A processing instruction (PI) is an SGML and XML node type, which may occur anywhere in a document, intended to carry instructions to the application.^[1][2]

Processing instructions are exposed in the Document Object Model as Node.PROCESSING_INSTRUCTION_NODE, and they can be used in XPath and XQuery with the 'processing-instruction()' command.

An SGML processing instruction is enclosed within <? and >.^[3]

An XML processing instruction is enclosed within <? and ?>, and contains a target and optionally some content, which is the node value, that cannot contain the sequence ?>.^[4]

<?PITarget PIContent?>

The XML Declaration at the beginning of an XML document (shown below) is another example of a processing instruction,^[5] however it may not technically be considered one.^[6]

<?xml version="1.0" encoding="UTF-8" ?>

The most common use of a processing instruction is to request the XML document be rendered using a stylesheet using the 'xml-stylesheet' target, which was standardized in 1999.^[7] It can be used for both XSLT and CSS stylesheets.

<?xml-stylesheet type="text/xsl" href="style.xsl"?>

<?xml-stylesheet type="text/css" href="style.css"?>

The DocBook XSLT stylesheets understand a number of processing instructions to override the default behaviour.^[8]

A draft specification for Robots exclusion standard rules inside XML documents uses processing instructions.^[9]