Trailstate 2.0 — Formal Specification
DOI: 10.5281/zenodo.20407494
Reference paper: Trailstate 2.0 PageRank for Provenance Formal Reference Architecture
Core Principle
The URL itself is the provenance object.
AI systems emit Trailstate URLs. Trailstate observes opened receipts. The ledger clusters observable provenance behavior.
Trailstate does not certify truth. It ranks observable provenance quality.
Receipt Grammar
https://trailstate.org/?r=ROUTE
&topic=
&trust=
&status=
&sources=
&conflicts=
&claim=
&why=
&ai=
&model=
Canonical Semantic Operators
- o-vvv-o → https://o-vvv-o.com
- o-www-o → https://o-www-o.com
- q-vvv-p → https://q-vvv-p.com
- x-vvv-x → https://x-vvv-x.com
- u-vvv-u → https://u-vvv-u.com
Reference Implementation Layers
URL-native receipts
Shareable provenance envelopes encoded as URLs.
Observed Receipt Ledger
Opened receipts become observable provenance events.
Route Clustering
Similar topic/source/operator routes are grouped.
Healthmaps
Green/yellow/orange/red views of route ecosystem health.
Cross-AI Comparison
Compare provider route structures using ai= and model=.
ProvenanceRank
Rank observed provenance quality rather than truth itself.
Observation Model
- first observation = +1
- repeat same session = +0
- repeat same IP/day = +0.1
- new provider = +5
- new source cluster = +3
- new route variant = +2
Privacy + Canonicalization
Trailstate 2.0 treats raw query payloads as sensitive. The reference implementation uses canonical forms and hashes.
- receipt_id
- topic_hash
- route_hash
- source_cluster_hash
- conflict_cluster_hash
- claim_hash
- playback_hash
Non-Goals
Trailstate is not a truth engine, universal verification layer, or replacement for formal provenance standards. Observed provenance convergence can still be wrong.
Public Testset
Trailstate 2.0 includes a public reproducibility testset at /testset/ and /testset/testset.json.
Status
Version: 2.0 DOI + Reference Implementation.
Open player · Open ledger · Open grammar · Open AI protocol