aboutsummaryrefslogtreecommitdiff
path: root/misc
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--misc/w3c-xml-1.1.html1598
1 files changed, 1598 insertions, 0 deletions
diff --git a/misc/w3c-xml-1.1.html b/misc/w3c-xml-1.1.html
new file mode 100644
index 0000000..6a9211a
--- /dev/null
+++ b/misc/w3c-xml-1.1.html
@@ -0,0 +1,1598 @@
1<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html lang="EN" xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html;charset=UTF-8" /><title>Extensible Markup Language (XML) 1.1</title><style type="text/css">
2code { font-family: monospace; }
3
4div.constraint,
5div.issue,
6div.note,
7div.notice { margin-left: 2em; }
8
9li p { margin-top: 0.3em;
10 margin-bottom: 0.3em; }
11
12div.exampleInner pre { margin-left: 1em;
13 margin-top: 0em; margin-bottom: 0em}
14div.exampleOuter {border: 4px double gray;
15 margin: 0em; padding: 0em}
16div.exampleInner { background-color: #d5dee3;
17 border-top-width: 4px;
18 border-top-style: double;
19 border-top-color: #d3d3d3;
20 border-bottom-width: 4px;
21 border-bottom-style: double;
22 border-bottom-color: #d3d3d3;
23 padding: 4px; margin: 0em }
24div.exampleWrapper { margin: 4px }
25div.exampleHeader { font-weight: bold;
26 margin: 4px}
27
28em.rfc2119 { text-transform: lowercase;
29 font-variant: small-caps;
30 font-style: normal; }
31</style><link rel="stylesheet" type="text/css" href="http://www.w3.org/StyleSheets/TR/W3C-REC.css" /></head><body><div class="head"><p><a href="http://www.w3.org/"><img src="http://www.w3.org/Icons/w3c_home" alt="W3C" height="48" width="72" /></a></p>
32<h1><a name="title" id="title" />Extensible Markup Language (XML) 1.1</h1>
33<h2><a name="w3c-doctype" id="w3c-doctype" />W3C Recommendation 04
34 February 2004, edited in place 15 April 2004</h2><dl><dt>This version:</dt><dd><a href="http://www.w3.org/TR/2004/REC-xml11-20040204/">http://www.w3.org/TR/2004/REC-xml11-20040204/</a></dd><dt>Latest version:</dt><dd><a href="http://www.w3.org/TR/xml11">http://www.w3.org/TR/xml11</a></dd><dt>Previous version:</dt><dd><a href="http://www.w3.org/TR/2003/PR-xml11-20031105/">http://www.w3.org/TR/2003/PR-xml11-20031105/</a></dd><dt>Editors:</dt><dd>Tim Bray, Textuality and Netscape <a href="mailto:tbray@textuality.com">&lt;tbray@textuality.com&gt;</a></dd><dd>Jean Paoli, Microsoft <a href="mailto:jeanpa@microsoft.com">&lt;jeanpa@microsoft.com&gt;</a></dd><dd>C. M. Sperberg-McQueen, W3C <a href="mailto:cmsmcq@w3.org">&lt;cmsmcq@w3.org&gt;</a></dd><dd>Eve Maler, Sun Microsystems, Inc. <a href="mailto:elm@east.sun.com">&lt;eve.maler@east.sun.com&gt;</a></dd><dd>François Yergeau <a href="mailto:fyergeau@alis.com">&lt;fyergeau@alis.com&gt;</a></dd><dd>John Cowan <a href="mailto:cowan@ccil.org">&lt;cowan@ccil.org&gt;</a></dd></dl><p>Please refer to the <a href="http://www.w3.org/XML/xml-V11-1e-errata"><strong>errata</strong></a> for this document, which may include some normative corrections.</p><p>This document is also available in these non-normative formats: <a href="REC-xml11-20040204.xml">XML</a> and <a href="REC-xml11-20040204-review.html">XHTML with color-coded revision indicators</a>.</p><p>See also <a href="http://www.w3.org/2003/03/Translations/byTechnology?technology=xml11"><strong>translations</strong></a>.</p><p class="copyright"><a href="http://www.w3.org/Consortium/Legal/ipr-notice#Copyright">Copyright</a> © 2004 <a href="http://www.w3.org/"><acronym title="World Wide Web Consortium">W3C</acronym></a><sup>®</sup> (<a href="http://www.csail.mit.edu/"><acronym title="Massachusetts Institute of Technology">MIT</acronym></a>, <a href="http://www.ercim.org/"><acronym title="European Research Consortium for Informatics and Mathematics">ERCIM</acronym></a>, <a href="http://www.keio.ac.jp/">Keio</a>), All Rights Reserved. W3C <a href="http://www.w3.org/Consortium/Legal/ipr-notice#Legal_Disclaimer">liability</a>, <a href="http://www.w3.org/Consortium/Legal/ipr-notice#W3C_Trademarks">trademark</a>, <a href="http://www.w3.org/Consortium/Legal/copyright-documents">document use</a> and <a href="http://www.w3.org/Consortium/Legal/copyright-software">software licensing</a> rules apply.</p></div><hr /><div> <h2><a name="abstract" id="abstract" />Abstract</h2><p>The Extensible Markup Language (XML) is a subset of SGML that is completely
35described in this document. Its goal is to enable generic SGML to be served,
36received, and processed on the Web in the way that is now possible with HTML.
37XML has been designed for ease of implementation and for interoperability
38with both SGML and HTML.</p></div><div> <h2><a name="status" id="status" />Status of this Document</h2><p><em>This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the <a href="http://www.w3.org/TR/">W3C technical reports index</a> at http://www.w3.org/TR/.</em></p><p>This document is a <a href="http://www.w3.org/2003/06/Process-20030618/tr.html#RecsW3C">Recommendation</a> of the W3C.
39It has been reviewed by W3C Members and other interested parties, and has
40been endorsed by the Director as a W3C Recommendation. It is a stable document and may be used as reference material or cited as a normative reference from another document. W3C's role in making the
41Recommendation is to draw attention to the specification and to promote its widespread deployment.
42This enhances the functionality and interoperability of the Web.</p><p>This document specifies a syntax created by subsetting an existing, widely
43used international text processing standard (Standard Generalized Markup Language,
44ISO 8879:1986(E) as amended and corrected) for use on the World Wide Web.
45It is a product of the <a
46 href="http://www.w3.org/XML/Activity.html">W3C XML
47 Activity</a>.</p>
48
49<p>On 15 April 2004, this document was edited in place to add two
50missing spaces to <a
51href="http://www.w3.org/TR/2004/REC-xml11-20040204/Overview.html#NT-document">production
52[1]</a> in section 2.1</p>
53
54<p>The English version of this specification is the only normative version. However,
55for translations of this document, see <a href="http://www.w3.org/2003/03/Translations/byTechnology?technology=xml11">http://www.w3.org/2003/03/Translations/byTechnology?technology=xml11</a>.
56</p><p>Documentation of intellectual property possibly relevant to this recommendation
57may be found at the Working Group's public
58<a href="http://www.w3.org/2002/08/xmlcore-IPR-statements">IPR disclosure page</a>.</p><p>An implementation report for XML 1.1 is available at <a href="http://www.w3.org/XML/2002/09/xml11-implementation.html">http://www.w3.org/XML/2002/09/xml11-implementation.html</a>.</p><p>Please report errors in this document to <a href="mailto:xml-editor@w3.org">xml-editor@w3.org</a>; <a href="http://lists.w3.org/Archives/Public/xml-editor">archives</a> are available. The errata list for this edition is available
59at <a href="http://www.w3.org/XML/xml-V11-1e-errata">http://www.w3.org/XML/xml-V11-1e-errata</a>.</p><p>A <a href="http://www.w3.org/XML/Test/">Test Suite</a> is maintained to help assessing conformance to this specification.</p></div><div class="toc"> <h2><a name="contents" id="contents" />Table of Contents</h2><p class="toc">1 <a href="#sec-intro">Introduction</a><br />     1.1 <a href="#sec-origin-goals">Origin and Goals</a><br />     1.2 <a href="#sec-terminology">Terminology</a><br />     1.3 <a href="#sec-xml11">Rationale and list of changes for XML 1.1</a><br /> 2 <a href="#sec-documents">Documents</a><br />     2.1 <a href="#sec-well-formed">Well-Formed XML Documents</a><br />     2.2 <a href="#charsets">Characters</a><br />     2.3 <a href="#sec-common-syn">Common Syntactic Constructs</a><br />     2.4 <a href="#syntax">Character Data and Markup</a><br />     2.5 <a href="#sec-comments">Comments</a><br />     2.6 <a href="#sec-pi">Processing Instructions</a><br />     2.7 <a href="#sec-cdata-sect">CDATA Sections</a><br />     2.8 <a href="#sec-prolog-dtd">Prolog and Document Type Declaration</a><br />     2.9 <a href="#sec-rmd">Standalone Document Declaration</a><br />     2.10 <a href="#sec-white-space">White Space Handling</a><br />     2.11 <a href="#sec-line-ends">End-of-Line Handling</a><br />     2.12 <a href="#sec-lang-tag">Language Identification</a><br />     2.13 <a href="#sec-normalization-checking">Normalization Checking</a><br /> 3 <a href="#sec-logical-struct">Logical Structures</a><br />     3.1 <a href="#sec-starttags">Start-Tags, End-Tags, and Empty-Element Tags</a><br />     3.2 <a href="#elemdecls">Element Type Declarations</a><br />         3.2.1 <a href="#sec-element-content">Element Content</a><br />         3.2.2 <a href="#sec-mixed-content">Mixed Content</a><br />     3.3 <a href="#attdecls">Attribute-List Declarations</a><br />         3.3.1 <a href="#sec-attribute-types">Attribute Types</a><br />         3.3.2 <a href="#sec-attr-defaults">Attribute Defaults</a><br />         3.3.3 <a href="#AVNormalize">Attribute-Value Normalization</a><br />     3.4 <a href="#sec-condition-sect">Conditional Sections</a><br /> 4 <a href="#sec-physical-struct">Physical Structures</a><br />     4.1 <a href="#sec-references">Character and Entity References</a><br />     4.2 <a href="#sec-entity-decl">Entity Declarations</a><br />         4.2.1 <a href="#sec-internal-ent">Internal Entities</a><br />         4.2.2 <a href="#sec-external-ent">External Entities</a><br />     4.3 <a href="#TextEntities">Parsed Entities</a><br />         4.3.1 <a href="#sec-TextDecl">The Text Declaration</a><br />         4.3.2 <a href="#wf-entities">Well-Formed Parsed Entities</a><br />         4.3.3 <a href="#charencoding">Character Encoding in Entities</a><br />         4.3.4 <a href="#sec-version-info">Version Information in Entities</a><br />     4.4 <a href="#entproc">XML Processor Treatment of Entities and References</a><br />         4.4.1 <a href="#not-recognized">Not Recognized</a><br />         4.4.2 <a href="#included">Included</a><br />         4.4.3 <a href="#include-if-valid">Included If Validating</a><br />         4.4.4 <a href="#forbidden">Forbidden</a><br />         4.4.5 <a href="#inliteral">Included in Literal</a><br />         4.4.6 <a href="#notify">Notify</a><br />         4.4.7 <a href="#bypass">Bypassed</a><br />         4.4.8 <a href="#as-PE">Included as PE</a><br />         4.4.9 <a href="#error">Error</a><br />     4.5 <a href="#intern-replacement">Construction of Entity Replacement Text</a><br />     4.6 <a href="#sec-predefined-ent">Predefined Entities</a><br />     4.7 <a href="#Notations">Notation Declarations</a><br />     4.8 <a href="#sec-doc-entity">Document Entity</a><br /> 5 <a href="#sec-conformance">Conformance</a><br />     5.1 <a href="#proc-types">Validating and Non-Validating Processors</a><br />     5.2 <a href="#safe-behavior">Using XML Processors</a><br /> 6 <a href="#sec-notation">Notation</a><br /> </p> <h3><a name="appendices" id="appendices" />Appendices</h3><p class="toc">A <a href="#sec-bibliography">References</a><br />     A.1 <a href="#sec-existing-stds">Normative References</a><br />     A.2 <a href="#null">Other References</a><br /> B <a href="#sec-CharNorm">Definitions for Character Normalization</a><br /> C <a href="#sec-entexpand">Expansion of Entity and Character References</a> (Non-Normative)<br /> D <a href="#determinism">Deterministic Content Models</a> (Non-Normative)<br /> E <a href="#sec-guessing">Autodetection of Character Encodings</a> (Non-Normative)<br />     E.1 <a href="#sec-guessing-no-ext-info">Detection Without External Encoding Information</a><br />     E.2 <a href="#sec-guessing-with-ext-info">Priorities in the Presence of External Encoding Information</a><br /> F <a href="#sec-xml-wg">W3C XML Working Group</a> (Non-Normative)<br /> G <a href="#sec-core-wg">W3C XML Core Working Group</a> (Non-Normative)<br /> H <a href="#prod-notes">Production Notes</a> (Non-Normative)<br /> I <a href="#sec-suggested-names">Suggestions for XML Names</a> (Non-Normative)<br /> </p></div><hr /><div class="body"><div class="div1"> <h2><a name="sec-intro" id="sec-intro" />1 Introduction</h2><p>Extensible Markup Language, abbreviated XML, describes a class of data
