Standards

Library classification that sorts knowledge into ten main classes with decimal subdivisions, first published by Melvil Dewey in 1876.

Library classification adapted from Dewey's system by Otlet and La Fontaine to index all recorded knowledge; still maintained today.

S.R. Ranganathan's library classification, which builds subject numbers from facets joined by colons.

Sequential number given by the US military to every object it tracks in orbit, starting with Sputnik 1's rocket (1) and satellite (2). It ran out of five digits on 11 July 2026, forcing the 'Alpha-5' workaround and a move to nine-digit formats.

Identifier for a spacecraft by launch year, launch number and piece, such as 1990-037B for the Hubble Space Telescope; in use since 1963, replacing a Greek-letter scheme from Harvard.

MAchine-Readable Cataloging: formats for library catalogue records, developed by Henriette Avram at the Library of Congress between 1965 and 1968 so that libraries could share records by computer; a US national standard by 1971 and international by 1973. MARC 21 (1999) is still the predominant format.

The US grocery industry's product barcode and numbering standard, chosen on 3 April 1973 from IBM's design; the first scan was a pack of Wrigley's gum in Troy, Ohio, on 26 June 1974.

GS1's family of product numbers, joining the American UPC with the European Article Number from 1977; it is the number in retail barcodes and the base of the GS1 system.

NORAD's fixed-width text format for an object's orbit, read with the SGP4 model published in 1980; still the most shared orbital data, though its five-digit catalogue field overflowed and the CCSDS Orbit Mean-Elements Message has been used alongside it since 2020.

12-character code identifying a security for clearing, reporting and settlement, first used in 1981 and standardised as ISO 6166 after the G30 recommended it in 1989.

The geodetic reference system of 1984 behind GPS: the shape of the Earth and origin against which most latitudes and longitudes on the Web, from the W3C Geo vocabulary to GeoJSON, are given.

Guidelines for encoding texts for scholarly study, from a 1987 meeting of humanities and computing scholars at Vassar College; P3 appeared in 1994, P4 moved to XML in 2002 and P5 followed in 2007. It is the standard markup of the digital humanities.

The UN's standard messages and syntax for electronic data interchange between businesses, such as orders and invoices, approved as ISO 9735 in 1987 after merging with the American proposals: machine-readable business documents a decade before the Web.

HyperText Markup Language, the markup of Web pages, whose anchor element carries the Web's links. Berners-Lee first described it in the informal 'HTML Tags' document in late 1991; HTML 2.0 was published as RFC 1866 in November 1995.

Hypertext Transfer Protocol: the Web's request–response protocol for fetching a resource by its URL. Berners-Lee's first version was written up in 1991; HTTP/1.0 and 1.1 followed as IETF RFCs. Looking up a URI to get data about the thing it names depends on it.

System for assigning and resolving persistent identifiers, developed and operated by CNRI; DOIs are Handle System handles under the prefix 10.

ISO's family of standards for describing manufactured products so CAD, manufacturing and archive systems can exchange them, begun in 1984 and first published in 1994–95; its EXPRESS schema language is an early example of a formal shared data model across an industry.

Format for describing the structure and metadata of websites, developed by R.V. Guha at Apple and later submitted to the W3C with Tim Bray as an XML application.

Persistent Uniform Resource Locator: a web address that points at a resolver instead of the resource, so the target can move without breaking links. OCLC introduced it in 1995 and ran purl.org, where Dublin Core and many other vocabularies keep their namespaces, until it handed the service to the Internet Archive in September 2016.

Small set of metadata elements, such as title, creator and date, for describing resources on the Web; it began at a 1995 workshop in Dublin, Ohio.

Simple HTML Ontology Extensions: an early language for writing ontologies and annotating web pages with machine-readable claims, developed at the University of Maryland around 1996; among the first ontology languages made for the Web.

Platform for Internet Content Selection: W3C labels describing the content of web pages, made to let parents filter what children see; the label specification became a Recommendation on 31 October 1996. The RDF specification calls it a predecessor of RDF.

Markup standard for archival finding aids, begun at Berkeley in the early 1990s and released in 1998; it is maintained by the Society of American Archivists and the Library of Congress, with EAD3 from 2015.

Functional Requirements for Bibliographic Records: IFLA's conceptual model of the bibliographic universe as entities and relationships, separating the work from its expressions, manifestations and items; its final report was published in 1998.

Format for business and financial reports whose taxonomies define each reporting concept and its relations, begun by Charles Hoffman in 1998; regulators, securities commissions and tax agencies worldwide receive filings in it, and Inline XBRL embeds the tags in an ordinary HTML report.

W3C markup language for structured documents and data, a Recommendation since February 1998; RDF's first official syntax was written in it.

