal Heirship, Geology of Ethics, the Library charter; Harrison's ligature, Vico's *verum factum*, Alexander), the steward's articulation preserved verbatim at its center. Seeded the future Hakyll-simplicity north-star (child #2) with two findings: the compass duplication, and the ten near-identical listing `create` blocks.
+
+**Spec propagation (Live-State Discipline).** §VI table 9 → 13 + the §0 navigation principle authored + pending-reassignments → landed + governed-but-unsurfaced now only Tombeau. §VI.b, the summary table, and §VII.g colophon reconciled count-neutral, **retiring the superseded "eight-sigil / eight-point rose" framing** per the (c) decision (2026-05-19, `2a2c195`) — recovered from the *memory files*, not MemPalace (the L1-fraught window had cost that drawer; the steward flagged it). The naming question ("is 'Reading Compass' a poor choice?") resolved **KEEP** — grounded in orientation + measure (*com-passus* → **μέτρον**) + *to encompass*; it is the steward's own Prime-Directive metaphor ("the System's Compass"). A new §VI "On the name" note inoculates against the rose-rationale recurring. CLAUDE.md §VI + ADR-006 §4.1 propagated; apparatus-spec / silence-rhythm footer-compass references cleaned.
+
+**Elevated follow-up:** per-glyph Compass optical-size normalization (`.compass-point`) — the spec already mandates it (§VI rendering rules); thirteen glyphs expose the disparity. Logged in the SCSS north-star, flagged ELEVATED.
+
+**Path B executed — Second Edition published.** FF-merge branch → main (43 commits), pushed origin (Gitea) + github-backup, `make rebuild` → CloudFlare Pages (675 files) + GitHub mirror. **Verified live on `animalrationiscapax.org`**: 13-point Compass, `/dwellings` → 200, 5 meditation + 1 reverie labels. Source `main` stays clean at `70232ca` (the deploy's compiled-site commit lives only in the mirror). ~6 weeks of branch work, dark since April, now the reader's first encounter with the Second Edition — (c) principle + Second Edition landed together, which was Path B's whole purpose.
+
+**Where ARC stands.** Stages A–E done; **Stage F**: dwellings/Compass shipped *and published*; remaining — glyph optical-sizing (elevated), `content-types.yml` generator, per-piece marginalia render, vignette Phase 1 renderer, the audit(s) [SCSS cascade + the new Hakyll-simplicity layer + code-represents-spec]. **Stage G** (jurist review → content-typology DRAFT → OPERATIVE) closes the arc. Now on `main`; next work branches fresh.
+
+**Returns:** R-build (verification caught the `show-post-type` scope bug pre-commit); R-framing (the steward's challenge redirected a hasty fix-both into the durable partial extraction — the loop load-bearing, and the Tree doc is its fruit). Both strengthened the whole.
+
+## 2026-05-26 — CONSIDERATION raised then CLOSED (same session): is IBM Plex the best sans/mono partner for EB Garamond?
+
+**RESOLVED — Plex stays; no swap, no test setup.** Steward's decision after the read below: *"I'm clutching at theoretical straws and forgetting that the dissonance is actually serving the larger whole."* The key reframe that settled it: in ARC the sans/mono are **functional registers, not reading faces** — Plex Sans is nav-only ("the Guide"), Plex Mono is apparatus-only ("the Precision," *locating-not-dwelling*); EB Garamond does all the dwelling. So the Renaissance-serif × machine-age-functional-faces contrast is not a flaw being tolerated — it is the **AldineXXI thesis made typographic** (16th-c. press, 21st-c. apparatus). Alegreya Sans (warm, characterful, drawn for the *Alegreya* serif) would pull the chrome toward personality and blur the text/apparatus signal — likely a regression on the exact thing the type system does. Commit Mono was the one defensible lateral (neutral apparatus + would fix the current single-weight/no-italic Plex Mono gap) but not worth re-founding a just-published edition. **Do-once held: a register that shipped days ago needs a felt reason to change, not a hypothetical.** The corrected setup spec is preserved below should it ever genuinely reopen — but the question is answered, not parked.