60objects called <a title="XML Document" href="#dt-xml-doc">XML documents</a> and partially
61describes the behavior of computer programs which process them. XML is an
62application profile or restricted form of SGML, the Standard Generalized Markup
63Language <a href="#ISO8879">[ISO 8879]</a>. By construction, XML documents are conforming
64SGML documents.</p><p>XML documents are made up of storage units called <a title="Entity" href="#dt-entity">entities</a>,
65which contain either parsed or unparsed data. Parsed data is made up of <a title="Character" href="#dt-character">characters</a>, some of which form <a title="Character Data" href="#dt-chardata">character
66data</a>, and some of which form <a title="Markup" href="#dt-markup">markup</a>.
67Markup encodes a description of the document's storage layout and logical
68structure. XML provides a mechanism to impose constraints on the storage layout
69and logical structure.</p><p>[<a name="dt-xml-proc" id="dt-xml-proc" title="XML Processor">Definition</a>: A software module called
70an <b>XML processor</b> is used to read XML documents and provide access
71to their content and structure.] [<a name="dt-app" id="dt-app" title="Application">Definition</a>: It
72is assumed that an XML processor is doing its work on behalf of another module,
73called the <b>application</b>.] This specification describes
74the required behavior of an XML processor in terms of how it must read XML
75data and the information it must provide to the application.</p><div class="div2"> <h3><a name="sec-origin-goals" id="sec-origin-goals" />1.1 Origin and Goals</h3><p>XML was developed by an XML Working Group (originally known as the SGML
76Editorial Review Board) formed under the auspices of the World Wide Web Consortium
77(W3C) in 1996. It was chaired by Jon Bosak of Sun Microsystems with the active
78participation of an XML Special Interest Group (previously known as the SGML
79Working Group) also organized by the W3C. The membership of the XML Working
80Group is given in an appendix. Dan Connolly served as the Working Group's contact with
81the W3C.</p><p>The design goals for XML are:</p><ol type="1"><li><p>XML shall be straightforwardly usable over the Internet.</p></li><li><p>XML shall support a wide variety of applications.</p></li><li><p>XML shall be compatible with SGML.</p></li><li><p>It shall be easy to write programs which process XML documents.</p></li><li><p>The number of optional features in XML is to be kept to the absolute
82minimum, ideally zero.</p></li><li><p>XML documents should be human-legible and reasonably clear.</p></li><li><p>The XML design should be prepared quickly.</p></li><li><p>The design of XML shall be formal and concise.</p></li><li><p>XML documents shall be easy to create.</p></li><li><p>Terseness in XML markup is of minimal importance.</p></li></ol><p>This specification, together with associated standards (Unicode
83<a href="#Unicode">[Unicode]</a> and ISO/IEC 10646 <a href="#ISO10646">[ISO/IEC 10646]</a>
84for characters, Internet RFC 3066 <a href="#RFC1766">[IETF RFC 3066]</a> for
85language identification tags, ISO 639 <a href="#ISO639">[ISO 639]</a>
86for language name codes, and ISO 3166 <a href="#ISO3166">[ISO 3166]</a> for
87country name codes), provides all the information necessary to
88understand XML Version 1.1 and construct computer
89programs to process it.</p><p>This version of the XML specification may be distributed freely, as long as
90all text and legal notices remain intact.</p></div><div class="div2"> <h3><a name="sec-terminology" id="sec-terminology" />1.2 Terminology</h3><p>The terminology used to describe XML documents is defined in the body of
91this specification. <span class="mustard">The key words <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">SHALL</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">SHALL NOT</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD NOT</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">RECOMMENDED</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>, and <em class="rfc2119" title="Keyword in RFC 2119 context">OPTIONAL</em>, when <em class="rfc2119" title="Keyword in RFC 2119 context">EMPHASIZED</em>, are to be interpreted as described in <a href="#rfc2119">[IETF RFC 2119]</a>. In addition, </span>the terms defined in the following list are used in building
92those definitions and in describing the actions of an XML processor:</p><dl><dt class="label">error</dt><dd><p>[<a name="dt-error" id="dt-error" title="Error">Definition</a>: A violation of the rules of this specification;
93results are undefined. <span class="mustard">Unless otherwise specified, failure to observe a prescription of this specification indicated by one of the keywords <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em>, <em class="rfc2119" title="Keyword in RFC 2119 context">SHALL</em> and <em class="rfc2119" title="Keyword in RFC 2119 context">SHALL NOT</em> is an error.</span> Conforming software <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> detect and report an error
94and <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> recover from it.]</p></dd><dt class="label">fatal error</dt><dd><p>[<a name="dt-fatal" id="dt-fatal" title="Fatal Error">Definition</a>: An error which a conforming <a title="XML Processor" href="#dt-xml-proc">XML processor</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> detect and report to the application.
95After encountering a fatal error, the processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> continue processing the
96data to search for further errors and <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> report such errors to the application.
97In order to support correction of errors, the processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> make unprocessed
98data from the document (with intermingled character data and markup) available
99to the application. Once a fatal error is detected, however, the processor
100<em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> continue normal processing (i.e., it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> continue to pass character
101data and information about the document's logical structure to the application
102in the normal way).]</p></dd><dt class="label">at user option</dt><dd><p>[<a name="dt-atuseroption" id="dt-atuseroption" title="At user option">Definition</a>: Conforming software
103<em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> or <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> (depending on the modal verb in the sentence) behave as described;
104if it does, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> provide users a means to enable or disable the behavior
105described.]</p></dd><dt class="label">validity constraint</dt><dd><p>[<a name="dt-vc" id="dt-vc" title="Validity constraint">Definition</a>: A rule which applies to
106all <a title="Validity" href="#dt-valid">valid</a> XML documents. Violations of validity
107constraints are errors; they <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>, at user option, be reported by <a title="Validating Processor" href="#dt-validating">validating XML processors</a>.]</p></dd><dt class="label">well-formedness constraint</dt><dd><p>[<a name="dt-wfc" id="dt-wfc" title="Well-formedness constraint">Definition</a>: A rule which applies
108to all <a title="Well-Formed" href="#dt-wellformed">well-formed</a> XML documents. Violations
109of well-formedness constraints are <a title="Fatal Error" href="#dt-fatal">fatal errors</a>.]</p></dd><dt class="label">match</dt><dd><p>[<a name="dt-match" id="dt-match" title="match">Definition</a>: (Of strings or names:) Two strings
110or names being compared <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be identical. Characters with multiple possible
111representations in Unicode (e.g. characters with both precomposed and
112base+diacritic forms) match only if they have the same representation in both
113strings. No
114case folding is performed. (Of strings and rules in the grammar:) A string
115matches a grammatical production if it belongs to the language generated by
116that production. (Of content and content models:) An element matches its declaration
117when it conforms in the fashion described in the constraint <b>[VC: <a href="#elementvalid">Element Valid</a>]</b>.]</p></dd><dt class="label">for compatibility</dt><dd><p>[<a name="dt-compat" id="dt-compat" title="For Compatibility">Definition</a>: Marks
118a sentence describing a feature of XML included solely to ensure
119that XML remains compatible with SGML.]</p></dd><dt class="label">for interoperability</dt><dd><p>[<a name="dt-interop" id="dt-interop" title="For interoperability">Definition</a>: Marks
120a sentence describing a non-binding recommendation included to increase
121the chances that XML documents can be processed by the existing installed
122base of SGML processors which predate the WebSGML Adaptations Annex to ISO 8879.]</p></dd></dl><p></p></div><div class="div2"> <h3><a name="sec-xml11" id="sec-xml11" />1.3 Rationale and list of changes for XML 1.1</h3><p>The W3C's XML 1.0 Recommendation was first issued in 1998, and
123despite the issuance of many errata culminating in a Third Edition
124of 2004, has remained (by intention) unchanged with respect to what
125is well-formed XML and what is not. This stability has been
126extremely useful for interoperability. However, the Unicode
127Standard on which XML 1.0 relies for character specifications has
128not remained static, evolving from version 2.0 to version 4.0 and
129beyond. Characters not present in Unicode 2.0 may already be used
130in XML 1.0 character data. However, they are not allowed in XML
131names such as element type names, attribute names, enumerated
132attribute values, processing instruction targets, and so on. In
133addition, some characters that should have been permitted in XML
134names were not, due to oversights and inconsistencies in Unicode
1352.0.</p><p>The overall philosophy of names has changed since XML 1.0.