Ontology for museum and cultural heritage documentation from ICOM's documentation committee, first issued in 1999 and an ISO standard since 2006. It describes objects through the events in their history, such as production, acquisition and movement, so that records from different institutions can be integrated.

Data model that describes resources as subject–predicate–object statements identified by URIs. It became a W3C Recommendation in 1999 and was revised as RDF 1.1 in 2014.

Family of XML formats for syndicating site updates. Netscape's RSS 0.90 (1999) and RSS 1.0 were RDF-based; RSS 2.0 is plain XML.

Persistent identifier for documents and data, introduced in 2000. A DOI names the object rather than where it is stored, and resolves through doi.org; it is built on the Handle System and fits within the URI system.

RDF vocabulary for describing people, their online accounts and who they know; one of the first widely used Semantic Web vocabularies.

ISO standard for representing knowledge as topics, associations and occurrences; similar in expressive power to RDF and developed alongside it.

Ontology language formed in 2001 by merging the American DAML with the European OIL; a thin layer over RDF Schema with description-logic semantics, and the direct basis of OWL.

Logic-based clinical terminology of medical concepts and their relations, released in January 2002 and maintained by SNOMED International.

W3C Recommendation of 16 April 2002 for websites to state, in machine-readable form, what they do with visitors' data, so browsers could compare it with users' preferences. Few sites or browsers implemented it, and the W3C made it obsolete in 2018.

ISO standard for integrating the life-cycle data of process plants, including oil and gas facilities, with a generic data model (Part 2, 2003) and reference data libraries; Part 8 implements it in OWL and RDF.

Microformat from 2003 for marking up how the person behind a link relates to you, such as friend, colleague or met, using the rel attribute; Google's Social Graph API crawled it.

Small RDF vocabulary from January 2003 for points with latitude, longitude and altitude, made in the W3C Semantic Web Interest Group; never a W3C standard, it became the most used way to give a thing a position in RDF.

XML standard for legislative, judicial and executive documents, begun in 2004 by the UN for African parliaments and adopted by OASIS as LegalDocML in August 2018; its identifiers follow the FRBR model of work, expression and manifestation.

W3C language for writing ontologies on the Web, based on description logics. OWL 2 followed in 2009.

W3C vocabulary for describing classes and properties of RDF data; the first schema language for RDF.

W3C Technical Architecture Group's account of the Web's architecture: identifying resources by URI, interacting with them through representations, and data formats; a W3C Recommendation since 15 December 2004.

Conventions for marking up people, events, reviews and similar data with plain HTML class names, without a separate data format.

Protocol created by Brad Fitzpatrick in 2005 that lets people sign in to many sites with an identity they control, a URL rather than a site-specific account; later versions built on OAuth.

Open Knowledge Foundation's definition of 'open' for content and data: anyone can freely access, use, modify and share it, subject at most to requirements that preserve provenance and openness. First drafted in 2005.

Formal ontology joining the librarians' FRBR model with the museums' CIDOC CRM, begun in 2006, so that bibliographic and museum records share one model.

RDF vocabulary for music, first released in 2007, describing artists, works, recordings and releases, and the performances and recordings that connect them as events in time. It reuses FRBR's split between a work and its expressions, and FOAF for people.

GS1 standard, first ratified in April 2007, for sharing events in a product's journey through supply chains: what was seen, when, where and why. Version 2.0 (June 2022) added JSON-LD, a REST API and GS1 Digital Link URIs.

RDF vocabulary for describing posts, threads, forums and their authors, so that conversations on blogs, forums and mailing lists can be linked across sites; a W3C Member Submission of June 2007.

JSON format for geographic features, first specified by a community group in 2008 and standardised as RFC 7946 in August 2016, which dropped other coordinate systems in favour of WGS 84 alone; the developers' favourite, with no RDF in it.

Martin Hepp's web vocabulary for products, prices and offers, adopted by Google, Yahoo, Bing and retailers such as Best Buy, and merged into schema.org in November 2012 as its commerce vocabulary.

Query language for RDF data, a W3C Recommendation since January 2008; SPARQL 1.1 followed in 2013.

Creative Commons' RDF vocabulary for describing licences and attaching them to works, written into web pages with RDFa; submitted to the W3C in May 2008.

W3C syntax for embedding RDF statements in HTML attributes, first a Recommendation in October 2008.

Creative Commons tool, released in 2009, by which a rightsholder waives all copyright and database rights in a work as far as the law allows, putting it in effect into the public domain. Wikidata's structured data is published under it.

Attributes for nesting machine-readable data in HTML, added to the WHATWG's HTML specification in 2009 as a simpler alternative to RDFa and microformats; schema.org launched with it in 2011.

