Glossary

Name: Magellan — Audit-Ready Mapping & Metadata Control Layer
Availability: PreOrder

Plain-English definitions for the 35 terms used across the Magellan platform, the white paper, the feature pages, and the grant submissions. Each entry has a stable URL — link directly to /glossary#<term-slug>.

Filter:

Core Concepts

Entity#entity: A named collection of records that share the same structure. In Magellan, an entity is similar to a table in a relational database (e.g., the FHIR Patient resource or the OMOP person table). Entities are the unit of mapping at the schema level.
Attribute#attribute: A named, typed field on an entity. An attribute is similar to a column in a relational database (e.g., Patient.birthDate or person.year_of_birth). Attributes are the unit of mapping at the field level.
Schema#schema: A versioned definition of one or more entities and their attributes (e.g., FHIR R4, OMOP CDM v5.4, USCDI v3, i2b2 star schema). Magellan treats schemas as first-class, versioned assets.
Mapping#mapping: A versioned, machine-readable relationship between a source attribute (or entity) and a target attribute (or entity), including any transformation logic, value-set crosswalk, and provenance.
Knowledge Graph#knowledge-graph: The internal representation in which schemas, entities, attributes, mappings, value sets, and decisions are stored as nodes and edges, enabling semantic search and impact analysis across versions.
Value Set#value-set: A controlled vocabulary or code list (e.g., LOINC, SNOMED CT, RxNorm, ICD-10-CM) referenced by an attribute. Value-set alignment is a distinct mapping problem from attribute alignment.

Transformation Debt

Transformation Debt#transformation-debt: The accumulated cost of data-transformation logic that exists only as scripts, stored procedures, spreadsheets, and the memory of the people who wrote them. It cannot be inspected, diffed, or reviewed, so every new integration re-pays the same work and every audit question becomes reconstruction. Magellan retires it by making the mapping a versioned, approved artifact. See Transformation Debt, and the Record That Pays It Off.
Technical Debt (Data Transformation)#data-technical-debt: The slice of technical debt that code reviews never reach: debt that accumulates every time data is transformed without a governed record. Magellan addresses seven forms of it — see The Technical Debt Magellan Retires. It does not address application-code, infrastructure, or data-quality debt.
Re-mapping Debt#re-mapping-debt: The repeated cost of rebuilding a mapping when a source or target schema changes because nobody recorded which schema versions the original mapping was built against. Retired in Magellan because each mapping version pins the exact source schema(s) and target schema(s) versions it references.
Undocumented-Logic Debt#undocumented-logic-debt: Transformation rules buried inside ETL jobs or pipeline code, where they are a side effect rather than an artifact. Retired when the mapping itself is the governed, reviewable record.
Audit Archaeology#audit-archaeology: Forensic reconstruction of who decided a mapping, when, and why — performed after an auditor or regulator asks. Replaced in Magellan by a ULID-keyed, SHA-256 hash-chained audit ledger and exportable evidence packs (the next module being built).
Mapping as Artifact#mapping-as-artifact: Magellan's core principle: a mapping is a versioned, approved, reproducible decision that can be inspected years later — not a byproduct of a pipeline or a row in a spreadsheet.

Decisions & Audit

Semantic Match#semantic-match: Magellan's mapping engine that proposes candidate attribute-to-attribute (and entity-to-entity) correspondences using embeddings, lexical signals, and value-set evidence, then scores each candidate with a calibrated confidence.
Confidence Score#confidence-score: A calibrated probability (0.00–1.00) assigned to each proposed mapping. Scores above the high-confidence threshold may be auto-accepted under policy; scores in the review band are routed to a human informaticist.
Decision#decision: A discrete, reviewable event in which a proposed mapping is accepted, modified, rejected, or deferred. Every decision is written to the audit ledger.
Audit Ledger#audit-ledger: An append-only, hash-chained record of every decision, including the inputs (source/target schema versions, candidate set, confidence), the actor (human or model), the policy version in force, and a SHA-256 link to the prior entry.
Hash Chain#hash-chain: A cryptographic technique in which each ledger entry includes the SHA-256 hash of the previous entry, so any retroactive modification breaks the chain and is detectable on verification.

Identifiers & Provenance

ULID#ulid: Universally Unique Lexicographically Sortable Identifier. A 128-bit identifier whose leading bits encode a millisecond timestamp, so identifiers sort in creation order. Magellan uses ULIDs (never UUIDs) for all decision and ledger identifiers.
Provenance#provenance: The recorded lineage of a decision or data element: who or what produced it, from which inputs, under which policy and model version, and at what time.