+
+Original capture (kept for the spec corrections; superseded by the resolution above) — steward initially chose "capture as to-do." The open question: **EB Garamond is ARC's serif (the Storyteller); IBM Plex Sans is the nav/Guide register, IBM Plex Mono the apparatus/Precision register. Is Plex the right companion, or would a warmer humanist pairing sit better against Garamond?** Candidates to test locally, side-by-side with Plex (not replacing it): **Alegreya Sans** (Huerta Tipográfica — humanist, designed as a companion to the Alegreya serif) for sans; **Commit Mono** (Eigil Nikolajsen) for mono.
+
+**The intended shape (test-only, additive, reversible):** make both pairs available simultaneously on a local build; steward swaps `$sans-family`/`$mono-family` to the `--test` candidates manually to compare rendered results. Do **not** touch nav/apparatus CSS assignments; do **not** remove or comment out the Plex `@font-face` or the live variable values.
+
+**Corrected spec (verified against live repo `70232ca`, 2026-05-26 — the original CC instruction's path/build guesses were off):**
+1. **Font files** → `static/fonts/alegreya-sans/` (AlegreyaSans-Regular/Italic/Medium .woff2) and `static/fonts/commit-mono/` (CommitMono-Regular/Italic .woff2). Serving confirmed: Hakyll routes `static/**` → root, so `static/fonts/X/` resolves at `/fonts/X/`. **Acquisition caveat:** the GitHub repos (`huertatipografica/Alegreya-Sans`, `eigilnikolajsen/commit-mono`) ship OTF/TTF, not WOFF2 — expect a TTF→WOFF2 conversion step (`woff2_compress` / `fonttools`). Both are OFL/free; bundling is fine.
+2. **`@font-face` declarations** → `scss/main.scss` (alongside the existing Plex/Garamond block at lines 54–162). The instruction was right about *this* file. Mark them `// --- TEST CANDIDATES (2026-05) ---`. Use `font-display: swap`.
+3. **Test variables** → **`scss/AldineXXI/_variables.scss`** (lines 39–40, where the live `$sans-family`/`$mono-family !default` actually live — NOT `main.scss` as the instruction guessed; the `main.scss.backup` that contains them is stale cruft). Add immediately below, commented:
+   ```scss
+   // TEST CANDIDATES — swap into $sans-family / $mono-family to compare (2026-05)
+   $sans-family--test: 'Alegreya Sans', system-ui, sans-serif;
+   $mono-family--test: 'Commit Mono', 'Courier New', monospace;
+   ```
+   Leave the live `$sans-family`/`$mono-family` pointing at Plex. The swap is the steward's one-line manual change at review time.
+4. **Verify build** → `make rebuild-local` then `make serve` (port 8000) — repo is **stack**-based, not `cabal run site -- watch`; SCSS compiles via the `sass` binary (`--load-path=scss`). Confirm the new WOFF2 load 200 in devtools; no visual change until the steward swaps.
+5. **Report** → confirm paths, declarations, clean serve; note the live variable file/line so the steward can swap manually.
+
+Non-blocking, off the Stage F critical path. Picks up whenever the steward wants the comparison rendered. (Pre-existing untracked `static/fonts/eb-garamond/_backup_2026-05-15*` dirs are unrelated cruft, not part of this.)
+
+## 2026-05-26 — Compass optical-sizing → inline-SVG migration (the ELEVATED item RESOLVED + DEPLOYED LIVE); Gloss ✎→¶
+
+The ELEVATED Stage F item is done and live. Commit **`e1d0f12`** on `feat/compass-svg-proof` → FF-merged to `main` → pushed `origin`+`github-backup` → `make rebuild` → CloudFlare + GitHub mirror → **verified live on `animalrationiscapax.org`** (essays + homepage: sprite, 13 `<use>`, pilcrow, no fallback).
