D-Open Code Sovereign Licence · Core Architecture

The World Has Never Had
a Library Like This.

Designed to hold over one million bibliographic records within forty megabytes of memory. Every field traceable to its source, human, machine, or collective consensus. The D-Code Model is the technical core of the World Library: the first cataloguing engine in history designed from the ground up for cryptographic provenance, permanent memory, and governmental-grade sovereignty.

1M+ Framework capacity
<40 MB Memory envelope
BIBFRAME Compatible
Native Blockchain storage
The Vision

Built for Civilization Scale.

This is not a library plugin. It is the architectural foundation for any governmental-grade decentralized application: public registries, civic records, sovereign data commons, and more.

Sovereign Licensing & Enforcement

Designed to run alongside the D-Open Code Sovereign Licence and natively enforce the D-SAFE Label. Data usage that violates ethical or legal boundaries is structurally impossible, not just prohibited in writing.

Protection of Public Domain Knowledge

The base state of every catalogue lives on Arweave permanent storage. Every edit is a cryptographically signed patch on an EVM blockchain. History cannot be quietly erased or altered.

International Law by Default

Strict schema versioning, per-field provenance tracking, and content-addressed storage provide built-in compliance with international archival, transparency, and data provenance laws, with no additional configuration required.

Heritage

Standing on Two Thousand Years of Library Science.

The D-Code Model descends from the international bibliographic modeling tradition: IFLA's FRBR/LRM entity analysis and the Library of Congress BIBFRAME 2.0 vocabulary. It does not replace these standards. It extends them into the blockchain era.

Every concept from the international standard maps directly to the model, and every generated application inherits full compatibility with the global library ecosystem automatically.

International Standard D-Code Model Realization
bf:Work: the intellectual creation Root ORM record (e.g. Book)
bf:Instance: physical / digital embodiment Embedded InstanceData structs, up to 32 per language (PDF, EPUB, scan, AI translation…)
bf:Item: holdings Arweave transaction IDs (content-addressed, permanent)
DDC / LCC classification Native deweyDecimal + locCallNumber properties
Language coding ISO 639-3 throughout, mapped to compact 8-bit wire identifiers
Agent relations (authorship) Typed, bidirectional authors / authorOf graph edges
The Refinement

One Record. All Realities.

Classical library systems assign a single canonical description to each work, encoding one cultural viewpoint. Consider an ancient text: its original language is Latin to one scholarly tradition, Greek to another, Ge'ez to a third. All three answers are correct. Classical systems force a choice. The D-Code Model does not.

One work record holds parallel, equally legitimate language realities simultaneously. book.getOriginalLang('eng') → 'lat' and book.getOriginalLang('amh') → 'gez' coexist without conflict. A single-language catalogue is simply the degenerate case of a universal one.

Architecture

Four Layers That Never Fail Each Other.

Each sub-system has one job, one invariant, and one contract with the layer above it. Correctness is not a target, it is a structural property.

Main Thread
01

Schema Generator

The CLI compiler that turns human-readable schema definitions into strongly-typed ORM classes. If a definition violates a structural rule, it is a build failure, not a runtime surprise discovered in production.

Compile-time enforcement Zero hand-written parsers
02

In-Memory Database

A 10,000-record LRU cache that fits within a mobile browser tab's memory budget. Lazy reactivity attaches only to records actually rendered on screen. Cache misses re-hydrate from the worker automatically, without blocking the UI.

<40 MB Mobile-first LRU eviction
Web Worker RPC
Background Worker
03

Binary Transport

Every edit becomes a 256-bit EVM word, a bit-packed binary patch encoding the operation, object type, object ID, language, and value. Edits are debounced and deduplicated: only the latest edit per field per object reaches the chain.

256-bit precision Idempotent Blockchain-ready
04

Sidecar Decoders

Columnar binary files that decode in O(n) time on the worker thread. The format is self-describing, it embeds its own field descriptor, so every historical catalogue file remains parseable across all future schema versions, indefinitely.

Self-describing format O(n) decode Forever compatible
Design Assurance

Verified Before It Shipped.