Health Standards

USCDI#uscdi: United States Core Data for Interoperability. The HHS-designated minimum set of health data classes and elements for nationwide interoperability.
FHIR#fhir: Fast Healthcare Interoperability Resources. The HL7 standard for exchanging healthcare information electronically; FHIR R4 is the version referenced in this proposal.
OMOP CDM#omop-cdm: Observational Medical Outcomes Partnership Common Data Model, maintained by OHDSI. The standard analytic schema used by research networks; v5.4 is the version referenced in this proposal.
i2b2#i2b2: Informatics for Integrating Biology and the Bedside. A widely deployed clinical research data warehouse and star-schema data model.
FAIR#fair: Findable, Accessible, Interoperable, Reusable — the data-stewardship principles published by Wilkinson et al. (2016) and adopted by NIH and NSF for funded research data.

Regulatory & Policy

21 CFR Part 11#21-cfr-part-11: The U.S. Food and Drug Administration regulation governing electronic records and electronic signatures, applicable to FDA-regulated clinical, manufacturing, and post-market data systems.
HIPAA §164.312(b)#hipaa-164-312-b: The HIPAA Security Rule technical safeguard requiring covered entities to implement hardware, software, and procedural mechanisms that record and examine activity in information systems containing electronic protected health information.
CMMC L2 §3.3#cmmc-l2-3-3: Cybersecurity Maturity Model Certification, Level 2, Audit & Accountability family — the Department of Defense framework that operationalizes the 14 control families in NIST SP 800-171 for any contractor handling Controlled Unclassified Information (CUI). The §3.3 family (Audit & Accountability) requires that systems create, protect, and retain the audit records needed to monitor, analyze, investigate, and report unlawful or unauthorized activity. Magellan's hash-chained Audit & Trust Ledger satisfies the four core §3.3 controls: AU-2, AU-3, AU-9, and AU-11. Magellan does not claim CMMC certification — it produces the AU-family evidence a prime contractor's assessor needs.
CMMC §3.3 AU-2 — Event Logging#cmmc-au-2: Audit & Accountability control requiring the organization to identify the types of events the system must log (e.g., authentication, privilege use, data access, configuration change) and to keep that event catalog under review. In Magellan, every accepted/modified/rejected mapping decision, every schema or mapping version change, every threshold-profile update, and every actor authentication is a logged event class written to the ledger.
CMMC §3.3 AU-3 — Content of Audit Records#cmmc-au-3: Requires each audit record to capture enough detail to reconstruct the event: what happened, when (timestamp), where (system/component), the source, the outcome, and the identity of the associated user or process. Magellan ledger entries carry a ULID (millisecond-precise, lexicographically sortable), source/target paths and ids, confidence score, algorithm class, active threshold-profile version, model identifier, API version, and the authenticated actor with role-at-time-of-action.
CMMC §3.3 AU-9 — Protection of Audit Information#cmmc-au-9: Requires audit information and audit tools to be protected from unauthorized access, modification, and deletion. Magellan satisfies AU-9 with an append-only, SHA-256 hash-chained ledger: each entry binds the SHA-256 of the prior entry, so any retroactive insert, edit, or delete breaks the chain and is detectable on verification. Ledger writes are scoped by RLS and role-at-time-of-action.
CMMC §3.3 AU-11 — Audit Record Retention#cmmc-au-11: Requires audit records to be retained for a defined period consistent with records-retention policy to support after-the-fact investigations and to meet regulatory requirements. Magellan's ledger is append-only by design — entries are never overwritten or pruned — and retention windows are configurable per tenant to align with HIPAA, 21 CFR Part 11, and DoD CUI requirements.
NIST SP 800-53 Rev. 5#nist-sp-800-53-rev-5: The National Institute of Standards and Technology catalog of security and privacy controls for federal information systems and organizations, referenced by FedRAMP and many federal authorizations.
OMB M-24-10#omb-m-24-10: The Office of Management and Budget memorandum on advancing governance, innovation, and risk management for federal agencies' use of artificial intelligence, applicable to AI systems supporting federal missions.
Zero Trust#zero-trust: A security model that assumes no implicit trust between components and verifies every access request; relevant federal directives include OMB M-22-09 and CISA's Zero Trust Maturity Model.

Programs

SBIR#sbir: Small Business Innovation Research. The federal program under which Magellan's NIH Phase I application is being prepared for submission; Phase I funds feasibility, Phase II funds development, Phase III is commercialization.