Schema.org

Schema.org organizes all entities into a multi-level hierarchy rooted at Thing. More specific types inherit properties from their ancestors.

• Thing → CreativeWork → Article → NewsArticle
• Each level adds domain-specific properties
• Over 800 core types defined in the vocabulary
• Enables both broad and granular entity description

A NewsArticle inherits author from CreativeWork and name from Thing, while adding its own dateline property.

Entities are described through properties rather than nested structures. Each property has a defined domain (types it applies to) and range (expected value type).

• Properties can apply to multiple types
• Expected types include Text, Number, Date, or other Schema.org types
• Properties like author can accept both Person and Organization
• Supports multiple values via array syntax

This design allows flexible, partial descriptions without requiring complete entity graphs.

Schema.org uses enumerations to constrain property values to predefined sets, ensuring semantic consistency across implementations.

• DayOfWeek: Monday, Tuesday, Wednesday, etc.
• ItemAvailability: InStock, OutOfStock, PreOrder
• PaymentMethod: Cash, CreditCard, Cryptocurrency
• EventStatusType: EventScheduled, EventPostponed, EventCancelled

Enumerations prevent free-text ambiguity and enable reliable machine interpretation of constrained value spaces.

Schema.org vocabulary is syntax-agnostic. It can be expressed in three formats recognized by major search engines:

• JSON-LD: A JavaScript object injected into a <script> tag, kept separate from HTML markup
• Microdata: HTML attributes (itemscope, itemprop) woven directly into existing tags
• RDFa: Attribute-based syntax extending HTML5 for linked data embedding

JSON-LD is Google's recommended format due to its clean separation of data and presentation.

The vocabulary is partitioned into hosted extensions that cover specific verticals while remaining part of the core namespace.

• bib.schema.org: Bibliographic and citation data
• auto.schema.org: Vehicle listings and specifications
• health-lifesci.schema.org: Medical and life science entities
• pending.schema.org: Proposed types under community review

Extensions allow domain-specific evolution without destabilizing the core vocabulary.

Schema.org is stewarded by a consortium of search engines—Google, Microsoft, Yahoo, and Yandex—with public participation through:

• Open GitHub repository for proposals and issue tracking
• Mailing lists for discussion of new types and properties
• Regular releases (currently v27.0+) incorporating community feedback
• The pending namespace as a staging area for experimental terms

This governance model balances industry alignment with open evolution.

The automated process of checking a data instance against a defined schema to ensure it conforms to required structures, data types, and constraints. For Schema.org, this involves validating JSON-LD against the vocabulary's expected ranges and domains.

• SHACL: Shapes Constraint Language for validating RDF graphs.
• Google's Rich Results Test: A practical validator that parses Schema.org and checks for required/recommended properties.
• Domain/Range Checks: Ensures a price property points to a Number or MonetaryAmount, not a Person.

A formal, explicit specification of a shared conceptualization. Schema.org is an ontology that defines the types, properties, and interrelationships of entities on the web.

• Classes: Abstract categories like Thing > Organization > LocalBusiness.
• Properties: Relationships like founder, address, or aggregateRating.
• Inference: Because LocalBusiness is a subclass of Organization, it inherits all properties of Organization.

A method of publishing structured data so that it can be interlinked and become more useful through semantic queries. Schema.org enables Linked Data by using URIs as identifiers.

• Entity Reconciliation: Using sameAs to link a local business entity to its Wikidata or Wikipedia entry.
• De-referencable URIs: A @id like https://example.com/entity#org should ideally resolve to useful data.
• Graph Connectivity: Connecting a Product to its Manufacturer via a URI creates a machine-readable knowledge graph.

The process of modifying a data schema's structure while maintaining backward and forward compatibility. Schema.org evolves through a community process, adding new types and properties without breaking existing markup.

• Backward Compatibility: New properties are additive; old parsers ignore unknown properties.
• Deprecation: Superseded terms like cookTime (replaced by cookTime on HowToStep) are retained but discouraged.
• Versioning: Schema.org uses a rolling release model rather than strict semantic versioning.

A formal representation of content types, their attributes, and relationships. A content model maps business concepts to Schema.org types.

• Mapping: A 'Blog Post' content model maps to Article or BlogPosting.
• Constraints: Defining cardinality—a Product must have exactly one name but can have multiple offers.
• Enrichment: Adding domain-specific properties like hasPart to model complex technical documentation.