W3C vocabulary for publishing thesauri, classification schemes and subject headings as Linked Data; a Recommendation since August 2009.

Second version of OWL, adding profiles such as OWL 2 EL, QL and RL for tractable reasoning; a W3C Recommendation since October 2009.

Facebook's 2010 vocabulary of meta tags that turn a web page into a rich object when shared; the first version was based on RDFa.

The film and television industry's registry of identifiers, founded in October 2010 by MovieLabs, CableLabs, Comcast and Rovi. Every title, edit and version gets a DOI, so studios, broadcasters and streaming services can refer to the same thing.

Identifier for court decisions across the EU, in the form ECLI:country:court:year:number, with a minimum set of shared metadata; set out in EU Council conclusions published in April 2011.

Shared vocabulary for structured data in web pages, launched in 2011 by Bing, Google and Yahoo, with Yandex joining later that year.

Library of Congress model for bibliographic data based on Linked Data, designed to replace the MARC formats; work began in 2012 and version 2.0 followed in 2016.

OGC standard, version 1.0 from 2012 and 1.1 from January 2024, that adds a vocabulary for geometries to RDF and spatial functions and relations to SPARQL, so queries can ask what is within, near or touching what.

Shared web APIs for digitised images, proposed in late 2011 by research libraries including the British Library, Stanford, the Bodleian and the BnF. Its Image API (2012) serves zoomable images, and its Presentation API (2013) describes books and artworks as JSON-LD manifests, so any viewer can show any library's images.

Twitter's meta tags, introduced in 2012, that control how a linked page is previewed in posts; like Open Graph, a platform vocabulary that most websites publish.

Identifiers and an ontology for national and EU legislation, so each law and version has a stable web address and machine-readable metadata; the EU Council invited its adoption in October 2012.

20-character identifier for every company, fund or government body that takes part in financial transactions, defined in ISO 17442. The G20 called for it in 2011 after the 2008 crisis showed that regulators could not tell who was exposed to whom; the first LEIs were issued in December 2012. Its records say who is who and who owns whom, and are published openly, including as RDF.

Second version of SPARQL, adding updates, aggregates, subqueries and federated queries; a W3C Recommendation since March 2013.

W3C family of specifications for recording provenance: who made something, from what, and how. PROV-DM is the model, PROV-O its OWL ontology; both 2013.

IETF binary data format with a data model close to JSON's, designed for small code and message size on constrained devices; first published as RFC 7049 in October 2013 and revised as Internet Standard RFC 8949 in December 2020.

OWL ontology of financial contracts, instruments, parties and their relations, begun after the 2008 crisis as a shared vocabulary for analysis and regulatory reporting; developed by the EDM Council with the Object Management Group and in an open community process since 2020.

Open identifier for financial instruments of every kind, from shares to currencies; it began as Bloomberg's open symbology in 2009 and became an Object Management Group standard in 2014, free to use, unlike most proprietary securities codes.

W3C vocabulary for describing datasets in data catalogues, used by government open-data portals; a Recommendation since January 2014.

JSON-based syntax for Linked Data that lets ordinary JSON be read as RDF; a W3C Recommendation since January 2014.

Compact text syntax for writing RDF triples by hand, widely used long before it became a W3C Recommendation in February 2014.

Open encoding, published by Google in October 2014, that turns coordinates into short codes such as 8FVC9G8F+6X; a free alternative to proprietary address grids.

Small OWL ontology for smart appliances and the Internet of Things, begun in 2014–15 in a study for the European Commission and continued by ETSI, with extensions for energy, buildings, agriculture and other sectors.

W3C Recommendation of 26 February 2015 setting rules for reading and writing Linked Data with ordinary HTTP operations, so resources and containers behave like a familiar REST API; Solid builds on it.

Query language for web APIs developed at Facebook from 2012 and published in 2015; it lets a client ask for exactly the JSON it needs, and became the common choice where Linked Data APIs were not adopted.

Community profile of the CIDOC CRM for describing artworks and other heritage objects, in JSON-LD and using the Getty Vocabularies, so that museum data is practical for developers; started in 2016.

The International Council on Archives' standard for archival description as a graph of records, agents, activities and their relations, with a conceptual model and an ontology (RiC-O). Work began in 2012, a draft followed in 2016, and version 1.0 was released in 2023–2024; it extends and replaces the ICA's four earlier description standards, including ISAD(G).

IFLA's 2017 conceptual model of bibliographic information, which consolidates FRBR and its companion models for authority and subject data.

W3C Recommendation (2017) for telling any URL that you have linked to it. The receiving site checks that the link really exists and can then show it, giving the Web's one-way links a way back.

W3C Recommendation of February 2017 for describing annotations, such as comments, highlights and tags, in JSON-LD so they can be shared between tools; IIIF uses it for annotations on images.

