synthesis: shard spectrum across nine systems; fold into SHARD-WP-0002 (T11 eleven spectra, new T16)

Cross-dive synthesis (research/260614-shard-spectrum-synthesis/) reading
the two Nelson conceptual systems, four engines, and three modern tools
across each other: a shard family matrix and eleven capability spectra
(addressing, content identity, structure, history, native query,
translation, attachment mode, operational envelope, access grant, write
granularity, content types) — positions anchored at both ends by a real
system, federation ops degrading by position. Through-lines:
files-canonical/index-derived wins; fine-grained addressing is adoptable;
transclusion=clone=embed is one primitive; structure/history federate iff
in-text; attach mode is a per-binding choice; Notion proves the platform
can enforce no-silent-mutation.

Folded into SHARD-WP-0002: T11 reframed around the eleven spectra; T12
page model stretched four ways (prose + typed records + non-Markdown
assets + query-defined); T13 adds Notion supplement; T14 rewritten as the
three-mode attachment taxonomy (file-store / in-engine-host /
external-API); T15 adds lossy-with-fidelity-report; new T16 (addressing,
content identity, dimensional/query navigation). UC coverage UC-34-43 ->
UC-34-59. Workplan now 16 tasks. No new UCs (synthesis only).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

research/260614-shard-spectrum-synthesis/README.md

@@ -0,0 +1,47 @@
+# 260614 — Shard spectrum synthesis (one capability model across nine systems)
+
+Date: 2026-06-14
+
+## What this is
+
+A **synthesis** (no new external research) that reads the nine studied systems *across*
+each other and distills them into a single comparative model feeding the **shard adapter
+contract** (`SHARD-WP-0002`).
+
+The nine: two conceptual ancestors (**Xanadu**, **ZigZag**), four engines (**XWiki,
+TWiki, Foswiki** + the wiki-engines landscape), and three modern tools (**Roam,
+Obsidian, Notion**), against the federation/origin research.
+
+Centerpieces:
+- **The shard family matrix** — every candidate backend × {substrate, attach mode,
+  addressing, structure, history, query, transclusion, →Markdown, write granularity},
+  with Xanadu/ZigZag as the ideal anchors.
+- **The eleven capability spectra** — the claim that the adapter contract should model
+  *positions on spectra* (addressing, content identity, structure, history, query,
+  translation, attachment mode, operational envelope, access grant, write granularity,
+  content types), each anchored at both ends by a real system, with federation ops
+  degrading by position.
+- **UC-44–UC-59 → workplan task mapping**, including one genuinely new thread (T16:
+  addressing, content identity, dimensional/query navigation).
+
+## Contents
+
+| Path | Role |
+|------|------|
+| `findings.md` | Family matrix, the eleven spectra, cross-cutting through-lines, UC→task fold-in, recommendations/decisions, escalated open questions |
+
+## Status
+
+Synthesis complete. No new use cases (consolidation only). Feeds `SHARD-WP-0002`: T11
+reframed around the eleven spectra; T12 page-model breadth (prose + typed records +
+non-Markdown assets + query-defined); T13 history confirmed; T14 three-mode attachment
+taxonomy; T15 lossy-with-fidelity-report; **new T16** (addressing, content identity,
+dimensional/query navigation). UC coverage in the workplan extended UC-34–UC-43 →
+UC-34–UC-59.
+
+**Through-lines recorded:** files-canonical/index-derived is the winning architecture;
+fine-grained addressing is real and adoptable (adopt native, fingerprint otherwise);
+transclusion ⇄ clone ⇄ embed is one primitive; structure/history federate iff in text;
+attach mode is a per-binding capability-gated choice; Notion proves the platform can
+*enforce* no-silent-mutation via scoped grants.
+</content>