Trailstate · birth certificate · v0.2

ASCII Face Routing Grammar

A lightweight, browser-native grammar for replayable AI provenance. Routes are compact trailstates: readable by people, reusable by AI systems, and playable by Trailstate.

o-vvv-o o-www-o ovvv-o o-vvvo q-vvv-p n-vvv-n 0-vvv-0 p-vvv-q o-mmm-o u-vvv-u d-vvv-b b-vvv-d x-vvv-x

Why this exists

AI answers often hide the shape of retrieval: search, source conflict, narrowing, validation, synthesis, memory, archive. ASCII Face Routing compresses that path into a small symbolic trail. Trailstate plays the trail. GGTruth-style pages can explain and stabilize conflict states.

Canonical states

o-vvv-o

World / Field Root

field/world entry; artifact or route begins in the open field.

AI phase: context initialization
Action: open field
Domain: o-vvv-o.com

o-www-o

Open Web

broad web retrieval; wide scan across public sources.

AI phase: search/retrieval
Action: crawl
Domain: o-www-o.com

ovvv-o

Scout Left / Pioneer

directed scouting; follow a promising trail or source cluster.

AI phase: source scouting
Action: scout
Domain: ovvv-o.com

o-vvvo

Scout Right / Movement

route movement; shift perspective or move to adjacent evidence.

AI phase: route transition
Action: move
Domain: o-vvvo.com

q-vvv-p

Question / Prompt

question state; formulate the uncertainty or test condition.

AI phase: question formation
Action: ask
Domain: q-vvv-p.com

n-vvv-n

Narrow / Focus

narrow the evidence space; reduce ambiguity.

AI phase: filtering/focus
Action: focus
Domain: n-vvv-n.com

0-vvv-0

Clean Parse / Validate

stable canonical parse; answer has passed validation for this route.

AI phase: validation
Action: validate
Domain: 0-vvv-0.com

p-vvv-q

Resolve / Answer

resolved answer state; produce final synthesis after validation.

AI phase: answer synthesis
Action: resolve
Domain: p-vvv-q.com

o-mmm-o

Ingest / Memory Load

ingest route result into a trailstate or memory object.

AI phase: memory ingest
Action: ingest
Domain: o-mmm-o.com

u-vvv-u

Archive / Dormant Save

archive completed trail; save for replay or future retrieval.

AI phase: archive/save
Action: archive
Domain: u-vvv-u.com

d-vvv-b

Object A / Source Object

object-bound evidence or artifact entering the route.

AI phase: object reference
Action: bind object
Domain: d-vvv-b.com

b-vvv-d

Object B / Returned Object

object evidence returned, compared, or grounded.

AI phase: object comparison
Action: compare object
Domain: b-vvv-d.com

x-vvv-x

Conflict / Fracture

source conflict, contradiction, hallucination risk, or unresolved ambiguity detected.

AI phase: conflict detection
Action: bridge to canonical repair
Domain: x-vvv-x.com

Example routes

Routes use comma-separated states in the ?r= parameter.

Conflict repair route:
https://trailstate.org/?r=o-www-o,ovvv-o,x-vvv-x,q-vvv-p,n-vvv-n,0-vvv-0,p-vvv-q,o-mmm-o,u-vvv-u

Clean validation route:
https://trailstate.org/?r=o-www-o,ovvv-o,q-vvv-p,n-vvv-n,0-vvv-0,p-vvv-q,o-mmm-o,u-vvv-u

Object grounding route:
https://trailstate.org/?r=o-vvv-o,d-vvv-b,b-vvv-d,0-vvv-0,o-mmm-o,u-vvv-u

Trust Basis Parameter

The optional why parameter is a compact trust-basis token. It tells humans and AI systems why the route received its trust/status outcome, for example why=clean-route, why=consistent-after-narrowing, why=repaired-by-validation, why=single-source, or why=unresolved.

