58 lines
5.5 KiB
Markdown
58 lines
5.5 KiB
Markdown
# Lint Report — Sources (2026-07-18)
|
|
|
|
#lint-report
|
|
|
|
## Scope
|
|
|
|
Health check of the **12 source summary pages** in `wiki/sources/*` (per user request "run a lint on sources"). Cross-page checks (inbound links, contradiction coverage) were run against the whole vault where a source page is one end of the relationship. This is the vault's **first lint**.
|
|
|
|
Checks run: raw↔summary coverage · page-type tags · template-section completeness · outbound link validity · inbound linking / orphans · cross-source contradiction coverage · stale/superseded claims · metadata completeness · evidence provenance.
|
|
|
|
## Verdict
|
|
|
|
**The source layer is healthy.** No orphans, no broken links, no missing or mismatched tags, complete raw↔summary coverage, and provenance is well-maintained. Five findings, all Low/Info; two were fixed in this pass, three are flagged.
|
|
|
|
## Passing Checks
|
|
|
|
- **Coverage is complete and 1:1.** 12 raw files in `raw/sources/` → 12 summaries in `wiki/sources/`. No un-ingested raw source; no summary without a backing raw file.
|
|
- **Page-type tags correct.** All 12 pages open with `#source` as the first tag; topical tags follow. No folder/tag mismatch.
|
|
- **Template sections present.** All 12 carry Source Metadata, Core Claims, Key Evidence / Details, Connections, Open Questions, and Change Impact on Wiki. (`konspekt-aphorisms` uses "Core Claims (the aphorisms)" — compliant.)
|
|
- **No broken outbound links.** Every `[[…]]` target in every source page resolves to an existing page.
|
|
- **No orphan sources.** Every source has ≥1 inbound link from a content page (concept/entity/comparison/query/overview), plus index/overview catalog links.
|
|
- **Provenance chain intact.** The primary→secondary demotion of [[2026-06-15-selling-development-services-in-the-ai-era]] is recorded on the page and reflected in concept-page citations; its resolved open question is struck through, not deleted (history preserved). Primaries carry accurate "provenance role" notes. The `konspekt-aphorisms` "Indian dev" stereotype is quoted for fidelity with an explicit non-reuse caveat — compliant with the citation policy.
|
|
|
|
## Findings
|
|
|
|
### F1 — Content-marketing-timing contradiction was under-recorded (Low) — ✅ fixed
|
|
|
|
The new source [[2026-07-18-information-is-free-implementation-is-paid]] makes organic content the *primary* lead engine from day one — a third, most-aggressive position on content timing. [[client-acquisition-channels]] still recorded the split as two-way (Tony vs AB Analytics). Non-negotiable rule #5 requires contradictions be recorded explicitly.
|
|
**Fix:** upgraded the Contradictions bullet in [[client-acquisition-channels]] to the three-way split, noting the "audience-dependent" escape is weaker here (all three are US SMB coaching sources).
|
|
|
|
### F2 — Missing reciprocal link + inaccurate Change Impact on the LinkedIn template (Low) — ✅ fixed
|
|
|
|
[[2026-06-15-linkedin-mail-template]]'s Change Impact claims it is "linked from [[client-acquisition-channels]] and [[sales-discipline]]," but `sales-discipline` did **not** link it — leaving the template inbound-linked from only one content page (its weakest link in the vault).
|
|
**Fix:** added the link from [[sales-discipline]] ("customize every first touch" → the concrete instance is the template), making the claim true and strengthening the weak link.
|
|
|
|
### F3 — `meta-analysis` is the most weakly inbound-linked source (Low) — flagged
|
|
|
|
[[2026-06-15-meta-analysis-selling-dev-in-ai-era]] is inbound-linked from only [[overview]] among content pages (plus index/log). Acceptable for a second-order map that mostly *points at* the primaries, but it is the least-connected source. Optional future strengthening: cite it from the concept pages whose contradictions it first surfaced (content-timing, team-size).
|
|
|
|
### F4 — Filename date-prefix convention is inconsistent for undated videos (Info) — flagged
|
|
|
|
Three conventions are in use for the date prefix: **ingestion date** for the two US anchors (`2026-07-17-design…`, `2026-07-18-information…`), **batch date** for the June-15 set, and **file date** for `2026-07-06-sebastian…`. Each page documents its own date in metadata, so nothing is misleading — but the naming rule doesn't specify which date to use when a source is undated. Consider codifying: "undated source → use ingestion date." No file renames recommended now (links would churn).
|
|
|
|
### F5 — Corpus remains single-school and unverified (Info) — flagged (standing gap)
|
|
|
|
The 12th source adds no statistics but also no verification and no dissent: 11 of 12 sources are the same "sell dear / give-value / productize" school; all quantitative claims across the corpus remain attributable-but-unverified (promotional videos). Already tracked in [[overview]] and [[index]]. The highest-value next ingest is still an **adversarial / failure-case** source; a warm partial-dissent already exists in [[sebastian]] (channels only).
|
|
|
|
## Changes Made This Pass
|
|
|
|
- [[client-acquisition-channels]] — content-timing contradiction upgraded to three-way (F1).
|
|
- [[sales-discipline]] — added reciprocal link to [[2026-06-15-linkedin-mail-template]] (F2).
|
|
|
|
## Recommended Next Operations
|
|
|
|
1. Ingest an **adversarial source** (F5) — the one gap that materially changes vault confidence.
|
|
2. A **full-wiki lint** (this pass covered sources only) — to check concept/entity pages for the same categories, plus high-mention concepts lacking pages (e.g. *referrals* is a high-value channel with metrics but no dedicated page).
|
|
3. Optional low-priority tidy: F3 (link meta-analysis from concept pages), F4 (codify the date-prefix rule).
|