136Whereas XML 1.0 provided a rigid definition of names, wherein
137everything that was not permitted was forbidden, XML 1.1 names are
138designed so that everything that is not forbidden (for a specific
139reason) is permitted. Since Unicode will continue to grow past
140version 4.0, further changes to XML can be avoided by allowing
141almost any character, including those not yet assigned, in
142names.</p><p>In addition, XML 1.0 attempts to adapt to the line-end
143conventions of various modern operating systems, but discriminates
144against the conventions used on IBM and IBM-compatible mainframes.
145As a result, XML documents on mainframes are not plain text files
146according to the local conventions. XML 1.0 documents generated on
147mainframes must either violate the local line-end conventions, or
148employ otherwise unnecessary translation phases before parsing and
149after generation. Allowing straightforward interoperability is
150particularly important when data stores are shared between
151mainframe and non-mainframe systems (as opposed to being copied
152from one to the other). Therefore XML 1.1 adds NEL (#x85) to the
153list of line-end characters. For completeness, the Unicode line
154separator character, #x2028, is also supported.
155</p><p>Finally, there is considerable demand to define a standard representation
156of arbitrary Unicode characters in XML documents. Therefore, XML 1.1
157allows the use of character references to the control characters #x1 through
158#x1F, most of which are forbidden in XML 1.0. For reasons of robustness,
159however, these characters still cannot be used directly in documents. In
160order to improve the robustness of character encoding detection, the additional
161control characters #x7F through #x9F, which were freely allowed in XML 1.0
162documents, now must also appear only as character references. (Whitespace
163characters are of course exempt.) The minor sacrifice of backward compatibility
164is considered not significant. Due to potential problems with APIs,
165#x0 is still forbidden both directly and as a character reference.
166</p><p>Finally, XML 1.1 defines a set of constraints called "full
167normalization" on XML documents, which document creators
168<em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> adhere to, and document processors
169<em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> verify. Using fully normalized documents
170ensures that identity comparisons of names, attribute values, and
171character content can be made correctly by simple binary comparison of
172Unicode strings.</p><p>A new XML version, rather than a set of errata to XML 1.0, is
173being created because the changes affect the definition of
174well-formed documents. XML 1.0 processors must continue to reject
175documents that contain new characters in XML names, new line-end
176conventions, and references to control characters. The distinction between XML 1.0 and XML 1.1 documents
177is indicated by the version number information in the XML
178declaration at the start of each document.
179</p></div></div><div class="div1"> <h2><a name="sec-documents" id="sec-documents" />2 Documents</h2><p>[<a name="dt-xml-doc" id="dt-xml-doc" title="XML Document">Definition</a>: A data object is an <b>XML
180document</b> if it is <a title="Well-Formed" href="#dt-wellformed">well-formed</a>,
181as defined in this specification. A well-formed XML document <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> in addition
182be <a title="Validity" href="#dt-valid">valid</a> if it meets certain further constraints.]</p><p>Each XML document has both a logical and a physical structure. Physically,
183the document is composed of units called <a title="Entity" href="#dt-entity">entities</a>.
184An entity <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> <a title="Entity Reference" href="#dt-entref">refer</a> to other entities to
185cause their inclusion in the document. A document begins in a "root"
186or <a title="Document Entity" href="#dt-docent">document entity</a>. Logically, the document
187is composed of declarations, elements, comments, character references, and
188processing instructions, all of which are indicated in the document by explicit
189markup. The logical and physical structures <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> nest properly, as described
190in <a href="#wf-entities"><b>4.3.2 Well-Formed Parsed Entities</b></a>.</p><div class="div2"> <h3><a name="sec-well-formed" id="sec-well-formed" />2.1 Well-Formed XML Documents</h3><p>[<a name="dt-wellformed" id="dt-wellformed" title="Well-Formed">Definition</a>: A textual object is a <b>well-formed</b>
191XML document if:]</p><ol type="1"><li><p>Taken as a whole, it matches the production labeled <a href="#NT-document">document</a>.</p></li><li><p>It meets all the well-formedness constraints given in this specification.</p></li><li><p>Each of the <a title="Text Entity" href="#dt-parsedent">parsed entities</a>
192which is referenced directly or indirectly within the document is <a
193title="Well-Formed"
194href="#dt-wellformed">well-formed</a>.</p></li></ol> <h5><a
195name="document" id="document" />Document</h5><table class="scrap"
196summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-document"
197id="NT-document"
198/>[1]   </td><td><code>document</code></td><td>   ::=   </td><td><code><a
199href="#NT-prolog">prolog</a> <a href="#NT-element">element</a> <a
200href="#NT-Misc">Misc</a>* - <a href="#NT-Char">Char</a>* <a
201href="#NT-RestrictedChar">RestrictedChar</a> <a href="#NT-Char">Char</a>*</code></td></tr></tbody></table><p>Matching the <a href="#NT-document">document</a> production implies that:</p><ol type="1"><li><p>It contains one or more <a title="Element" href="#dt-element">elements</a>.</p></li><li><p>[<a name="dt-root" id="dt-root" title="Root Element">Definition</a>: There is exactly one element,
202called the <b>root</b>, or document element, no part of which appears
203in the <a title="Content" href="#dt-content">content</a> of any other element.] For
204all other elements, if the <a title="Start-Tag" href="#dt-stag">start-tag</a> is in
205the content of another element, the <a title="End Tag" href="#dt-etag">end-tag</a>
206is in the content of the same element. More simply stated, the elements,
207delimited by start- and end-tags, nest properly within each other.</p></li></ol><p>[<a name="dt-parentchild" id="dt-parentchild" title="Parent/Child">Definition</a>: As a consequence of this,
208for each non-root element <code>C</code> in the document, there is one other element <code>P</code>
209in the document such that <code>C</code> is in the content of <code>P</code>, but
210is not in the content of any other element that is in the content of <code>P</code>. <code>P</code>
211is referred to as the <b>parent</b> of <code>C</code>, and <code>C</code> as
212a <b>child</b> of <code>P</code>.]</p></div><div class="div2"> <h3><a name="charsets" id="charsets" />2.2 Characters</h3><p>[<a name="dt-text" id="dt-text" title="Text">Definition</a>: A parsed entity contains <b>text</b>,
213a sequence of <a title="Character" href="#dt-character">characters</a>, which may
214represent markup or character data.] [<a name="dt-character" id="dt-character" title="Character">Definition</a>: A <b>character</b>
215is an atomic unit of text as specified by <span>ISO/IEC 10646 <a href="#ISO10646">[ISO/IEC 10646]</a></span>. Legal characters are tab, carriage
216return, line feed, and the legal characters
217of Unicode and ISO/IEC 10646. The
218versions of these standards cited in <a href="#sec-existing-stds"><b>A.1 Normative References</b></a> were
219current at the time this document was prepared. New characters may be added
220to these standards by amendments or new editions. Consequently, XML processors
221<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> accept any character in the range specified for <a href="#NT-Char">Char</a>.]</p> <h5><a name="char32" id="char32" />Character Range</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-Char" id="NT-Char" />[2]   </td><td><code>Char</code></td><td>   ::=   </td><td><code>[#x1-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]</code></td><td><i>/* any Unicode character, excluding the surrogate blocks, FFFE, and FFFF. */</i></td></tr><tr valign="baseline"><td><a name="NT-RestrictedChar" id="NT-RestrictedChar" />[2a]   </td><td><code>RestrictedChar</code></td><td>   ::=   </td><td><code>[#x1-#x8] | [#xB-#xC] | [#xE-#x1F] | [#x7F-#x84] | [#x86-#x9F]</code></td></tr></tbody></table><p>The mechanism for encoding character code points into bit patterns <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>
222vary from entity to entity. All XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> accept the UTF-8 and UTF-16
223encodings of <span> Unicode
224<a href="#Unicode">[Unicode]</a></span>;
225the mechanisms for signaling which of the two is in use,
226or for bringing other encodings into play, are discussed later, in <a href="#charencoding"><b>4.3.3 Character Encoding in Entities</b></a>.</p><div class="note"><p class="prefix"><b>Note:</b></p><p>Document authors are encouraged to avoid
227"compatibility characters", as defined
228in Unicode <a href="#Unicode">[Unicode]</a>.