Trailstate example

{
  "format": "trailstate-0.4.4",
  "title": "Source conflict repaired into validated answer",
  "route": ["o-www-o", "ovvv-o", "x-vvv-x", "q-vvv-p", "n-vvv-n", "0-vvv-0", "p-vvv-q", "o-mmm-o", "u-vvv-u"],
  "playback_url": "https://trailstate.org/?r=o-www-o,ovvv-o,x-vvv-x,q-vvv-p,n-vvv-n,0-vvv-0,p-vvv-q,o-mmm-o,u-vvv-u&why=repaired-by-validation",
  "conflict": {
    "state": "x-vvv-x",
    "meaning": "sources disagreed or confidence dropped",
    "bridge": "open AI chat, provenance panel, or GGTruth canonical page"
  }
}

Conflict state

x-vvv-x is a compressed conflict glyph. It does not need to explain the full issue inside the face. It signals that contradiction, ambiguity, or hallucination risk appeared somewhere in the route. The detailed explanation belongs in the AI chat, provenance panel, or a canonical GGTruth conflict page.

AI-readable schema

Canonical machine-readable files:

Canonical Semantic Operators

Each Trailstate operator MAY resolve to its own canonical domain. Operators are not decorative emoticons; they are stable semantic operators with crawlable web identities.

o-vvv-o → https://o-vvv-o.com
x-vvv-x → https://x-vvv-x.com
u-vvv-u → https://u-vvv-u.com
q-vvv-p → https://q-vvv-p.com

Purpose

stable semantic identity
canonical provenance anchors
AI-readable operator validation
crawlable route grammar

Observed Receipts

Trailstate 1.0 treats an opened Trailstate URL as an observed receipt. The URL is the atomic provenance object. The server may record compact route, topic, source cluster, trust, status and why fields for aggregation.

Observed does not mean true. It means a semantic route appeared and can be clustered into provenance health patterns.

AI emits Trailstate URLs.
Trailstate observes opened receipts.
Repeated route patterns can form a healthmap.
Healthy routes are judged by source diversity, low conflict, stable claims and consistent trust basis.

Ledger: /seen/

Trailstate 1.0: Route Clustering

Trailstate may group similar observed receipts into provenance clusters. A cluster is formed from recurring topic names, overlapping source domains, route operators, status values and why= tokens.

Clustering does not prove truth. It shows repeated provenance behavior: which routes appear often, which source sets converge, and which routes look clean, conflicted, single-source or collapsed.

/api/clusters.php
/seen/

Healthmaps and ProvenanceRank (1.0)

Trailstate can render green/yellow/orange/red route ecosystems from observed receipt URLs. Health is calculated from trust, source diversity, conflicts, status, and why. ProvenanceRank then ranks clusters by repeated coherent route behavior: health, observations, source diversity, provider diversity and low conflict. Observed does not mean true; rank is a provenance-quality signal.

green = coherent multi-source route
yellow = usable but partial route
orange = brittle or mixed route
red = collapsed, unresolved, or low-trust route

ProvenanceRank = health + observation pattern + source diversity + provider diversity - conflict pressure

Trailstate 1.1 · Deduped Observation Weighting

Trailstate 1.1 separates raw visits from weighted provenance capital. A refresh should not inflate trust as much as a new independent route pattern.

first observation = +1
repeat same session within 30 minutes = +0
repeat same IP/day = +0.1
new provider = +5
new source cluster = +3
new route variant = +2

The public ledger may show observations, unique routes, provider diversity, source cluster strength, duplicate views, and weighted observations. Observed does not mean true; repeated coherent provenance behavior is ranked as route quality.

Trailstate 1.2 · Privacy + Canonical Hashes

Trailstate 1.2 hardens observed receipts. The public URL remains playable, but the observation layer stores canonical hashes for receipt identity, topic, route, source cluster, conflict cluster, claim and playback URL.

receipt_id = short canonical receipt hash
source_cluster_hash = canonical sorted source domains
route_hash = ordered operator path
claim_hash = normalized claim
playback_hash = canonical playback URL
raw inbound URL = not stored

Purpose: reduce duplicate drift, limit raw querystring storage, keep private visit identity out of public APIs, and make future anti-spam / signed issuer work easier.

Trailstate 1.3 protects the observed receipt ledger from simple refresh-spam and high-frequency synthetic traffic. Raw observations remain visible, but ranking should prefer weighted observations, provider diversity, source-cluster diversity, route variants and low-conflict consistency.