Common Graph

Identity System / Design

Foundational architecture proposal — "Frames All the Way Down" — for a peer-to-peer network in which every assertion is a signed semantic frame, every coherent thing is an item, and the network is organized by trust rather than servers. The proposal rests on two pillars: (1) a semantic commons grounded in fifty years of computational linguistics (WordNet, CILI, FrameNet, VerbNet) used as the substrates data model — keys are grounded meanings rather than strings, with meaning resolved at write time by the asserter; (2) a peer-to-peer layer in which trust drives routing, moderation, replication, code execution, and discovery — generalized from the PGP web of trust. Structural properties claimed include peer review on every public assertion, moderation as views rather than policy, applications interchangeable on shared data, and software distributed through a trust graph rather than a registry. Note: domain specificity criterion read broadly here — Common Graphs primary contribution is to the data + trust layer rather than to identity strictly, but the architectural-proposal-with-paper-and-early-implementation profile maps cleanly onto ID Sys

Community

Details

License Reference implementation at github.com/joshualibrarian/common-graph (license per repository); paper itself CC BY 4.0
Dev Status 🔨 WIP
Dev Status Detail Pre-spec / Concept — 36-page paper published Apr 2026 (CC BY 4.0); reference implementation in active development; explicitly noted as "not ready for use beyond development"
Owner Joshua Chambers (independent researcher / developer; joshualibrarian on GitHub)
Country Not specified
Start Year 2026
Stack Reference implementation in active development; details not finalized
Funding Independent / not specified
Last Investigated Jul 1, 2026

Use Case Domains

Identity System / Design Attributes

Origins Joshua Chambers — independent research synthesizing computational linguistics and PGP-style web of trust
Database TBD — Common Graph proposes a substrate where every assertion is a signed semantic frame and every coherent thing is an item; concrete database/storage model is part of the reference implementation under active development
Query Language TBD — frames as semantic data structure suggest graph-query semantics; specifics in the reference implementation rather than the paper
Data Formats Signed semantic frames (proposed); items as coherent things; meaning grounded in WordNet / CILI / FrameNet / VerbNet vocabularies rather than strings
Mobile Support TBD — design is platform-agnostic; implementation status pre-deployment
Web Support TBD — implementation status pre-deployment
Native Apps TBD — no shipping reference application; reference implementation under active development per release notes
Terms Paper CC BY 4.0; reference implementation per repository
Funds Unknown — Joshua Chambers as independent researcher; specific funding for ongoing work not publicly disclosed
Based On PGP web of trust (generalized); WordNet, Collaborative Interlingual Index (CILI), FrameNet, VerbNet (semantic commons foundation); public-key cryptography
Authentication & Identity Public-key cryptography; trust-graph-based authentication where each runtime computes its own trust scores rather than relying on a global trust authority
Storage Model Distributed peer-to-peer with trust-driven replication; each peer stores what its trust graph indicates is relevant
Interoperability Designed for application interchangeability on shared data — applications operate over a common semantic graph rather than maintaining isolated databases
Data Portability By architectural commitment — data not tied to applications because meanings are grounded in shared computational-linguistics substrate (WordNet, CILI, FrameNet, VerbNet) rather than platform-defined schemas; trust-anchored authorship
Governance & Decision Making Trust-graph-based — moderation as views computed by each runtime rather than as centralized policy; each runtime computes its own trust scores; no global trust score; no platform-defined schema; peer review on every public assertion as a structural property of the design
Identity Standards Self-defined; rooted in PGP web-of-trust generalization rather than W3C DID or other identity standards
DID Methods Supported TBD / N/A — the proposal questions the platform-mediated identifier paradigm itself rather than implementing a DID method; could potentially bridge to DID methods if conformance becomes a priority but this is not the architectural direction
Key Management PGP-derived patterns — web-of-trust key verification generalized to drive routing, moderation, replication, code execution, and discovery; specifics in the reference implementation
Credential Types Signed semantic frames as the universal credential primitive — every assertion is a signed frame; every coherent thing is an item; uniformity at the data-model level rather than separate credential schema systems
Verification Method Cryptographic signature on every assertion (semantic frame); trust-graph traversal for endorsement-based verification
Privacy Features Per-runtime trust computation (no global trust score); routing privacy claimed (paper includes detailed analysis)
Authentication Methods Public-key signatures on semantic frames; trust-graph-based identity assertions
Revocation Mechanism Trust-graph-based — revocation expressed as trust-graph updates rather than registry-based revocation lists; each runtime's local trust computation reflects updated trust assertions
Agent Types Supported Humans (primary use case); applicable to any party with a key pair and trust relationships; potentially extensible to institutions and software agents
Wallet/Client Types TBD — no shipping wallet/client; reference implementation pre-deployment
Recovery Mechanisms TBD — PGP-style key management precedents available; specifics in the reference implementation
Compliance / Regulations Not evaluated — pre-spec architectural proposal at concept stage
Credential Exchange Protocols Frames-as-credentials — every assertion is a signed semantic frame; exchange happens via the trust-organized P2P layer; specific protocol bindings TBD
Trust Framework Web-of-trust generalized to drive all major network functions (routing, moderation, replication, code execution, discovery) — no central authority, no platform-defined schema, no global trust score
Cost Model Free — paper licensed CC BY 4.0; reference implementation open source; no platform fees by architectural commitment
Censorship Resistance Inherent to design — trust-driven routing replaces server-mediated routing; moderation is local view-based rather than policy-enforced