W3C vocabulary and JSON-LD format for describing social activities such as posts, likes and follows; a Recommendation since May 2017.

W3C language for validating RDF data against shapes, closer to schema checking than to OWL's open-world reasoning; a Recommendation since July 2017.

Joint W3C and OGC guidance, first published in 2017, for putting spatial data on the Web rather than in downloads; its first practice is to give places and spatial things persistent HTTP URIs.

Joint W3C and OGC ontology, a Recommendation of 19 October 2017, for describing sensors, actuators, observations and samples, with the lightweight SOSA core for simple use; a 2023 edition has been in draft since September 2025.

W3C Recommendation (2018) for decentralised social networking, used by Mastodon and the wider Fediverse. Its messages are Activity Streams 2.0 in JSON-LD, so the Fediverse runs on Linked Data.

Policy language for stating permissions, prohibitions and duties over content and data; the W3C's ODRL Information Model 2.2 became a Recommendation in February 2018.

GS1 standard for writing product identifiers such as GTINs as web URIs, so that a barcode or QR code on a pack resolves to information about the product; ratified in August 2018. It puts Linked Data identifiers on everyday goods.

W3C data model for tamper-evident digital credentials, such as diplomas or licences, that can be checked cryptographically. Version 1.0 became a Recommendation in November 2019, and version 2.0, with its Data Integrity and cryptosuite specifications, in May 2025.

W3C Recommendation of 9 April 2020 that describes a device's properties, actions and events in JSON that can be read as JSON-LD, so applications can work across Internet of Things platforms.

C2PA's format for cryptographically signed provenance records ('manifests') attached to images, video and other media, recording who made them and how they were edited; version 1.0 appeared in December 2021.

Protocol behind Bluesky. Every account is a DID with a domain name as its handle. Its schema language, Lexicon, deliberately avoids RDF and JSON-LD in favour of strict, versioned schemas.

W3C specification for identifiers that their subject can create and control without a central registry. DID 1.0 became a Recommendation in July 2022; DID 1.1 reached Candidate Recommendation in March 2026.

Vocabulary for describing how personal and other data is processed, why and on what legal basis, made by a W3C community group from 2018; version 1 was published in December 2022. It is designed to support legal requirements such as the GDPR.

ISO standard query language for property graphs, published in April 2024 and partly derived from Neo4j's Cypher; the property-graph counterpart to SPARQL.

W3C algorithm that gives any RDF dataset a single canonical form, so the same graph always hashes and signs the same way regardless of blank-node labels or order; a Recommendation since May 2024.

Open protocol introduced by Anthropic in November 2024 for connecting AI models to external tools and data sources; donated to the Linux Foundation's Agentic AI Foundation in December 2025.

Revision of SHACL by the W3C Data Shapes Working Group, in drafts since March 2025, adding modules for node expressions, rules, user interfaces, profiling and, from September 2026, inference rules that derive new triples from existing ones.

W3C specification for attaching cryptographic proofs to data such as verifiable credentials, using pluggable cryptosuites; a Recommendation since May 2025.

W3C specification, a first public draft since March 2026, for giving applications secure, permissioned access to data people store elsewhere, in an interoperable way.

Revision of RDF that adds triple terms, so a statement can itself be the object of another statement; it grew out of RDF-star. Its core spec became a W3C Candidate Recommendation in April 2026.

Binary serialization of JSON-LD in CBOR that uses JSON-LD contexts to compress terms and values, reporting results more than 60% better than general-purpose compression; meant for storage- and bandwidth-constrained settings. A W3C JSON-LD Working Group First Public Working Draft since 28 April 2026.

The electricity industry's shared model of power networks, adopted by the IEC, so grid operators' software can exchange data about lines, equipment and measurements; its profiles are exchanged as RDF, including between Europe's transmission operators.

Identifier for each individual physical object, not just each product type, designed to be carried on RFID tags; it came out of MIT's Auto-ID Center and is maintained by GS1.

Europeana's RDF model for cultural heritage metadata. It reuses Dublin Core, SKOS and OAI-ORE so that each provider's own description is kept alongside Europeana's additions; it replaced the flat Europeana Semantic Elements.

GS1's terms for products, organisations and locations published as Linked Data and designed as an extension to schema.org.

Data model for linking data by content hash rather than by location; the data layer under IPFS, and the basis of AT Protocol's repository format.

AT Protocol's schema language for records, API endpoints and event streams. Each schema is a JSON file named by a reverse-DNS identifier (NSID); it resembles JSON Schema and OpenAPI rather than RDF vocabularies.

Open Data Commons' copyleft licence for databases: anyone may share, adapt and use a database as long as adapted versions stay open. OpenStreetMap moved to it in 2012.