Why this document exists

A 19-axis investigation with 200+ content pages fails if readers can't find the argument. Most accountability sites are flat navigation + dated-list landing pages — that works for ~50 pages, breaks at 200+. This document records the specific IA decisions TENET5 made to keep the investigation navigable, and why. It's also a template: any other investigator building a large evidence base can apply the same pillars.

PILLAR 01 · OVERVIEW-FIRST

Curated reading path: shock → finding → structural → action

A first-time visitor gets 11 pages in 4 phases in ~45 minutes via reading-path.html, not a blank index. Ben Shneiderman's "overview first, zoom and filter, details on demand" (Visual Information Seeking Mantra, 1996) applied: the 4 phases are the overview; each phase's 2-3 pages zoom in; the per-page primary-source citations are details-on-demand.

Phase 1 · Shock

Concrete evidence. Visceral entry.

Phase 2 · Finding

The cross-axis pattern.

Phase 3 · Structural

Why the system misses it.

Phase 4 · Action

What citizens can do.

Research note: Shneiderman (1996), "The Eyes Have It" conference proceedings. Also Morville & Rosenfeld, Information Architecture for the World Wide Web (O'Reilly). Curated reading paths outperform flat indexes for multi-document investigations — see ProPublica's long-form design patterns.
A flat A-Z list of 200+ pages produces paralysis of choice. A curated sequence produces momentum. The page is a suggestion, not a requirement — links to axes-index.html always allow the power-user jump.
PILLAR 02 · PROGRESSIVE DISCLOSURE

TL;DR at top, primary sources at bottom

Every analysis page starts with a 2-3 sentence summary, then the structural finding, then per-source detail, then the receipts. Readers who only have 30 seconds get the thesis; readers who have 10 minutes get the evidence chain; readers who want to verify get the SHA-256 Merkle anchors.

Research note: Nielsen Norman Group research on progressive disclosure (2006, revisited 2022): "Hide advanced functionality behind a 'More' link or second screen." For accountability journalism, the "advanced" layer is the primary-source verification chain; the first layer is the structural finding.
Standard news sites bury the lede in prose. Investigation sites need the opposite — the thesis IS the lede, the story is the evidence.
PILLAR 03 · PROVENANCE CHAIN

Every claim has a source. Every source has a hash. Every hash is committed.

IA for evidence-based journalism requires provenance to be structurally inseparable from the claim. On TENET5:

Research note: Kahn & Cerf (1988), "A Scheme for Long-Term Information Integrity" — Merkle trees as tamper-evident verification. Also the IETF W3C Verifiable Credentials specification (2022): provenance is a first-class property, not a footnote.
Investigative reporting that doesn't publish its receipts is asking readers to trust, not verify. We publish the receipts.
PILLAR 04 · PAGE-CHAIN WALKTHROUGH

Automated reading sequence across the PAGE_SEQUENCE

193 curated pages in PAGE_SEQUENCE + 5 presentation phases per page. When a reader starts the walkthrough on any page, the engine:

  1. Narrates every [data-narrate] block in order (mean ~8 per page).
  2. On page-end, sets sessionStorage.liril_autopilot = true.
  3. Calls window.__TENET5_NEXT_PAGE() to advance.
  4. Next page auto-starts narration on arrival (autopilot flag persists).
  5. At PAGE_SEQUENCE end, shows "Tour complete" overlay.
Research note: Continuous-reading flow is documented in e-book / audiobook UX research (e.g., Nielsen, "Reading Online Copy"). For investigation sites, it matches podcast consumption patterns better than page-by-page clicking — the reader chooses to engage, the system handles navigation.
Autoplay defaults to ON (user can disable with A key persisted in localStorage.tenet5_wt_autoplay). Chain advances even without autopilot via liril-walkthrough.js's own advanceToNextPageWalkthrough() at sessionStorage.liril_autopilot.
PILLAR 05 · ACCESSIBILITY & NARRATION PARITY

Every visual presentation has an audio + text equivalent

Every [data-narrate] attribute carries a screen-reader- friendly narration of the visual content. Voice synthesis (Clara via LIRIL voice profile) reads sections aloud. Transcript side-panel (T key) shows all narration as scrollable text for deaf / hard-of-hearing readers. Closed captions (C key) display current sentence while speech plays. Reading time estimate (~170 wpm × speed multiplier) tells users how long the page takes.