This architecture was not written first and rationalized later. It was specified, audited, and hardened before and during implementation.

01

Multi-Reviewer Audit

The schema compiler, codec, and runtime underwent repeated independent review rounds, two reviewers cross-auditing each other's output, with human arbitration at every disagreement. Every finding formally triaged and regression-tested.

02

Compile-Time Invariants

Architectural rules are not conventions, they are build failures. Redefining a base property, declaring conflicting layouts, reordering instance fields, or using unknown enum values all abort compilation with explanatory diagnostics.

03

Encode-Time Range Verification

Every fixed-width field of every bit-packed EVM word is range-checked at the moment of serialization. An out-of-range value throws rather than silently corrupting the adjacent bits of a 256-bit word.

04

Versioned Wire Protocol

Bit layouts are never hardcoded. They are resolved per payload from an append-only, delta-versioned layout dictionary. Every historical on-chain patch remains decodable forever, regardless of future schema evolution.

05

Built to Last

Code written with AI, audited by humans, and battle-tested in production. Every critical behaviour is locked down before it ships.

06

Written Governance

The non-negotiable compatibility rules, append-only enums, append-only version history, deploy-before-broadcast ordering, are codified in writing and enforced by guard comments at every edit entry point in the codebase.

Provenance

Every Field Knows Its Own History.

Most databases record what a value is. The D-Code Model records who set it, how it was established, and whether a human has verified it: at the individual field level, across every language simultaneously.

A dedicated provenance tier stores, per property and per language, one of four states:

Unset No value has been established for this field in this language
AI-Generated A machine produced a value, awaiting human review
Human-Verified A human expert has confirmed the value is correct
Community Consensus Multiple independent contributors have independently agreed

Two bits per field. The entire edit history of every record is reconstructible from the Arweave base state and the ordered on-chain patch log. Nothing is ever lost. Nothing can be silently changed.

Universal by Construction

Every Application Built on This Licence Is Already Compatible With Every Other.

The code generation pipeline is built directly into the D-Open Code Sovereign Licence. Compatibility is not a feature. It is a structural property of the licence itself.

Instant Interoperability

Anyone building a new application using this Licence automatically becomes 100% compatible with all data already published under it. No translation layers. No migration scripts. No negotiation required.

Universal Data Sharing

Applications can automatically share any newly published data across the entire ecosystem of existing applications. Zero friction. Zero coordination overhead. The network grows stronger with every participant.

Infinite Model Generation

Because the engine is schema-driven end to end, any future model (Person, Recording, PublicRecord) inherits the full verified design with zero hand-written parser or codec code.

Global Data Universality

By aligning on this generated schema foundation, we establish true worldwide universality in data structure, the prerequisite for a genuine global sovereign knowledge commons.

Open Recruitment

Code for the Next 10 Generations.

The World Library is built by its community. The best developers (those most aligned with the values of permanent, sovereign public knowledge) compete openly for a place on the build team. When you contribute, your name is recorded on the blockchain forever.

01

Compete on Merit

Register, choose a bounty from the public project list, fork the repo, and submit your solution. Automated pipelines verify compliance. Human judges rank efficiency and longevity. The leaderboard is public.

02

Smart-Contract Hiring

Teams are not hired by a manager: they are activated automatically when the community funds a project past a predefined threshold. No human gatekeeping. No politics. The treasury triggers the contract; the best-ranked team answers.

03

Recognized and Paid

The work of a recognized expert cannot be expected for free. Top contributors are offered paid contracts: Architect, Maintainer, and Legacy roles with monthly retainers and governance seats as the project grows.

04

Leave Your Mark

Every contribution is permanently recorded on-chain. When you build a block of the World Library, your name is written into it forever, visible to every reader for the next 250 years.

Join the Leaderboard ↗

All contributions are CC0. No tokens. No NFTs. Just permanent, public knowledge And the respect of a community that builds to last.

Available Now

This infrastructure is available to institutions, governments, and developers worldwide.

The D-Code Model and the D-Open Code Sovereign Licence are open to any organization committed to the permanent, sovereign preservation of public knowledge.