| Field | Value |
|---|---|
| Standard | Lexical Markup Framework (LMF) — Part 1: Core model |
| ISO Document | ISO 24613-1:2024 |
| ISO Adoption | ISO/IEC JTC 1 / TC 37 / SC 4 — Language resource management |
| Multi-part Family | ISO 24613-2:2024 (MRD), -3:2024 (ETF), -4:2024 (TEI), -5 (LBX), -6:2024 (Syntax & Semantics), -7 (LBR) |
| Authority | International Organization for Standardization under TC 37/SC 4 (Terminology and other language and content resources) |
| npm Package | @amlhubs/lmf-core |
| npm Version | 0.0.1 |
| Peer Dependencies | @amlhubs/uml@^0.0.2, @amlhubs/mof@^0.0.1 |
| License | MIT |
ISO 24613-1:2024 — Language resource management — Lexical Markup Framework (LMF) — Part 1: Core model — is the international standard that defines the abstract metamodel for representing machine-readable lexical resources. LMF is the reference model that grounds every modern monolingual, bilingual, and multilingual computational lexicon, NLP morphological lexicon, and machine-translation termbase that aspires to ISO compliance. The 2024 edition is the multi-part successor to the foundational ISO 24613:2008 first edition, expanded by a 60-expert international panel into seven complementary parts, of which Part 1 — the Core model implemented here — is the load-bearing foundation that every other part extends through the LMF package-merge mechanism.
The @amlhubs/lmf-core npm package repackages the LMF Core Model as extensible TypeScript interfaces and base classes following the @amlhubs Three-Layer Pattern: every metaclass surfaces as IFoo<T> (generic interface), AbstractFoo<const T> (inference-site abstract class), and Foo{Instance} (registered concrete). The package surfaces the Core metaclasses spanning the four LMF Core packages — Information (LexicalResource, GlobalInformation), Lexicon (Lexicon, LexicalEntry, Form, FormRepresentation), Sense (Sense, SenseRelation, Definition, Statement, SenseExample), and CrossLingual (Equivalent, Translation, TranslationSet, MonolingualExternalRef) — together with the LMF graph-of-features substrate (FeatureStructure, FeatureValue, DataCategory) on which every annotation in the Core depends. Every interface carries a JSDoc header citing the precise ISO 24613-1:2024 clause and section that defines it, making each symbol an auditable projection of the ISO standard rather than an internal invention.
Adopting LMF through a typed package produces an immediate interoperability dividend. Every modern lexical-resource management platform — ELEXIS infrastructure, Lexonomy, DBnary, the European Language Grid, the LREC/LRE Map deposits, and every academic-press dictionary digitization project — converges on the LMF metamodel as the machine-readable target for monolingual and multilingual lexicons. A lexicon expressed against the metaclasses @amlhubs/lmf-core exports is, by construction, portable to every one of those platforms without a custom converter. Ventures that would otherwise spend quarters re-implementing per-platform lexicon exporters — a Wiktionary-to-LMF converter, a Princeton WordNet-to-LMF converter, a custom Wolof-French lexicon-to-LMF converter — amortize that engineering cost to zero.
ISO adoption turns internal architectural decisions into regulator-recognized artifacts. ISO 24613 sits inside the ISO/TC 37 (Terminology and other language and content resources) family that includes ISO 12620 (Data Categories), ISO 16642 (Terminological Markup Framework, TMF), ISO 24611 (MAF — Morpho-Syntactic Annotation Framework), ISO 24612 (LAF — Linguistic Annotation Framework), and ISO 24617 (SemAF — Semantic Annotation Framework). Government-procurement frameworks for terminology services, EU translation-memory contracts, and the W3C Best Practices for Multilingual Linked Open Data all cite ISO 24613 as the metamodel of record. A regulated vendor — a public-administration translation service subject to EU multilingualism requirements, a healthcare-terminology platform subject to ISO 17115 (Categorial structures), a legal-translation service subject to national-language statutes — presents the same LMF surface to an auditor without translating internal jargon into ISO language.
The second business lever is agentic runtime leverage. Ageni's Probabilistic Reduction Engine consumes the LMF metamodel as the deterministic substrate over which large-language-model reasoning operates on lexical resources. When an agent writes source code against ILexicalEntry, IForm, ISense, and IDefinition, the TypeScript compiler evaluates whether the proposed lexical structure is a well-formed LMF lexicon at the same moment the compiler evaluates whether the code itself is well-formed — the two correctness checks collapse into one tsc pass. Structural hallucinations that would otherwise slip past a natural-language review (inventing a Sense without a parent LexicalEntry, mis-attributing a Form to a Sense rather than a LexicalEntry, violating the FeatureStructure containment constraint) are caught at compile time, and every surviving interface reference traces to a §-section of the ISO standard through the JSDoc header. The agent's reasoning surface is reduced from the open set of possible English sentences about lexicons to the closed set of typed LMF metaclass compositions.
