(This section is non-normative.)

There are several document schema definition languages in common use today that can be used to specify one or more validation processes performed against Extensible Markup Language (XML) documents. Some schema languages provide their own syntax for associating schemas with documents (DTD, W3C XML Schema) and some languages (RELAX NG, Schematron) do not provide schema association mechanisms at all. The purpose of this specification is to define a common, schema-agnostic syntax for associating schema documents written in any schema definition language with a given XML document.

This specification defines the syntax and processing expectations for an xml-model processing instruction. Such processing instructions associate one or more schemas with the XML document in which they are present. The associated schemas may be written in any schema definition language. Applications can use the associated schemas for any purpose including those such as document validation, content completion in interactive editors, or creating models for data binding. Presence of an xml-model processing instruction is not in itself an instruction to any processor to validate the document, nor is it a statement that the document is not to be processed without validation. It is a declarative statement of a relationship between the document and one or more external schemas. This specification does not prescribe what, if anything, an application does with an xml-model processing instruction. The presence of an xml-model processing instruction referencing a DTD does not affect the validity of the document which contains it.PBG2011-04-13 The preceding two sentences attempt to clarify how an xml-model PI referencing a DTD might affect application processing of the document. This issue was raised in comment JP-0001 during the ISO review.

It should be noted that this specification is not meant as a replacement for other technologies that provide more general and indirect schema association features like NVDL and XProc. This specification is complementary technology which can be used when it is necessary to store ad-hoc schema associations directly inside XML document.

All diagrams, examples, and notes in this specification are non-normative, as are all sections explicitly marked non-normative. Everything else in this specification is normative.

The key words must, must not, should, should not and may in the normative parts of this document are to be interpreted as described in RFC 2119. These words do not appear in all uppercase letters in this specification .

This specification is defined with reference to the vocabulary for XML provided by the XML Information Set as well as the rules for parsing pseudo-attributes from a string as defined in the Associating Style Sheets with XML documents Recommendation .

The productions in this specification use the same notation as used in the XML Recommendation. Tokens in the grammar and terms used in this specification that are not defined in this specification are defined in the XML Recommendation or the Associating Style Sheets Recommendation.

A processing instruction information item is said to be a potential xml-model processing instruction if it has the [target] property xml-model and it is in the [children] property of a document information item and appears before the element information item of the document information item's [children] property.

For such potential xml-model processing instructions, xml-model processors must report to the application the parsing result of invoking the rules for parsing pseudo-attributes from a string, using the processing instruction information item's [content] property as the string.

A potential xml-model processing instruction is said to be an xml-model processing instruction if the parsing result is not an error when invoking the rules for parsing pseudo-attributes from a string, using the processing instruction information item's [content] property as the string.

Documents must not contain processing instruction information items with the [target] property xml-model that are not xml-model processing instructions.PBG2011-04-12The rewording in the previous sentence addresses a comment from Tony Graham to xml-editors.

An xml-model processor must process all xml-model processing instructions properly and must pass on to the application the full parsing result for each xml-model processing instruction.

An xml-model processing instruction will match the following production:

Documents may specify the following pseudo-attributes on xml-model processing instructions, unless otherwise stated:

To allow for extensibility, documents may specify other pseudo-attributes on xml-model processing instructions.

This specification provides a way to associate multiple schemas with a given XML document. Furthermore, there exist other ways certain schemas can be associated with a given XML document. Regardless of the association method, this specification does not prescribe the processing order when multiple schemas are associated with a given XML document.

In particular, this specification does not define the interaction of xml-model processing instructions with xsi:schemaLocation and xsi:noNamespaceSchemaLocation attributes which provide hints for locating schema in W3C XML Schema. Applications supporting both xml-model processing instructions and xsi:schemaLocation/xsi:noNamespaceSchemaLocation attributes may provide means for specifying which information takes precedence.