Research note: WCAG 2.1 AA compliance is the floor. The actual goal is sensory parity: every piece of information available visually is also available via audio, text, or keyboard. See the limitations & critiques page for the full accessibility statement.
Accountability journalism that doesn't reach users with disabilities isn't accountability journalism. It's an inaccessible evidence dump.
PILLAR 06 · NO TRACKING. NO ANALYTICS. NO CHAT.

Privacy by design, not by policy

Zero third-party trackers. No Google Analytics. No Facebook Pixel. No server-side open/click tracking on the .eml campaigns. Page state (walkthrough progress, sent-campaign marks) lives in localStorage on the user's device — never transmitted anywhere. Even the LIRIL-authored-advice dashboard reads from static JSON files, not a live API.

Research note: Cavoukian (2009), "Privacy by Design: The 7 Foundational Principles." The site is an investigation — surveilling the readers of an investigation into surveillance would defeat the investigation's own premise.
LIRIL (R6-Q1) suggested adding campaign tracking pixels + A/B testing. REJECTED on this principle. The roadmap documents the rejection so future investigators can see what we declined and why.
PILLAR 07 · TRANSPARENT AI USE

When LIRIL drafts, LIRIL is credited — and verified

Every LIRIL-drafted artifact (campaign letters, interpretations, analyses) is attributed with subject (tenet5.liril.infer), elapsed milliseconds, timestamp, and prompt. When LIRIL hallucinates, the hallucination is documented on the roadmap before being fixed. Every commit that LIRIL influenced is co-authored in the Git metadata. Human review is the canonical gate — the LIRIL roadmap's 4 red REJECTED rows (R5-Q3, R6 triple-rejection, etc.) show the gate catching real errors.

Research note: Partnership on AI (2020), "Responsible AI Disclosure." Attribution + verification + rejection logs are three of the eight pillars of responsible ML disclosure.
"The AI wrote it" is not a defense or an excuse. It's an attribution. The human investigator chose to ship or reject, and the roadmap logs that choice.
PILLAR 08 · RECOVERY & AUDIT TRAIL

Every change is atomic, Git-committed, and publishable

The whole site is a GitHub repo at github.com/TENET-5/TENET-5.github.io. Every change is a commit with a descriptive message. Every session's work is traceable. Every dossier file's SHA-256 is recomputable. Every campaign's correction history lives in the data file itself. Recovery from a bad state is always a git revert away.

Research note: O'Reilly, Version Control with Git (Loeliger & McCullough). For journalism specifically, see Nieman Lab 2023: "Git as a journalism primitive."
An investigation that can't be rolled back to a verified state is an investigation without an undo button. We have one.

Anti-patterns — what TENET5 deliberately does NOT do

No infinite scroll home feed. Infinite scroll erodes sense of position. Each page has a finite, knowable scope; the reading-path.html makes the whole-site scope knowable too.
No paywall. No email capture. No account required. Every claim, every dossier JSON, every Merkle receipt is public URL- addressable.
No algorithmic personalization. Everyone sees the same ordering, the same phases, the same citations. Personalization would fragment the provenance chain.
No comments. No reactions. No chat widgets. This is documentary accountability journalism, not a social platform. Reader engagement is letters to oversight bodies, not likes on posts.
No ad tech. Literally zero. The site has no revenue model because revenue would distort incentives. It's public-good infrastructure on free GitHub Pages hosting.

Reading-time footprint

Total site narration: 76,954 words across 1,669 narration blocks (measured by scripts/scan_narration_integrity.py). At 170 words/min, that's ~7.5 hours of end-to-end voiceover. The reading-path.html curated subset is ~45 minutes. The state-of-investigation.html executive summary is ~8 minutes.

Using this as a template

To apply these pillars to your own investigation site:

  1. Fork github.com/TENET-5/TENET-5.github.io
  2. Replace data/*_grover_decisionmakers.json with your own axis dossiers
  3. Replace page content; keep the structure (reading-path, axes-index, state-of-investigation)
  4. Keep the shell.js load chain as-is — it's the canonical walkthrough stack
  5. Publish Merkle receipts from tools/_merkle_*.py scripts — all parameterized by SYSTEM_SEED
  6. Open-source everything. Provenance is worthless if it's private.
See also: State of Investigation · Reading Path · Axes Index · Limitations & Critiques · LIRIL Roadmap · Methodology
SYSTEM_SEED 118400 · author: human · commit trail on GitHub