The third lever is compounding reuse across the AML linguistic stack. LMF Core is the load-bearing foundation of the lexical-resource layer in the AML graph: future @amlhubs/lmf-mrd (Part 2 — Machine Readable Dictionary), @amlhubs/lmf-etf (Part 3 — Etymological Form), @amlhubs/lmf-tei (Part 4 — TEI serialisation), @amlhubs/lmf-syntax-semantics (Part 6 — Syntax and Semantics), @amlhubs/maf (ISO 24611 — Morpho-Syntactic Annotation Framework), @amlhubs/laf (ISO 24612 — Linguistic Annotation Framework), @amlhubs/semaf (ISO 24617 — Semantic Annotation Framework), and @amlhubs/ontolex (W3C OntoLex-Lemon) all extend or interoperate with the LMF Core metaclasses surfaced here. Every downstream ageni venture that reasons about words, lemmas, lexical entries, definitions, or translations consumes these same LMF Core metaclasses through their transitively-dependent packages. The engineering investment that produced @amlhubs/lmf-core is not charged to any single venture; it is amortized over every venture that ever extends it.
The fourth lever is composability across the OMG/ISO specification stack. LMF Core extends UML 2.5.1: every LMF metaclass realizes a UML IClass carrying owned attributes typed by UML IDataType/IPrimitiveType and connected by UML IAssociation instances. The @amlhubs/lmf-core package depends on @amlhubs/uml for those structural types and on @amlhubs/mof for the reflective machinery that lets agents query a lexicon's metamodel-shape at runtime. Owning the typed packages from UML through MOF through LMF — rooted in the single UML foundation that underpins the entire AML graph — gives the ageni platform one coherent metamodeling surface for both code and lexical resources rather than two loosely-coupled specifications, and makes every future agentic capability that touches a lexicon a composition of capabilities already expressed in the same formal language.
The package exports the LMF Core Model metaclasses grouped by ISO 24613-1:2024 clause. The complete enumeration lives in src/lmf-core.ts; the table below summarizes the groups and cites the authoritative clause.
| LMF Core Package | Clause | Metaclasses Surfaced |
|---|---|---|
| Information | §6 | ILexicalResource, IGlobalInformation |
| Lexicon | §7 | ILexicon, ILexicalEntry, IForm, IFormRepresentation, ILemma, IWordForm, IStem, IRelatedForm, IListOfComponents, IComponent, IListOfFormRepresentations |
| Sense | §8 | ISense, IDefinition, IStatement, ISenseExample, ISenseRelation, IPredicativeRepresentation, ISubjectField |
| CrossLingual | §9 | IEquivalent, ITranslation, ITranslationSet, IMonolingualExternalRef |
| Substrate | §5 | IFeatureStructure, IFeatureValue, IDataCategory, ILMFClass (root) |
Every interface is accompanied by an extensible base class with the same name minus the I prefix (e.g., LexicalResource, Lexicon, LexicalEntry). The full list and the JSDoc headers citing each ISO clause live at src/lmf-core.ts.
@amlhubs/lmf-core is a downstream consumer of UML and MOF. It depends on both. Future LMF Parts depend on it.
@amlhubs/uml (root, zero dependencies)
▲
│ peerDependency
├── @amlhubs/mof (reflective machinery over UML)
│ ▲
│ │ peerDependency
│ │
│ └─── @amlhubs/lmf-core (this package — ISO 24613-1:2024)
│ ▲
│ │ (future)
│ ├── @amlhubs/lmf-mrd (ISO 24613-2)
│ ├── @amlhubs/lmf-etf (ISO 24613-3)
│ ├── @amlhubs/lmf-tei (ISO 24613-4)
│ ├── @amlhubs/lmf-syntax-semantics (ISO 24613-6)
│ └── @amlhubs/maf (ISO 24611)
│
└── peerDependency reused by lmf-core for IPackage / IClass / IDataType
npm install @amlhubs/lmf-coreimport type {
ILexicalResource,
ILexicon,
ILexicalEntry,
IForm,
ISense,
IDefinition,
} from '@amlhubs/lmf-core';
// Declare an LMF Lexicon as a typed metamodel instance.
const myLexicon: ILexicon = {
elementId: 'LMF_Lexicon_FrenchCore',
name: 'French Core Lexicon',
language: 'fra',
lexicalEntryIds: ['LMF_LE_chien', 'LMF_LE_chat'],
// ... remaining LMF Core attributes
};The source artifact is src/lmf-core.ts. Every interface JSDoc header declares @standard ISO 24613-1:2024 and a @section §x.y reference.
- ISO 24613-1:2024 — LMF Part 1: Core model
- ISO 24613-2:2024 — LMF Part 2: Machine-Readable Dictionary (MRD)
- ISO 24613-3:2024 — LMF Part 3: Etymological extension (ETF)
- ISO 24613-4:2024 — LMF Part 4: TEI serialization
- ISO 24613-6:2024 — LMF Part 6: Syntax and Semantics
- LMF official site
- ISO/TC 37/SC 4 — Language resource management
| Version | Date | Change Summary |
|---|---|---|
| 0.0.1 | 2026-05-03 | Initial publish — ISO 24613-1:2024 LMF Core Model: Information, Lexicon, Sense, CrossLingual packages plus the LMF graph-of-features substrate. Three-Layer Pattern (IFoo / AbstractFoo / Foo{Instance}). Pure LMF Core; ETF, MRD, TEI, LBX, LBR (Parts 2-6) reserved for follow-on packages. |
MIT — public npm access under @amlhubs scope at npm.pkg.github.com.