229The characters defined in the following ranges are also
230discouraged. They are either control characters or permanently undefined Unicode
231characters:</p><div class="exampleInner"><pre>
232[#x7F-#x84], [#x86-#x9F], [#xFDD0-#xFDDF],
233[#1FFFE-#x1FFFF], [#2FFFE-#x2FFFF], [#3FFFE-#x3FFFF],
234[#4FFFE-#x4FFFF], [#5FFFE-#x5FFFF], [#6FFFE-#x6FFFF],
235[#7FFFE-#x7FFFF], [#8FFFE-#x8FFFF], [#9FFFE-#x9FFFF],
236[#AFFFE-#xAFFFF], [#BFFFE-#xBFFFF], [#CFFFE-#xCFFFF],
237[#DFFFE-#xDFFFF], [#EFFFE-#xEFFFF], [#FFFFE-#xFFFFF],
238[#10FFFE-#x10FFFF].</pre></div></div></div><div class="div2"> <h3><a name="sec-common-syn" id="sec-common-syn" />2.3 Common Syntactic Constructs</h3><p>This section defines some symbols used widely in the grammar.</p><p><a href="#NT-S">S</a> (white space) consists of one or more space (#x20)
239characters, carriage returns, line feeds, or tabs.</p> <h5><a name="white" id="white" />White Space</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-S" id="NT-S" />[3]   </td><td><code>S</code></td><td>   ::=   </td><td><code>(#x20 | #x9 | #xD | #xA)+</code></td></tr></tbody></table><div class="note"><p class="prefix"><b>Note:</b></p><p>The presence of #xD in the above production is
240maintained purely for backward compatibility with the
241<a href="http://www.w3.org/TR/1998/REC-xml-19980210">First Edition</a>.
242As explained in <a href="#sec-line-ends"><b>2.11 End-of-Line Handling</b></a>,
243all #xD characters literally present in an XML document
244are either removed or replaced by #xA characters before
245any other processing is done. The only way to get a #xD character to match this production is to
246use a character reference in an entity value literal.</p></div><p>[<a name="dt-name" id="dt-name" title="Name">Definition</a>: A <b>Name</b> is a token beginning
247with a letter or one of a few punctuation characters, and continuing with
248letters, digits, hyphens, underscores, colons, or full stops, together known
249as name characters.] Names beginning with the string "<code>xml</code>",
250or <span>with</span> any string which would match <code>(('X'|'x') ('M'|'m') ('L'|'l'))</code>,
251are reserved for standardization in this or future versions of this specification.</p><div class="note"><p class="prefix"><b>Note:</b></p><p>The
252Namespaces in XML Recommendation <a href="#xml-names">[XML Names]</a> assigns a meaning
253to names containing colon characters. Therefore, authors should not use the
254colon in XML names except for namespace purposes, but XML processors must
255accept the colon as a name character.</p></div><p>An <a href="#NT-Nmtoken">Nmtoken</a> (name token) is any mixture of name
256characters.</p><p>The first character of a Name <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be a NameStartChar, and any
257other characters <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be NameChars; this mechanism is used to
258prevent names from beginning with European (ASCII) digits or with
259basic combining characters. Almost all characters are permitted in
260names, except those which either are or reasonably could be used as
261delimiters. The intention is to be inclusive rather than exclusive,
262so that writing systems not yet encoded in Unicode can be used in
263XML names. See <a href="#sec-suggested-names"><b>I Suggestions for XML Names</b></a> for suggestions on the creation of
264names.</p><p>Document authors are encouraged to use names which are
265meaningful words or combinations of words in natural languages, and
266to avoid symbolic or white space characters in names. Note that
267COLON, HYPHEN-MINUS, FULL STOP (period), LOW LINE (underscore), and
268MIDDLE DOT are explicitly permitted.</p><p>The ASCII symbols and punctuation marks, along with a fairly
269large group of Unicode symbol characters, are excluded from names
270because they are more useful as delimiters in contexts where XML
271names are used outside XML documents; providing this group gives
272those contexts hard guarantees about what <em>cannot</em> be part of
273an XML name. The character #x037E, GREEK QUESTION MARK, is excluded
274because when normalized it becomes a semicolon, which could change
275the meaning of entity references.</p> <h5><a name="IDABN1S" id="IDABN1S" />Names and Tokens</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-NameStartChar" id="NT-NameStartChar" />[4]   </td><td><code>NameStartChar</code></td><td>   ::=   </td><td><code>":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-NameChar" id="NT-NameChar" />[4a]   </td><td><code>NameChar</code></td><td>   ::=   </td><td><code><a href="#NT-NameStartChar">NameStartChar</a> | "-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-Name" id="NT-Name" />[5]   </td><td><code>Name</code></td><td>   ::=   </td><td><code><a href="#NT-NameStartChar">NameStartChar</a> (<a href="#NT-NameChar">NameChar</a>)*</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-Names" id="NT-Names" />[6]   </td><td><code>Names</code></td><td>   ::=   </td><td><code><a href="#NT-Name">Name</a> (#x20 <a href="#NT-Name">Name</a>)*</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-Nmtoken" id="NT-Nmtoken" />[7]   </td><td><code>Nmtoken</code></td><td>   ::=   </td><td><code>(<a href="#NT-NameChar">NameChar</a>)+</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-Nmtokens" id="NT-Nmtokens" />[8]   </td><td><code>Nmtokens</code></td><td>   ::=   </td><td><code><a href="#NT-Nmtoken">Nmtoken</a> (#x20 <a href="#NT-Nmtoken">Nmtoken</a>)*</code></td></tr></tbody></table><div class="note"><p class="prefix"><b>Note:</b></p><p>The <a href="#NT-Names">Names</a>
276and <a href="#NT-Nmtokens">Nmtokens</a> productions are used to define the validity
277of tokenized attribute values after normalization (see <a href="#sec-attribute-types"><b>3.3.1 Attribute Types</b></a>).</p></div><p>Literal data is any quoted string not containing the quotation mark used
278as a delimiter for that string. Literals are used for specifying the content
279of internal entities (<a href="#NT-EntityValue">EntityValue</a>), the values
280of attributes (<a href="#NT-AttValue">AttValue</a>), and external identifiers
281(<a href="#NT-SystemLiteral">SystemLiteral</a>). Note that a <a href="#NT-SystemLiteral">SystemLiteral</a>
282can be parsed without scanning for markup.</p> <h5><a name="IDAFR1S" id="IDAFR1S" />Literals</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-EntityValue" id="NT-EntityValue" />[9]   </td><td><code>EntityValue</code></td><td>   ::=   </td><td><code>'"' ([^%&amp;"] | <a href="#NT-PEReference">PEReference</a>
283| <a href="#NT-Reference">Reference</a>)* '"' </code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>|  "'" ([^%&amp;'] | <a href="#NT-PEReference">PEReference</a> | <a href="#NT-Reference">Reference</a>)* "'"</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-AttValue" id="NT-AttValue" />[10]   </td><td><code>AttValue</code></td><td>   ::=   </td><td><code>'"' ([^&lt;&amp;"] | <a href="#NT-Reference">Reference</a>)*
284'"' </code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>|  "'" ([^&lt;&amp;'] | <a href="#NT-Reference">Reference</a>)*
285"'"</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-SystemLiteral" id="NT-SystemLiteral" />[11]   </td><td><code>SystemLiteral</code></td><td>   ::=   </td><td><code>('"' [^"]* '"') | ("'" [^']* "'") </code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-PubidLiteral" id="NT-PubidLiteral" />[12]   </td><td><code>PubidLiteral</code></td><td>   ::=   </td><td><code>'"' <a href="#NT-PubidChar">PubidChar</a>* '"'
286| "'" (<a href="#NT-PubidChar">PubidChar</a> - "'")* "'"</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-PubidChar" id="NT-PubidChar" />[13]   </td><td><code>PubidChar</code></td><td>   ::=   </td><td><code>#x20 | #xD | #xA | [a-zA-Z0-9] | [-'()+,./:=?;!*#@$_%]</code></td></tr></tbody></table><div class="note"><p class="prefix"><b>Note:</b></p><p>Although
287the <a href="#NT-EntityValue">EntityValue</a> production allows the definition
288of a general entity consisting of a single explicit <code>&lt;</code> in the literal
289(e.g., <code>&lt;!ENTITY mylt "&lt;"&gt;</code>), it is strongly advised to avoid
290this practice since any reference to that entity will cause a well-formedness
291error.</p></div></div><div class="div2"> <h3><a name="syntax" id="syntax" />2.4 Character Data and Markup</h3><p><a title="Text" href="#dt-text">Text</a> consists of intermingled <a title="Character Data" href="#dt-chardata">character data</a> and markup. [<a name="dt-markup" id="dt-markup" title="Markup">Definition</a>: <b>Markup</b> takes the form of <a title="Start-Tag" href="#dt-stag">start-tags</a>, <a title="End Tag" href="#dt-etag">end-tags</a>, <a title="Empty" href="#dt-empty">empty-element tags</a>, <a title="Entity Reference" href="#dt-entref">entity references</a>, <a title="Character Reference" href="#dt-charref">character
292references</a>, <a title="Comment" href="#dt-comment">comments</a>, <a title="CDATA Section" href="#dt-cdsection">CDATA section</a> delimiters, <a title="Document Type Declaration" href="#dt-doctype">document
293type declarations</a>, <a title="Processing instruction" href="#dt-pi">processing instructions</a>, <a href="#NT-XMLDecl">XML declarations</a>, <a href="#NT-TextDecl">text declarations</a>,
294and any white space that is at the top level of the document entity (that
295is, outside the document element and not inside any other markup).]</p><p>[<a name="dt-chardata" id="dt-chardata" title="Character Data">Definition</a>: All text that is not markup
296constitutes the <b>character data</b> of the document.]</p><p>The ampersand character (&amp;) and the left angle bracket (&lt;) <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> appear
297in their literal form<span class="mustard">, except</span> when used as markup delimiters, or
298within a <a title="Comment" href="#dt-comment">comment</a>, a <a title="Processing instruction" href="#dt-pi">processing
299instruction</a>, or a <a title="CDATA Section" href="#dt-cdsection">CDATA section</a>.
