Document schemas provide machine-understandable models that can be used to validate the structure and contents of electronically marked-up documents. The Document Type Definition (DTDs) language defined within ISO 8879 provides facilities for:

• defining the names used to identify document elements

• identifying where document elements may appear in the document structure (model)

• identifying which elements were optional or repeatable (without limiting repeatability)

• identifying which markup tags are optional when they can be inferred through the model

• assigning properties (attributes) to document elements that can be used to control their processing, or can contain information that needs to be processed in conjunction with element contents

• defining default values for attributes

• defining and naming repeatable segments of text (entities)

• identifying non-standard characters using user-assigned names or character numbers

• linking together different document structures defined in parallel sets of markup.

Document structures are defined in ISO 8879 in terms of "trees" of nested elements, though the standard also allows data sets to be defined as "graphs" of elements connected by means of unique identifiers and references to existing identifiers.

DTDs are not defined using the same components as marked-up document instances, being defined using a sparser notation. In SGML this notation can be preceded by a declaration of permitted character sets, which characters are assigned as control functions or otherwise ignored (shunned), which characters can be used within names, or to identify the boundaries of markup, which strings can be used to automatically identify markup points, and which optional functions are to be used within the DTDThe character definition rules predate the development of the ISO/IEC 10646 Universal Multiple-Octet Coded Character Set and are to some extent made redundant by this standard..

The W3C Extensible Markup Language (XML) uses an application profile of ISO 8879, known as WebSGML, togetheer with the ISO/IEC 10646 Universal Multiple-Octet Coded Character Set, to produce a simple-to-implement, streamable, application of ISO 8879 for use over the Internet and similar networks. XML document models can be defined using DTDs that conform to a well defined set of restrictions on the options of ISO 8879 so that vaildation can be performed within streamed networks.

Various organizations have developed techniques to manage the structure of documents using XML markup. Some of these further subset the facilities provided in ISO 8879, but many also provide functions over and above those allowed within SGML and XML DTDs, including:

• control of the minimum and maximum number of times an element can occur at a particular point in the document structure (e.g. to control cardinality)

• restriction of the contents of particular elements or attributes to those that conform to particular datatypes, patterns or internally defined lists of permitted elements

• provision for distinguishing the namespace of element and attribute names so that schema fragments can be used within other schemas without fear of name clashes

• identification of elements based on the path needed to reach them within the document structure

• validation of document structures by checking that elements conforming to particular paths exist

• provision of mechanisms for creating abstract stereotypes (similar to SGML architectural forms) that can be used to identify related classes of elements.

The following additional functionality has been identified as being required by users:

1. The ability to control the character set permitted within the contents of a particular type of element or attribute, or within specified sets of elements within the document model.

2. The ability to restrict the range of entries conforming to a particular use of a datatype within a specific element or attribute.

3. The ability to restrict element or attribute contents to values specified in either internally defined or externally defined lists of permitted values.

4. The ability to restrict the set of permitted values in one element or attribute based on the contents of another element or attribute (e.g. not Sex=Male and Condition=Pregnant).

5. The ability to generate compact forms of schemas that are easily readable by humans, and to use such compact representations to generate schemas or DTDs that can be used to validate documents.

6. The ability to visualize schemas using navigable diagrammatic representations.

The various parts of the DSDL standard are designed to be choreographed to satisfy a declarative expression of the validation requirements, without the need to use processes in a predefined sequence. Some parts of the standard will, however, need to be applied before others. For example, validation of the contents of a specific element will require prior identification of element boundaries and nesting, while the validation of the relationship between elements may require that the document structure be validated first so that the paths specified can be checked accordingly.

Consider the example in Two different orders of application of technology where two different results can be determined from two different applications of technology to the validation process: validating after or before processing XInclude. The order of these two steps may be critical in the correct processing of the information in the instance.

In a more complex example, consider A multi-step validation process where two different technologies must be applied to separate portions of the one document. In this case, one part of the input document must be validated by a W3C XML Schema while the other part of the input document must be validated by a RELAX-NG schema. The validation result, orchestrated by the DSDL framework, would express the consolidated validation of all steps.

The non-hierarchical links between information items in a structured resource can be identified by addressing the items within the document tree and then expressing the relationship between them. The addressing mechanism includes hierarchy-based paths of steps along the tree's branches to the information item being addressed.

Paths can be based on:

• a method of identifying information items dependent on:

  • the ancestry of the information item

  • the use of keys (e.g. references to unique identifier values)

  • an extensible basis for supporting mechanisms not currently available

• a method of describing the role of relationships that are not hierarchical.

A number of Parts within this International Standard utilize the concept of a path to address components of document instances.

At present users have no control over which set of characters can appear in a particular element or attribute value. For example, an element could have an xml:lang attribute indicating it is in English but contain Chinese or Sanskrit characters. This Part provides a mechanism for checking that the contents of an element or attribute are taken from a formally defined subset of the ISO/IEC 10646 Universal Character Set (UCS) that is the basis for XML encoded documents.

This Part provides a syntax for:

• defining named subsets of the ISO/IEC 10646 character set

• identifying which named character set shall be used to validate the content of a specific element or attribute

Declarative document manipulation allows default values to be assigned to specific parts of a data stream. This includes mechanisms for defining standard sequences of data that can be incorporated into document instances by reference to an identifying name, the provision of default content for elements and attributes for which no value is provided and the matching of local element and attribute names to those used in a specific schema.

This Part defines a syntax for describing simple modifications to be made to the information set of a DSDL document instance, without requiring the full power of a general-purpose transformation language such as XSLT.

This Part provides a syntax for:

• using XPath to address elements and attributes to be modified

• assigning a default value to the contents of a specific type of element or attribute

• defining named entities of predefined data elements that can be used to include template data within a document instance

• renaming elements and attributes in specific locations within the document model, including the assignment of element or attribute names to different namespaces

• the definition of replacement contents for specific elements or attributes

• removing elements or attributes from specific locations within the document model.

This Part specifies how the ISO 8879/XML Document Type Definition (DTD) syntax can be extended to validate documents that make full use of XML Namespaces and Part 5 Datatypes. This will ensure that the investment that individuals and organizations have made in DTD development and deployment over many years can be preserved. It will also help those converting between DTDs and other forms of schema languages.

The specification does not require documents using the schema language to violate XML's well-formedness or validity checks. It simply identifies names whose role can be considered to be that of specifying addtional validation rules to be applied to specific elements or attributes.