ENTITY Documentation
Architecture Guide

BTDU, knowledge graphs and vector indexes solve different parts of the information problem

A knowledge graph organizes explicit entities and relationships. A vector index supports similarity retrieval over numerical embeddings. BTDU instead focuses on reusable information atoms, bonds and compounds connected to ENTITY authority, provenance, rights and economic state. These approaches can coexist; they are not interchangeable.

Knowledge graph

Knowledge graphs are strong when systems need explicit nodes, typed relationships, ontologies and graph traversal. Their value comes from representing relational structure directly. They do not, by themselves, define sovereign identity, delegated authority, asset rights, settlement or recovery semantics.

Vector index

Vector databases and embedding indexes are strong when systems need approximate semantic similarity. They typically represent content as dense numerical vectors and retrieve nearby vectors. The index says that representations are similar under a model and metric; it does not inherently establish provenance, authority, title or economic entitlement.

BTDU

The Blackmore Technology Data Universe represents information using reusable atoms, bonds and compounds, while preserving governed relationships to ENTITY state. In the released v3.4.2 design, ADAM remains the deterministic atom/bond state engine, BTDU provides governed topology and indexes, ENTITY provides authority/rights/economics, and NIKI is a bounded reasoning projection rather than the source of authority.

LayerPrimary role
ENTITYAuthority, rights, provenance, economic state, recovery
ADAMDeterministic atom/bond state
BTDUGoverned information topology, compounds, indexes and manifests
NIKIBounded reasoning/projection over governed state

What BTDU is not

BTDU is not claimed as a universal replacement for databases, knowledge graphs, embeddings or raw-file storage. It is also not presented as generic compression. Controlled BTDU benchmarks compare specific semantic/index representations; they do not prove universal storage reduction for every workload.

Why the governance layer changes the question

A relation can be computationally useful without being legally or economically meaningful. ENTITY distinguishes temporary reasoning relationships from persistent governed relationships and requires authority before protected mutation. A topology edge therefore does not silently become ownership, permission or royalty.

Where combination makes sense

A deployment can use vectors for semantic retrieval, a graph for ontology/traversal, and BTDU/ENTITY for governed state, provenance and rights. The architectural question is not which database wins; it is which representation is authoritative for each kind of claim.

Challenge the boundary

External reviewers are invited to inspect src/40_BTDU/canonical_btdu.py and look for counterexamples involving authorization, exact-byte reconstruction, provenance separation, topology-created rights or state recovery.

Review BTDU independently →