300
301If they are needed elsewhere, they <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be <a title="escape" href="#dt-escape">escaped</a>
302using either <a title="Character Reference" href="#dt-charref">numeric character references</a>
303or the strings "<code>&amp;amp;</code>" and "<code>&amp;lt;</code>"
304respectively. The right angle bracket (&gt;) <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be represented using the string "<code>&amp;gt;</code>",
305and <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>, <a title="For Compatibility" href="#dt-compat">for compatibility</a>, be escaped
306using <span>either</span> "<code>&amp;gt;</code>" or a character reference when it
307appears in the string "<code>]]&gt;</code>" in content, when
308that string is not marking the end of a <a title="CDATA Section" href="#dt-cdsection">CDATA
309section</a>.</p><p>In the content of elements, character data is any string of characters
310which does not contain the start-delimiter of any markup or the
311CDATA-section-close delimiter,
312"<code>]]&gt;</code>".
313In a CDATA section,
314character data is any string of characters not including the CDATA-section-close
315delimiter.</p><p>To allow attribute values to contain both single and double quotes, the
316apostrophe or single-quote character (') <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be represented as "<code>&amp;apos;</code>",
317and the double-quote character (") as "<code>&amp;quot;</code>".</p> <h5><a name="IDASZ1S" id="IDASZ1S" />Character Data</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-CharData" id="NT-CharData" />[14]   </td><td><code>CharData</code></td><td>   ::=   </td><td><code>[^&lt;&amp;]* - ([^&lt;&amp;]* ']]&gt;' [^&lt;&amp;]*)</code></td></tr></tbody></table></div><div class="div2"> <h3><a name="sec-comments" id="sec-comments" />2.5 Comments</h3><p>[<a name="dt-comment" id="dt-comment" title="Comment">Definition</a>: <b>Comments</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> appear
318anywhere in a document outside other <a title="Markup" href="#dt-markup">markup</a>;
319in addition, they <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> appear within the document type declaration at places
320allowed by the grammar. They are not part of the document's <a title="Character Data" href="#dt-chardata">character
321data</a>; an XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>, but need not, make it possible for an
322application to retrieve the text of comments. <a title="For Compatibility" href="#dt-compat">For
323compatibility</a>, the string "<code>--</code>" (double-hyphen)
324<em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> occur within comments.] Parameter
325entity references <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> be</span> recognized within comments.</p> <h5><a name="IDAL11S" id="IDAL11S" />Comments</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-Comment" id="NT-Comment" />[15]   </td><td><code>Comment</code></td><td>   ::=   </td><td><code>'&lt;!--' ((<a href="#NT-Char">Char</a> - '-') | ('-'
326(<a href="#NT-Char">Char</a> - '-')))* '--&gt;'</code></td></tr></tbody></table><p>An example of a comment:</p><div class="exampleInner"><pre>&lt;!-- declarations for &lt;head&gt; &amp; &lt;body&gt; --&gt;</pre></div><p>Note
327that the grammar does not allow a comment ending in <code>---&gt;</code>. The
328following example is <em>not</em> well-formed.</p><div class="exampleInner"><pre>&lt;!-- B+, B, or B---&gt;</pre></div></div><div class="div2"> <h3><a name="sec-pi" id="sec-pi" />2.6 Processing Instructions</h3><p>[<a name="dt-pi" id="dt-pi" title="Processing instruction">Definition</a>: <b>Processing instructions</b>
329(PIs) allow documents to contain instructions for applications.]</p> <h5><a name="IDAD31S" id="IDAD31S" />Processing Instructions</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-PI" id="NT-PI" />[16]   </td><td><code>PI</code></td><td>   ::=   </td><td><code>'&lt;?' <a href="#NT-PITarget">PITarget</a> (<a href="#NT-S">S</a>
330(<a href="#NT-Char">Char</a>* - (<a href="#NT-Char">Char</a>* '?&gt;' <a href="#NT-Char">Char</a>*)))? '?&gt;'</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-PITarget" id="NT-PITarget" />[17]   </td><td><code>PITarget</code></td><td>   ::=   </td><td><code><a href="#NT-Name">Name</a> - (('X' | 'x') ('M' |
331'm') ('L' | 'l'))</code></td></tr></tbody></table><p>PIs are not part of the document's <a title="Character Data" href="#dt-chardata">character
332data</a>, but <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be passed through to the application. The PI begins
333with a target (<a href="#NT-PITarget">PITarget</a>) used to identify the application
334to which the instruction is directed. The target names "<code>XML</code>", "<code>xml</code>",
335and so on are reserved for standardization in this or future versions of this
336specification. The XML <a title="Notation" href="#dt-notation">Notation</a> mechanism
337<em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be used for formal declaration of PI targets. Parameter
338entity references <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> be</span> recognized within processing instructions.</p></div><div class="div2"> <h3><a name="sec-cdata-sect" id="sec-cdata-sect" />2.7 CDATA Sections</h3><p>[<a name="dt-cdsection" id="dt-cdsection" title="CDATA Section">Definition</a>: <b>CDATA sections</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> occur anywhere character data may occur; they are used to escape blocks
339of text containing characters which would otherwise be recognized as markup.
340CDATA sections begin with the string "<code>&lt;![CDATA[</code>"
341and end with the string "<code>]]&gt;</code>":]</p> <h5><a name="IDAOA2S" id="IDAOA2S" />CDATA Sections</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-CDSect" id="NT-CDSect" />[18]   </td><td><code>CDSect</code></td><td>   ::=   </td><td><code><a href="#NT-CDStart">CDStart</a> <a href="#NT-CData">CData</a> <a href="#NT-CDEnd">CDEnd</a></code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-CDStart" id="NT-CDStart" />[19]   </td><td><code>CDStart</code></td><td>   ::=   </td><td><code>'&lt;![CDATA['</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-CData" id="NT-CData" />[20]   </td><td><code>CData</code></td><td>   ::=   </td><td><code>(<a href="#NT-Char">Char</a>* - (<a href="#NT-Char">Char</a>*
342']]&gt;' <a href="#NT-Char">Char</a>*)) </code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-CDEnd" id="NT-CDEnd" />[21]   </td><td><code>CDEnd</code></td><td>   ::=   </td><td><code>']]&gt;'</code></td></tr></tbody></table><p>Within a CDATA section, only the <a href="#NT-CDEnd">CDEnd</a> string is
343recognized as markup, so that left angle brackets and ampersands may occur
344in their literal form; they need not (and cannot) be escaped using "<code>&amp;lt;</code>"
345and "<code>&amp;amp;</code>". CDATA sections cannot nest.</p><p>An example of a CDATA section, in which "<code>&lt;greeting&gt;</code>"
346and "<code>&lt;/greeting&gt;</code>" are recognized as <a title="Character Data" href="#dt-chardata">character data</a>, not <a title="Markup" href="#dt-markup">markup</a>:</p><div class="exampleInner"><pre>&lt;![CDATA[&lt;greeting&gt;Hello, world!&lt;/greeting&gt;]]&gt; </pre></div></div><div class="div2"> <h3><a name="sec-prolog-dtd" id="sec-prolog-dtd" />2.8 Prolog and Document Type Declaration</h3><p>[<a name="dt-xmldecl" id="dt-xmldecl" title="XML Declaration">Definition</a>: XML 1.1 documents <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
347begin with an <b>XML declaration</b> which specifies the version of
348XML being used.] For example, the following is a complete XML 1.1 document, <a title="Well-Formed" href="#dt-wellformed">well-formed</a> but not <a title="Validity" href="#dt-valid">valid</a>:</p><div class="exampleInner"><pre>&lt;?xml version="1.1"?&gt;
349&lt;greeting&gt;Hello, world!&lt;/greeting&gt; </pre></div><p>but the following is an XML 1.0 document because it
350does not have an XML declaration:</p><div class="exampleInner"><pre>&lt;greeting&gt;Hello, world!&lt;/greeting&gt;</pre></div><p>The function of the markup in an XML document is to describe its storage and
351logical structure and to associate <span>attribute
352name-value</span> pairs with its logical structures. XML provides a mechanism, the
353<a title="Document Type Declaration" href="#dt-doctype">document
354type declaration</a>, to define constraints on the logical structure
355and to support the use of predefined storage units. [<a name="dt-valid" id="dt-valid" title="Validity">Definition</a>: An XML document is <b>valid</b> if it has an associated
356document type declaration and if the document complies with the constraints
357expressed in it.]</p><p>The document type declaration <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> appear before the first <a title="Element" href="#dt-element">element</a>
358in the document.</p> <h5><a name="xmldoc" id="xmldoc" />Prolog</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-prolog" id="NT-prolog" />[22]   </td><td><code>prolog</code></td><td>   ::=   </td><td><code><a href="#NT-XMLDecl">XMLDecl</a> <a href="#NT-Misc">Misc</a>*
359(<a href="#NT-doctypedecl">doctypedecl</a> <a href="#NT-Misc">Misc</a>*)?</code></td></tr><tr valign="baseline"><td><a name="NT-XMLDecl" id="NT-XMLDecl" />[23]   </td><td><code>XMLDecl</code></td><td>   ::=   </td><td><code>'&lt;?xml' <a href="#NT-VersionInfo">VersionInfo</a> <a href="#NT-EncodingDecl">EncodingDecl</a>? <a href="#NT-SDDecl">SDDecl</a>? <a href="#NT-S">S</a>?'?&gt;'</code></td></tr><tr valign="baseline"><td><a name="NT-VersionInfo" id="NT-VersionInfo" />[24]   </td><td><code>VersionInfo</code></td><td>   ::=   </td><td><code><a href="#NT-S">S</a> 'version' <a href="#NT-Eq">Eq</a>
360("'" <a href="#NT-VersionNum">VersionNum</a> "'" | '"' <a href="#NT-VersionNum">VersionNum</a>
361'"')</code></td></tr><tr valign="baseline"><td><a name="NT-Eq" id="NT-Eq" />[25]   </td><td><code>Eq</code></td><td>   ::=   </td><td><code><a href="#NT-S">S</a>? '=' <a href="#NT-S">S</a>?</code></td></tr><tr valign="baseline"><td><a name="NT-VersionNum" id="NT-VersionNum" />[26]   </td><td><code>VersionNum</code></td><td>   ::=   </td><td><code>'1.1'</code></td></tr><tr valign="baseline"><td><a name="NT-Misc" id="NT-Misc" />[27]   </td><td><code>Misc</code></td><td>   ::=   </td><td><code><a href="#NT-Comment">Comment</a> | <a href="#NT-PI">PI</a>
362| <a href="#NT-S">S</a></code></td></tr></tbody></table><p>[<a name="dt-doctype" id="dt-doctype" title="Document Type Declaration">Definition</a>: The XML <b>document
363type declaration</b> contains or points to <a title="markup declaration" href="#dt-markupdecl">markup
364declarations</a> that provide a grammar for a class of documents. This
365grammar is known as a document type definition, or <b>DTD</b>. The document
366type declaration can point to an external subset (a special kind of <a title="External Entity" href="#dt-extent">external entity</a>) containing markup declarations,
367or can contain the markup declarations directly in an internal subset, or
368can do both. The DTD for a document consists of both subsets taken together.]</p><p>[<a name="dt-markupdecl" id="dt-markupdecl" title="markup declaration">Definition</a>: A <b>markup declaration</b>
369is an <a title="Element Type declaration" href="#dt-eldecl">element type declaration</a>, an <a title="Attribute-List Declaration" href="#dt-attdecl">attribute-list declaration</a>, an <a title="entity declaration" href="#dt-entdecl">entity
370declaration</a>, or a <a title="Notation Declaration" href="#dt-notdecl">notation declaration</a>.]
