XML Pocket Reference, 3rd Edition

The xml:id Attribute

The xml:id attribute is a method for guaranteeing proper ID processing.

Example

<message xml:id="i-35867"
>This is a message</message>

Description

Under XML 1.0, an ID is a unique identifier to aid in processing. You can annotate an element uniquely with an attribute of type ID, as in id="i-35867" (IDs can’t start with a number), which often assumes an associated DTD containing the attribute-list declaration <!ATTLIST message id ID #REQUIRED>. Likewise, XML Schema provides a mechanism for identifying markup as having type ID with the type="xs:ID" attribute, as with <xs:attribute name="id" type="xs:ID">. The problem is that non-validating yet conformant XML processors are not required to refer to or process an external subset DTD (one that exists outside of the XML document), and a correct schema may not be available, so processing IDs can be troublesome. Implementation of the xml:id attribute is an attempt to guarantee that ID processing will be consistent and reliable, whether the XML processor being used is validating or not. The xml:id mechanism is currently a W3C candidate recommendation, and it’s a development worth tracking. The upcoming specs for XPath 2.0—and hence XQuery and XSLT 2.0—also support xml:id.

See also

The xml:id spec: http://www.w3.org/TR/xml-id/