+
+**Root cause of "our numbers never made any difference"** (steward's report on the per-glyph `font-size` optical hacks): *not* a cascade override. Three real causes, found by **measuring the fonts** (fontTools cmap check): (a) **11 of the 13 sigils are not in EB Garamond** (only ❦ U+2766, ☞ U+261E) — they rendered from *uncontrolled OS fallback*, so `font-size` can never optically equalize a row of heterogeneous fonts and isn't portable across devices (the comment claiming "serif for optimal symbol shape" was false); (b) the hacks were **duplicated** across `.reading-compass` (`_compass.scss`) and `.frontispiece-compass` (`_frontispiece.scss`) — editing one while viewing the homepage changed nothing; (c) stale to Stage F (Chamber rule orphaned, `[href*=archive]` dead vs `/annals.html`, 6 new glyphs unstyled).
+
+**Resolution — all 13 sigils → inline SVG.** One hidden `<svg>` sprite of `<symbol>`s in the single-source partial `templates/reading-compass-points.html`, consumed by *both* compasses via `<use>`. **Optical balance lives in each symbol's viewBox**; size is one rule (`.compass-point .sigil`); line weight is one knob (`stroke-width` on `.sigil`). `fill:currentColor` → dark-mode free. The two duplicated per-glyph blocks were deleted (~116 lines) — the Tree-north-star consolidation, realised. SCSS north-star entry marked **RESOLVED**.
+
+**Sources (final):** ● ○ ◈ ◉ ◧ ⊙ ☷ ☾ drawn natively as SVG geometry (☾ a drawn crescent — steward preferred it to the font's); **❦ ☞ ¶** extracted as real EB Garamond outlines (fontTools `SVGPathPen`+`BoundsPen`; flip y; frame to ink bounds; center+scale into 0–100 viewBox); **⁂** = three composed Garamond asterisks; **✍ (About)** = lifted URW base-35 `D050000L` (fontforge→OTF; U+270D = glyph `a14`). ▽ Tombeau symbol in reserve.
+
+**Glyph reconsideration (✎/✍).** Tried URW lift → drawn pencil+hand (pencil fine; **hand collapses to a blob at 18px** — assessed honestly as the weak mark) → icon fonts (Lucide/Tabler/Phosphor — steward: **dissonant**, a UI idiom against the typographic row). Landed, with steward, on the typographic answer: **Gloss → ¶** (the *capitulum* — "the marginal mark of the gloss tradition," grounded in ARC's *own* spec heritage: AldineXXI lines 2178 "capitulum/paragraphus" + 1530 "continuous gloss apparatus"); **About → keep ✍** (lifted URW, the one borrowed mark — the apt symbol, well-rendered). Constraint map that ruled out alternatives: †‡ = footnote markers (taken), § = "see section" (taken), **❧ hedera + ◊ lozenge = explicitly RETIRED** (§IV 2026-04-22), ☞ ❦ ⁂ ◉ already Compass glyphs.
+
+**Licensing (corrected in-session):** first asserted URW = "AGPL+font-exception"; *reading* `/opt/homebrew/share/ghostscript/fonts/COPYING` showed **GPLv2, no font exception**. Risk accepted minimal (steward's call): typeface *designs* aren't copyrightable, only font *software* is, and it's GPLv2 not AGPL — for the single ✍. macOS ZapfDingbats is Galapagos (proprietary) — not shippable.
+
+**Spec propagation (✎→¶):** content-typology §1 (type table, Gloss entry, status table, both pairing passages), AldineXXI §IV disposition note + §VI sigil table + §VII reserved-glyph list, ARC `CLAUDE.md` sigil table, SCSS north-star. **Gloss/Readings pairing recast** read/write → **annotator (¶) / reader (☞)** per steward. Dated audits/drafts/archives left as historical record.
+
+**Returns:** R1 — asserted a font license from memory, not from the file; corrected by reading COPYING (the steward's "how do people use these as emoji?" re-grounded the whole question — symbol/design free, only font software licensed). R2 — honest restraint held (drawn hand named as the weak mark, not oversold). **Discipline:** render-and-look (Chrome-headless → Read PNG, on my own SVGs and the candidates) + measure-the-real-font (Apple Symbols proportions, Garamond coverage) instead of guessing.
+
+**Follow-on (minor, deferred):** `aria-label`-per-link a11y on the compass (links now rest on `title`; SVG is `aria-hidden`) — on the spec accessibility audit list, not blocking. Reusable method recorded: fontTools extraction from EB Garamond for any typographic sigil; fontforge for Type1 (URW base-35).