371These declarations <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be contained in whole or in part within <a title="Parameter entity" href="#dt-PE">parameter
372entities</a>, as described in the well-formedness and validity constraints
373below. For further
374information, see <a href="#sec-physical-struct"><b>4 Physical Structures</b></a>.</p> <h5><a name="dtd" id="dtd" />Document Type Definition</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-doctypedecl" id="NT-doctypedecl" />[28]   </td><td><code>doctypedecl</code></td><td>   ::=   </td><td><code>'&lt;!DOCTYPE' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a>
375(<a href="#NT-S">S</a> <a href="#NT-ExternalID">ExternalID</a>)? <a href="#NT-S">S</a>?
376('[' <a href="#NT-intSubset">intSubset</a> ']' <a href="#NT-S">S</a>?)? '&gt;'</code></td><td><a href="#vc-roottype">[VC: Root Element Type]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#ExtSubset">[WFC: External Subset]</a></td></tr><tr valign="baseline"><td><a name="NT-DeclSep" id="NT-DeclSep" />[28a]   </td><td><code>DeclSep</code></td><td>   ::=   </td><td><code><a href="#NT-PEReference">PEReference</a> | <a href="#NT-S">S</a></code></td><td><a href="#PE-between-Decls">[WFC: PE Between Declarations]</a></td></tr><tr valign="baseline"><td><a name="NT-intSubset" id="NT-intSubset" />[28b]   </td><td><code>intSubset</code></td><td>   ::=   </td><td><code>(<a href="#NT-markupdecl">markupdecl</a> | <a href="#NT-DeclSep">DeclSep</a>)*</code></td></tr><tr valign="baseline"><td><a name="NT-markupdecl" id="NT-markupdecl" />[29]   </td><td><code>markupdecl</code></td><td>   ::=   </td><td><code><a href="#NT-elementdecl">elementdecl</a> | <a href="#NT-AttlistDecl">AttlistDecl</a> | <a href="#NT-EntityDecl">EntityDecl</a>
377| <a href="#NT-NotationDecl">NotationDecl</a> | <a href="#NT-PI">PI</a> | <a href="#NT-Comment">Comment</a></code></td><td><a href="#vc-PEinMarkupDecl">[VC: Proper Declaration/PE Nesting]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#wfc-PEinInternalSubset">[WFC: PEs in Internal Subset]</a></td></tr></tbody></table><p>Note
378that it is possible to construct a well-formed document containing a <a href="#NT-doctypedecl">doctypedecl</a>
379that neither points to an external subset nor contains an internal subset.</p><p>The markup declarations <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be made up in whole or in part of the <a title="Replacement Text" href="#dt-repltext">replacement text</a> of <a title="Parameter entity" href="#dt-PE">parameter
380entities</a>. The productions later in this specification for individual
381nonterminals (<a href="#NT-elementdecl">elementdecl</a>, <a href="#NT-AttlistDecl">AttlistDecl</a>,
382and so on) describe the declarations <em>after</em> all the parameter
383entities have been <a title="Include" href="#dt-include">included</a>.</p><p>Parameter
384entity references are recognized anywhere in the DTD (internal and external
385subsets and external parameter entities), except in literals, processing instructions,
386comments, and the contents of ignored conditional sections (see <a href="#sec-condition-sect"><b>3.4 Conditional Sections</b></a>).
387They are also recognized in entity value literals. The use of parameter entities
388in the internal subset is restricted as described below.</p><div class="constraint"><p class="prefix"><a name="vc-roottype" id="vc-roottype" /><b>Validity constraint: Root Element Type</b></p><p>The <a href="#NT-Name">Name</a>
389in the document type declaration <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the element type of the <a title="Root Element" href="#dt-root">root element</a>.</p></div><div class="constraint"><p class="prefix"><a name="vc-PEinMarkupDecl" id="vc-PEinMarkupDecl" /><b>Validity constraint: Proper Declaration/PE Nesting</b></p><p>Parameter-entity <a title="Replacement Text" href="#dt-repltext">replacement text</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be properly nested with markup declarations. That is to say, if either
390the first character or the last character of a markup declaration (<a href="#NT-markupdecl">markupdecl</a>
391above) is contained in the replacement text for a <a title="Parameter-entity reference" href="#dt-PERef">parameter-entity
392reference</a>, both <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be contained in the same replacement text.</p></div><div class="constraint"><p class="prefix"><a name="wfc-PEinInternalSubset" id="wfc-PEinInternalSubset" /><b>Well-formedness constraint: PEs in Internal Subset</b></p><p>In
393the internal DTD subset, <a title="Parameter-entity reference" href="#dt-PERef">parameter-entity references</a> <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> occur within markup declarations; they <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> occur where markup declarations can occur</span>.
394(This does not apply to references that occur in external parameter entities
395or to the external subset.)</p></div><div class="constraint"><p class="prefix"><a name="ExtSubset" id="ExtSubset" /><b>Well-formedness constraint: External Subset</b></p><p>The external subset, if any, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the production for <a href="#NT-extSubset">extSubset</a>.</p></div><div class="constraint"><p class="prefix"><a name="PE-between-Decls" id="PE-between-Decls" /><b>Well-formedness constraint: PE Between Declarations</b></p><p>The replacement text of a parameter entity reference
396in a <a href="#NT-DeclSep">DeclSep</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the production <a href="#NT-extSubsetDecl">extSubsetDecl</a>.</p></div><p>Like the internal subset, the external subset and any external parameter
397entities referenced
398in a <a href="#NT-DeclSep">DeclSep</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> consist of a series of
399complete markup declarations of the types allowed by the non-terminal symbol <a href="#NT-markupdecl">markupdecl</a>, interspersed with white space or <a title="Parameter-entity reference" href="#dt-PERef">parameter-entity references</a>. However, portions of
400the contents of the external subset or of these
401external parameter entities <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> conditionally be ignored by using the <a title="conditional section" href="#dt-cond-section">conditional section</a> construct; this is not
402allowed in the internal subset<span> but is
403allowed in external parameter entities referenced in the internal subset</span>.</p> <h5><a name="ext-Subset" id="ext-Subset" />External Subset</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-extSubset" id="NT-extSubset" />[30]   </td><td><code>extSubset</code></td><td>   ::=   </td><td><code><a href="#NT-TextDecl">TextDecl</a>? <a href="#NT-extSubsetDecl">extSubsetDecl</a></code></td></tr><tr valign="baseline"><td><a name="NT-extSubsetDecl" id="NT-extSubsetDecl" />[31]   </td><td><code>extSubsetDecl</code></td><td>   ::=   </td><td><code>( <a href="#NT-markupdecl">markupdecl</a> | <a href="#NT-conditionalSect">conditionalSect</a> | <a href="#NT-DeclSep">DeclSep</a>)*</code></td></tr></tbody></table><p>The external subset and external parameter entities also differ from the
404internal subset in that in them, <a title="Parameter-entity reference" href="#dt-PERef">parameter-entity
405references</a> are permitted <em>within</em> markup declarations,
406not only <em>between</em> markup declarations.</p><p>An example of an XML document with a document type declaration:</p><div class="exampleInner"><pre>&lt;?xml version="1.1"?&gt;
407&lt;!DOCTYPE greeting SYSTEM "hello.dtd"&gt;
408&lt;greeting&gt;Hello, world!&lt;/greeting&gt; </pre></div><p>The <a title="System Identifier" href="#dt-sysid">system identifier</a> "<code>hello.dtd</code>"
409gives the address (a URI reference) of a DTD for the document.</p><p>The declarations can also be given locally, as in this example:</p><div class="exampleInner"><pre>&lt;?xml version="1.1" encoding="UTF-8" ?&gt;
410&lt;!DOCTYPE greeting [
411&lt;!ELEMENT greeting (#PCDATA)&gt;
412]&gt;
413&lt;greeting&gt;Hello, world!&lt;/greeting&gt;</pre></div><p>If both the external and internal subsets are used, the internal subset
414<span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> considered to occur before the external subset.
