Goal

The Abstract of the current working draft says:

This specification defines the XML Pointer Language (XPointer), the language to be used as a fragment identifier for any URI-reference that locates a resource of Internet media type text/xml or application/xml.

And RFC 2396 states:

The semantics of a fragment identifier is a property of the data resulting from a retrieval action, regardless of the type of URI used in the reference. Therefore, the format and interpretation of fragment identifiers is dependent on the media type [RFC2046] of the retrieval result.

The principle

XPointer

• Is based on XPath, an expression language to query XML trees
• Extends XPath types and add specific operations
• Returns a set of nodes, points or ranges within the document
• Should be able to express any user selection on the document

Architecture

XPath quick description

XPath describes the syntax and associated processing to select groups of nodes and attributes within an XML document tree.

• Is an expression language for XML trees
• Uses selection axis to build the search semantic
• Powerful, complex searches can be build
• Can be extended in a given context (variables, functions).
• Has a generic syntax and an abbreviated syntax for most used operations

XPath Context and Axis

XPath examples

• child::para selects the para element children
• child::* selects the element children
• /descendant::para all the para elements of the document
• /doc/chapter[5]/section[2] the second section of the fifth chapter of the doc elements under the document root
• /descendant::p[position()=last()] the last p in document order
• //note[@type="warning"] all note elements that has a type attribute with value warning

XPointer

XPointer

• Allow to address IDs with a simple HTML compatible syntax

  #Intro

• Reuse XPath and extends it
• Uses an extensible construct to embed the expression

  #xpointer(id(intro))

• Should allow to express any reasonable "user selection"
• Can return multiple non contiguous parts of the document

XPointer extensions to XPath

• XPath Node concept is extended to nodes, points and ranges and associated cast functions
• A string-range() function to address within (nodes) content
• The functions here() and origin() to refer to the XPointer expression location and hypertext traversal
• The functions start-point() and end-point() to address beginning and ending of locations.
• The unique() function to detect absence or multiple return values.
• Initialization of the XPath context
• Restriction: no variable used in XPointer though XPath defines them

XPointer Examples

• #xpointer(//*[local-name()='y' and namespace-uri()='http://www.foo.com/bar']) locates all the y elements from the namespace whose URI is http://www.foo.com/bar
• xpointer(id("sec2.1")//p[2] to id("sec2.2")//p[last()]) selects the range from the second p child of sec2.1 to the end of the last p in sec2.2
• string-range(//title,"XPointer") returns the set of ranges containing "XPointer" strings when they occur within title elements

A couple of issues

• How do we detect IDs when not validating ?
• Other fragment identifier syntaxes for XML based format ?

Current Status

XPath

• Is a W3C Recommendation
• No foreseen evolution
• Is currently used by XPointer and XSLT
• May be reused by XML Query and other future specifications

XPointer

• Is in Last Call status until Dec 27 1999
• Is considered solid and expected to go to REC within a few months
• Will be used by XLink (and all URI-References to XML data)

How you can help

• Review the Last Call Working Draft
• Implement it !
• Provide feedback if there is problems (also check the archive).
• Use it !