DTDs

Document Type Definitions (DTDs) are a schema language used to define the structure and legal content of SGML and XML documents. A DTD specifies which elements are permitted, the order and nesting of child elements, and the attributes that elements may have. It can also declare entities and notations used within the document.

DTDs originated with SGML, the ancestor of XML, and were carried into XML as a lightweight schema mechanism. In XML, a DTD can be declared as an internal subset within the DOCTYPE declaration or as an external subset stored in a separate file. Key declarations include <!ELEMENT> to define element content models, <!ATTLIST> to declare attributes (types and defaults), <!ENTITY> to declare general and parameter entities, and <!NOTATION> for identifying non-XML data.

An external DTD subset is referenced from the DOCTYPE via PUBLIC or SYSTEM identifiers, for example a DOCTYPE that points to an external DTD file. An internal subset allows defining elements and attributes directly inside the DOCTYPE. DTDs can also use parameter entities to reuse DTD fragments and general entities for text reuse within the document.

Validation and limitations: If a parser is configured to validate, a document can be checked against its DTD to ensure it is valid in addition to being well-formed. DTDs provide basic structural and attribute constraints but have a limited datatype system (for example CDATA, ID, IDREF, ENUMERATION) and lack the expressive power of XML Schema or RELAX NG. They also have limited native support for namespaces.

DTDs remain in use in legacy SGML/XML contexts and for simple validation tasks, though many modern XML projects prefer schema languages with richer data typing and flexibility.