Semantic Web
Wikidata
Free, multilingual, machine-readable knowledge base launched by the Wikimedia Foundation in October 2012. Organizes facts as items (QIDs) and properties (P-prefixed), exposed via a SPARQL endpoint, released under CC0, and consumed by Wikipedia infoboxes, search engines, and virtual assistants.
Semantic Web
Tim Berners-Lee's proposed extension of the web in which data carries machine-readable meaning via RDF triples, OWL ontologies, and SPARQL queries. Standardized by W3C and partially realized in linked-data and scientific-data ecosystems, but largely supplanted on the open web by schema.org and LLM-based extraction.
SPARQL
SPARQL is the W3C-standardized query language for RDF graphs, recommended as SPARQL 1.0 in 2008 and SPARQL 1.1 in 2013. Queries match patterns of triples and return tabular or graph results, and the language is also a protocol for HTTP-accessible query endpoints such as the Wikidata Query Service and DBpedia.
JSON-LD
JSON-LD (JavaScript Object Notation for Linked Data) is a JSON-based format for expressing linked data, standardized as a W3C Recommendation in January 2014. It uses reserved keywords like @context, @id and @type to let plain JSON documents be interpreted as RDF triples, and is the preferred encoding for Schema.org markup in HTML and the wire format for several W3C standards including Activity Streams and Verifiable Credentials.
Resource Description Framework (RDF)
RDF is a W3C standard data model, first recommended in 1999 and revised as RDF 1.1 in 2014, for describing resources on the web as subject-predicate-object triples. Triples form a directed labelled graph that can be merged across datasets, which is the foundation for Linked Data and the Semantic Web.
Google Knowledge Graph
The Google Knowledge Graph is a structured database of real-world entities and their relationships, launched in May 2012 with about 500 million entities and 3.5 billion facts. It is built from sources like Wikipedia, Wikidata, the former Freebase project and Schema.org markup crawled from the web, and powers knowledge panels, direct answers, carousels and voice-assistant responses.
Freebase (knowledge base)
Open collaborative knowledge base launched in 2007 by Metaweb, acquired by Google in 2010, and shut down in 2016 after its data was partially migrated to Wikidata. A direct conceptual ancestor of modern public knowledge graphs.
Schema.org Structured Data
Schema.org is a shared vocabulary for embedding machine-readable metadata in HTML, launched in June 2011 by Google, Microsoft and Yahoo! (with Yandex joining later). It supports Microdata, RDFa and JSON-LD encodings — with JSON-LD now preferred — and defines core types like Organization, Person, Article, Product, Event, Recipe and Review. Search engines use the markup to render rich results and to ingest entities into knowledge graphs; the vocabulary is now governed by an open community process.
Turtle (RDF Syntax)
Turtle (Terse RDF Triple Language) is a compact, human-readable syntax for RDF graphs, standardized as a W3C recommendation in February 2014. It is a superset of N-Triples and a subset of Notation3, uses the .ttl file extension and text/turtle media type, and shares syntax with the SPARQL query language.
Linked Data and RDF
Linked Data is a set of practices for publishing structured, machine-readable data on the web, articulated by Tim Berners-Lee in a 2006 design issues note as four short principles: name things with URIs, make those URIs dereference over HTTP, return useful RDF when they do, and link out to other URIs. The underlying data model is RDF, a W3C standard since 1999 that represents information as subject-predicate-object triples. Realized at scale in DBpedia, Wikidata, GeoNames, and the wider Linked Open Data cloud, the stack has thrived in libraries and cultural heritage but seen limited adoption elsewhere due to its steep learning curve.
DBpedia
DBpedia is a community project, started in 2007 at the Free University of Berlin and Leipzig University, that extracts structured data from Wikipedia infoboxes and categories and republishes it as RDF. It is one of the central hubs of the Linked Open Data cloud and exposes a public SPARQL endpoint.
OWL (Web Ontology Language)
OWL is the W3C-standardized Web Ontology Language for describing classes, properties, and relationships with formal logical semantics. It was first recommended in 2004 and revised as OWL 2 in 2009. It extends RDF Schema with constructs such as cardinality restrictions, disjointness, and equivalence, enabling automated reasoning.
RDFa
RDFa (Resource Description Framework in Attributes) is a W3C specification for expressing RDF triples through attributes embedded in HTML, XHTML or XML. RDFa 1.0 became a Recommendation in 2008 and RDFa 1.1 — including the simpler five-attribute RDFa Lite — in 2012. It is one of the three encodings supported by Schema.org, though JSON-LD has largely surpassed it in web-scale adoption.
Microdata (HTML)
Microdata is a WHATWG specification for embedding structured data directly inside HTML using attributes like itemscope, itemtype and itemprop. It was introduced by Ian Hickson in 2009 as part of HTML5 after RDFa was not adopted into the core HTML5 spec, and was the original encoding endorsed by Schema.org. Search engines still consume Microdata for rich results, though JSON-LD has overtaken it as the preferred encoding.
Linked Open Data Cloud
The interconnected ecosystem of RDF datasets published according to Berners-Lee's linked-data principles, visualized as the well-known LOD cloud diagram. Strongest in cultural-heritage catalogs, life-science databases, and Wikidata, where institutional payoffs justify the cost of ontology engineering.
Metacrap (Doctorow Essay)
Cory Doctorow's 2001 essay arguing that voluntary metadata schemes — including the Semantic Web — are undermined by seven recurring problems: people lie, are lazy, are bad self-observers, schemas embed bias, metrics distort behavior, and there is no single canonical description.
Rich Result (Search)
A rich result is an enhanced search-engine listing rendered from structured data on a page — typically Schema.org markup in JSON-LD, Microdata or RDFa. Common types include recipe cards, product listings, review stars, FAQ accordions, breadcrumbs, events and video thumbnails. Eligibility depends on required properties defined per type, and the supported catalogue changes over time as engines add or retire features.
Why the Semantic Web Underperformed Expectations
Tim Berners-Lee's 2001 vision of a machine-readable web powered by RDF, OWL, and SPARQL never produced the promised agentic future. High entry costs, weak publisher incentives, schema.org's lighter SEO-driven alternative, and LLMs that extract structure from prose directly all undercut it. The formal-semantics vision survived mainly in cultural-heritage linked data, life-science integration, and Wikidata, where institutional payoffs justify the ontology work.