415This has the effect that entity and attribute-list declarations in the internal
416subset take precedence over those in the external subset.</p><p>XML 1.1 processors <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> accept XML 1.0
417documents as well. If a document is well-formed or valid XML 1.0, and provided it
418does not contain any control characters
419in the range [#x7F-#x9F] other than as character escapes, it may be
420made well-formed or valid XML 1.1 respectively simply by changing the
421version number.</p></div><div class="div2"> <h3><a name="sec-rmd" id="sec-rmd" />2.9 Standalone Document Declaration</h3><p>Markup declarations can affect the content of the document, as passed from
422an <a title="XML Processor" href="#dt-xml-proc">XML processor</a> to an application; examples
423are attribute defaults and entity declarations. The standalone document declaration,
424which <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> appear as a component of the XML declaration, signals whether or
425not there are such declarations which appear external to the <a title="Document Entity" href="#dt-docent">document
426entity</a>
427or in parameter entities. [<a name="dt-extmkpdecl" id="dt-extmkpdecl" title="External Markup Declaration">Definition</a>: An <b>external
428markup declaration</b> is defined as a markup declaration occurring in
429the external subset or in a parameter entity (external or internal, the latter
430being included because non-validating processors are not required to read
431them).]</p> <h5><a name="fulldtd" id="fulldtd" />Standalone Document Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-SDDecl" id="NT-SDDecl" />[32]   </td><td><code>SDDecl</code></td><td>   ::=   </td><td><code>#x20+ 'standalone' <a href="#NT-Eq">Eq</a>
432(("'" ('yes' | 'no') "'") | ('"' ('yes' | 'no') '"')) </code></td><td><a href="#vc-check-rmd">[VC: Standalone Document Declaration]</a></td></tr></tbody></table><p>In a standalone document declaration, the value "yes" indicates
433that there are no <a title="External Markup Declaration" href="#dt-extmkpdecl">external markup declarations</a> which
434affect the information passed from the XML processor to the application. The
435value "no" indicates that there are or may be such external
436markup declarations. Note that the standalone document declaration only denotes
437the presence of external <em>declarations</em>; the presence, in a document,
438of references to external <em>entities</em>, when those entities are internally
439declared, does not change its standalone status.</p><p>If there are no external markup declarations, the standalone document declaration
440has no meaning. If there are external markup declarations but there is no
441standalone document declaration, the value "no" is assumed.</p><p>Any XML document for which <code>standalone="no"</code> holds can be converted
442algorithmically to a standalone document, which may be desirable for some
443network delivery applications.</p><div class="constraint"><p class="prefix"><a name="vc-check-rmd" id="vc-check-rmd" /><b>Validity constraint: Standalone Document Declaration</b></p><p>The
444standalone document declaration <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> have the value "no" if
445any external markup declarations contain declarations of:</p><ul><li><p>attributes with <a title="Attribute Default" href="#dt-default">default</a> values,
446if elements to which these attributes apply appear in the document without
447specifications of values for these attributes, or</p></li><li><p>entities (other than <code>amp</code>,
448<code>lt</code>,
449<code>gt</code>,
450<code>apos</code>,
451<code>quot</code>), if <a title="Entity Reference" href="#dt-entref">references</a>
452to those entities appear in the document, or</p></li><li><p>attributes with
453tokenized types, where the
454attribute appears in the document with a value such that
455<a href="#AVNormalize"><cite>normalization</cite></a>
456will produce a different value from that which would be produced
457in the absence of the declaration, or</p></li><li><p>element types with <a title="Element content" href="#dt-elemcontent">element content</a>,
458if white space occurs directly within any instance of those types.</p></li></ul></div><p>An example XML declaration with a standalone document declaration:</p><div class="exampleInner"><pre>&lt;?xml version="1.1" standalone='yes'?&gt;</pre></div></div><div class="div2"> <h3><a name="sec-white-space" id="sec-white-space" />2.10 White Space Handling</h3><p>In editing XML documents, it is often convenient to use "white space"
459(spaces, tabs, and blank lines)
460to set apart the markup for greater readability. Such white space is typically
461not intended for inclusion in the delivered version of the document. On the
462other hand, "significant" white space that should be preserved
463in the delivered version is common, for example in poetry and source code.</p><p>An <a title="XML Processor" href="#dt-xml-proc">XML processor</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> always pass
464all characters in a document that are not markup through to the application.
465A <a title="Validating Processor" href="#dt-validating"> validating XML processor</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> also
466inform the application which of these characters constitute white space appearing
467in <a title="Element content" href="#dt-elemcontent">element content</a>.</p><p>A special <a title="Attribute" href="#dt-attr">attribute</a> named <code>xml:space</code> <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be attached to an element to signal an intention that in that element,
468white space should be preserved by applications. In valid documents, this
469attribute, like any other, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be <a title="Attribute-List Declaration" href="#dt-attdecl">declared</a>
470if it is used. When declared, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be given as an <a title="Enumerated Attribute&#xA;Values" href="#dt-enumerated">enumerated
471type</a> whose values
472are one or both of "default" and "preserve".
473For example:</p><div class="exampleInner"><pre>&lt;!ATTLIST poem xml:space (default|preserve) 'preserve'&gt;
474&lt;!ATTLIST pre xml:space (preserve) #FIXED 'preserve'&gt;</pre></div><p>The value "default" signals that applications' default white-space
475processing modes are acceptable for this element; the value "preserve"
476indicates the intent that applications preserve all the white space. This
477declared intent is considered to apply to all elements within the content
478of the element where it is specified, unless <span>overridden</span> with
479another instance of the <code>xml:space</code> attribute. <span>This specification does not give meaning to any value of <code>xml:space</code> other than "default" and "preserve". It is an error for other values to be specified; the XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> report the error or <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> recover by ignoring the attribute specification or by reporting the (erroneous) value to the application. Applications may ignore or reject erroneous values.</span></p><p>The <a title="Root Element" href="#dt-root">root element</a> of any document is considered
480to have signaled no intentions as regards application space handling, unless
481it provides a value for this attribute or the attribute is declared with a
482default value.</p></div><div class="div2"> <h3><a name="sec-line-ends" id="sec-line-ends" />2.11 End-of-Line Handling</h3><p>XML <a title="Text Entity" href="#dt-parsedent">parsed entities</a> are often stored
483in computer files which, for editing convenience, are organized into lines.
484These lines are typically separated by some combination of the characters
485CARRIAGE RETURN (#xD) and LINE FEED (#xA).</p><p>To
486simplify the tasks of <a title="Application" href="#dt-app">applications</a>, the
487<span><a title="XML Processor" href="#dt-xml-proc">XML
488processor</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> behave as if it</span> normalized all line breaks in external parsed
489entities (including the document entity) on input, before parsing, by translating
490
491<span>all of the following to a single #xA character:</span></p><ol type="1"><li><p>the two-character sequence #xD #xA</p></li><li><p>the two-character sequence #xD #x85</p></li><li><p>the single character #x85</p></li><li><p>the single character #x2028</p></li><li><p>any #xD character that is not immediately followed by #xA or #x85.</p></li></ol><p> The characters #x85 and #x2028 cannot be reliably recognized and
492translated until an entity's encoding declaration (if present) has
493been read. Therefore, it is a fatal error to use them within the XML
494declaration or text declaration.
495</p></div><div class="div2"> <h3><a name="sec-lang-tag" id="sec-lang-tag" />2.12 Language Identification</h3><p>In document processing, it is often useful to identify the natural or formal
496language in which the content is written. A special <a title="Attribute" href="#dt-attr">attribute</a>
497named <code>xml:lang</code> <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be inserted in documents to specify the language
498used in the contents and attribute values of any element in an XML document.
499In valid documents, this attribute, like any other, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be <a title="Attribute-List Declaration" href="#dt-attdecl">declared</a>
500if it is used. The
501values of the attribute are language identifiers as defined by <a href="#RFC1766">[IETF RFC 3066]</a>, <cite>Tags
502for the Identification of Languages</cite>, or its successor<span>; in addition, the empty string <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be specified</span>.</p><p>(Productions 33 through 38 have been removed.)</p><p>For example:</p><div class="exampleInner"><pre>&lt;p xml:lang="en"&gt;The quick brown fox jumps over the lazy dog.&lt;/p&gt;
503&lt;p xml:lang="en-GB"&gt;What colour is it?&lt;/p&gt;
504&lt;p xml:lang="en-US"&gt;What color is it?&lt;/p&gt;
505&lt;sp who="Faust" desc='leise' xml:lang="de"&gt;
506&lt;l&gt;Habe nun, ach! Philosophie,&lt;/l&gt;
507&lt;l&gt;Juristerei, und Medizin&lt;/l&gt;
508&lt;l&gt;und leider auch Theologie&lt;/l&gt;
509&lt;l&gt;durchaus studiert mit hei&amp;#xDF;em Bem&amp;#xFC;h'n.&lt;/l&gt;
510&lt;/sp&gt;</pre></div><p>The intent declared with <code>xml:lang</code> is considered to apply to
511all attributes and content of the element where it is specified, unless overridden
512with an instance of <code>xml:lang</code> on another element within that content. <span>In particular, the empty value of <code>xml:lang</code> is used on an element B to override a specification of <code>xml:lang</code> on an enclosing element A, without specifying another language. Within B, it is considered that there is no language information available, just as if <code>xml:lang</code> had not been specified on B or any of its ancestors.</span></p><div class="note"><p class="prefix"><b>Note:</b></p><p>Language information may also be provided by external transport protocols (e.g. HTTP or
513MIME). When available, this information may be used by XML applications, but the more local
514information provided by <code>xml:lang</code> should be considered to override it.
515</p></div><p>A simple declaration for <code>xml:lang</code> might take the form</p><div class="exampleInner"><pre>xml:lang <span>CDATA</span> #IMPLIED</pre></div><p>but specific default values <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> also be given, if appropriate. In a collection
516of French poems for English students, with glosses and notes in English, the <code>xml:lang</code>
517attribute might be declared this way:</p><div class="exampleInner"><pre>&lt;!ATTLIST poem xml:lang <span>CDATA</span> 'fr'&gt;
518&lt;!ATTLIST gloss xml:lang <span>CDATA</span> 'en'&gt;
519&lt;!ATTLIST note xml:lang <span>CDATA</span> 'en'&gt;</pre></div></div><div class="div2"> <h3><a name="sec-normalization-checking" id="sec-normalization-checking" />2.13 Normalization Checking</h3><p>All XML <a title="Text Entity" href="#dt-parsedent"> parsed
520entities</a> (including <a title="Document Entity" href="#dt-docent"> document
521entities</a>) <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be <a title="fully normalized" href="#dt-fullnorm">fully
522normalized</a> as per the definition of
523<a href="#sec-CharNorm"><b>B Definitions for Character Normalization</b></a> supplemented by the following definitions of
524<em><a name="dt-relconst" id="dt-relconst" />relevant constructs</em> for XML:</p><ol type="1"><li><p>The <a title="Replacement Text" href="#dt-repltext">
525replacement text</a> of all <a title="Text Entity" href="#dt-parsedent">parsed
526entities</a></p></li><li><p>All text matching, in context, one of the following
527productions:</p><ol type="a"><li><p><a href="#NT-CData">
528CData</a></p></li><li><p><a href="#NT-CharData">
529CharData</a></p></li><li><p><a href="#NT-content">
530content</a></p></li><li><p><a href="#NT-Name"> Name</a></p></li><li><p><a href="#NT-Nmtoken">
531Nmtoken</a></p></li></ol></li></ol><p>However, a document is still well-formed even if it is not
532<a title="fully normalized" href="#dt-fullnorm">fully normalized</a>.
533XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> provide a user option to verify that the document being
534processed is in <a title="fully normalized" href="#dt-fullnorm">fully normalized</a> form, and report to the application whether
535it is or not. The option to not verify <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be chosen only when the
536input text is <a title="certified" href="#dt-certified">certified</a>,
537as defined by <a href="#sec-CharNorm"><b>B Definitions for Character Normalization</b></a>.</p><p>The verification of full normalization <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be carried out as if by
538first verifying that the entity is in <a title="include-normalized" href="#dt-inclnorm">include-normalized</a>
539form as defined by <a href="#sec-CharNorm"><b>B Definitions for Character Normalization</b></a> and by then verifying that none of the relevant
540constructs listed above begins (after character references are
541expanded) with a <a title="composing character" href="#dt-compchar">composing character</a> as defined by
542<a href="#sec-CharNorm"><b>B Definitions for Character Normalization</b></a>.
543Non-validating processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> ignore possible
544denormalizations that would be caused by inclusion of external
545entities that they do not read.</p><div class="note"><p class="prefix"><b>Note:</b></p><p>The <a title="composing character" href="#dt-compchar">composing character</a> are all
546Unicode characters of non-zero combining class, plus a small number
547of class-zero characters that nevertheless take part as a
548non-initial character in certain Unicode canonical
549decompositions. Since these characters are meant to follow
550base characters, restricting relevant constructs (including
551content) from beginning with a <a title="composing character" href="#dt-compchar">composing character</a> does not
552meaningfully diminish the expressiveness of XML.</p></div><p>If, while verifying full normalization, a processor encounters
553characters for which it cannot determine the normalization
554properties (i.e., characters introduced in a version of Unicode <a href="#Unicode">[Unicode]</a>
555later than the one used in the implementation of the processor),
556then the processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>, at user option, ignore any possible
557denormalizations caused by these characters. The option to ignore
558those denormalizations <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD NOT</em> be chosen by applications when
559reliability or security are critical.</p><p> XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> transform the input to be in
560<a title="fully normalized" href="#dt-fullnorm">fully normalized</a> form.
561XML applications that create XML 1.1 output
562from either XML 1.1 or XML 1.0 input <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> ensure that the output
563is <a title="fully normalized" href="#dt-fullnorm">fully normalized</a>; it is not necessary for internal processing
564forms to be <a title="fully normalized" href="#dt-fullnorm">fully normalized</a>.</p><p>The purpose of this section is to strongly encourage XML
565processors to ensure that the creators of XML documents have
566properly normalized them, so that XML applications can make tests
567such as identity comparisons of strings without having to worry
568about the different possible "spellings" of strings which
569Unicode allows.
570</p><p>When entities are in a non-Unicode encoding, if the processor
571transcodes them to Unicode, it <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> use a normalizing transcoder.
572</p></div></div><div class="div1"> <h2><a name="sec-logical-struct" id="sec-logical-struct" />3 Logical Structures</h2><p>[<a name="dt-element" id="dt-element" title="Element">Definition</a>: Each <a title="XML Document" href="#dt-xml-doc">XML
573document</a> contains one or more <b>elements</b>, the boundaries
574of which are either delimited by <a title="Start-Tag" href="#dt-stag">start-tags</a>
575and <a title="End Tag" href="#dt-etag">end-tags</a>, or, for <a title="Empty" href="#dt-empty">empty</a>
576elements, by an <a title="empty-element tag" href="#dt-eetag">empty-element tag</a>. Each
577element has a type, identified by name, sometimes called its "generic
578identifier" (GI), and <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> have a set of attribute specifications.]
579Each attribute specification has a <a title="Attribute Name" href="#dt-attrname">name</a>
580and a <a title="Attribute Value" href="#dt-attrval">value</a>.</p> <h5><a name="IDATJ3S" id="IDATJ3S" />Element</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-element" id="NT-element" />[39]   </td><td><code>element</code></td><td>   ::=   </td><td><code><a href="#NT-EmptyElemTag">EmptyElemTag</a></code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| <a href="#NT-STag">STag</a> <a href="#NT-content">content</a> <a href="#NT-ETag">ETag</a></code></td><td><a href="#GIMatch">[WFC: Element Type Match]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#elementvalid">[VC: Element Valid]</a></td></tr></tbody></table><p>This specification does not constrain the semantics, use, or (beyond syntax)
581names of the element types and attributes, except that names beginning with
582a match to <code>(('X'|'x')('M'|'m')('L'|'l'))</code> are reserved for standardization
583in this or future versions of this specification.</p><div class="constraint"><p class="prefix"><a name="GIMatch" id="GIMatch" /><b>Well-formedness constraint: Element Type Match</b></p><p>The <a href="#NT-Name">Name</a>
584in an element's end-tag <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the element type in the start-tag.</p></div><div class="constraint"><p class="prefix"><a name="elementvalid" id="elementvalid" /><b>Validity constraint: Element Valid</b></p><p>An element is valid
585if there is a declaration matching <a href="#NT-elementdecl">elementdecl</a>
586where the <a href="#NT-Name">Name</a> matches the element type, and one of
587the following holds:</p><ol type="1"><li><p>The declaration matches <b>EMPTY</b> and the element has no <a title="Content" href="#dt-content">content</a> <span>(not even entity
588references, comments, PIs or white space)</span>.</p></li><li><p>The declaration matches <a href="#NT-children">children</a> and the
589sequence of <a title="Parent/Child" href="#dt-parentchild">child elements</a> belongs
590to the language generated by the regular expression in the content model,
591with optional white space<span>, comments and
592PIs (i.e. markup matching production [27] <a href="#NT-Misc">Misc</a>)</span> between the
593start-tag and the first child element, between child elements, or between
594the last child element and the end-tag. Note that a CDATA section containing
595only white space <span>or a reference
596to an entity whose replacement text is character references expanding to white
597space</span> <span>do</span> not
598match the nonterminal <a href="#NT-S">S</a>, and
599hence cannot appear in these positions<span>; however, a
600reference to an internal entity with a literal value consisting of character
601references expanding to white space does match <a href="#NT-S">S</a>, since its
602replacement text is the white space resulting from expansion of the character
603references</span>.</p></li><li><p>The declaration matches <a href="#NT-Mixed">Mixed</a> and the content
604<span>(after replacing
605any entity references with their replacement text)</span> consists of
606<a title="Character Data" href="#dt-chardata">character data</a><span>,
607<a title="Comment" href="#dt-comment">comments</a>, <a title="Processing instruction" href="#dt-pi">PIs</a></span> and <a title="Parent/Child" href="#dt-parentchild">child elements</a> whose types match names in the
608content model.</p></li><li><p>The declaration matches <b>ANY</b>, and the
609<span>content
610<span>(after replacing
611any entity references with their replacement text)</span>
612consists of character data and <a title="Parent/Child" href="#dt-parentchild">child elements</a>
613whose types</span>
614have been declared.</p></li></ol></div><div class="div2"> <h3><a name="sec-starttags" id="sec-starttags" />3.1 Start-Tags, End-Tags, and Empty-Element Tags</h3><p>[<a name="dt-stag" id="dt-stag" title="Start-Tag">Definition</a>: The beginning of every non-empty
615XML element is marked by a <b>start-tag</b>.]</p> <h5><a name="IDA3O3S" id="IDA3O3S" />Start-tag</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-STag" id="NT-STag" />[40]   </td><td><code>STag</code></td><td>   ::=   </td><td><code>'&lt;' <a href="#NT-Name">Name</a> (<a href="#NT-S">S</a> <a href="#NT-Attribute">Attribute</a>)* <a href="#NT-S">S</a>? '&gt;'</code></td><td><a href="#uniqattspec">[WFC: Unique Att Spec]</a></td></tr><tr valign="baseline"><td><a name="NT-Attribute" id="NT-Attribute" />[41]   </td><td><code>Attribute</code></td><td>   ::=   </td><td><code><a href="#NT-Name">Name</a> <a href="#NT-Eq">Eq</a> <a href="#NT-AttValue">AttValue</a></code></td><td><a href="#ValueType">[VC: Attribute Value Type]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#NoExternalRefs">[WFC: No External Entity References]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#CleanAttrVals">[WFC: No &lt; in Attribute Values]</a></td></tr></tbody></table><p>The <a href="#NT-Name">Name</a> in the start- and end-tags gives the element's <b>type</b>. [<a name="dt-attr" id="dt-attr" title="Attribute">Definition</a>: The <a href="#NT-Name">Name</a>-<a href="#NT-AttValue">AttValue</a>
616pairs are referred to as the <b>attribute specifications</b> of the
617element], [<a name="dt-attrname" id="dt-attrname" title="Attribute Name">Definition</a>: with the <a href="#NT-Name">Name</a> in each pair referred to as the <b>attribute name</b>]
618and [<a name="dt-attrval" id="dt-attrval" title="Attribute Value">Definition</a>: the content of the <a href="#NT-AttValue">AttValue</a> (the text between the <code>'</code> or <code>"</code>
619delimiters) as the <b>attribute value</b>.] Note
620that the order of attribute specifications in a start-tag or empty-element
621tag is not significant.</p><div class="constraint"><p class="prefix"><a name="uniqattspec" id="uniqattspec" /><b>Well-formedness constraint: Unique Att Spec</b></p><p><span class="mustard">An attribute name
622<em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> appear more than once in the same start-tag or empty-element tag.</p></div><div class="constraint"><p class="prefix"><a name="ValueType" id="ValueType" /><b>Validity constraint: Attribute Value Type</b></p><p>The attribute <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
623have been declared; the value <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be of the type declared for it. (For attribute
624types, see <a href="#attdecls"><b>3.3 Attribute-List Declarations</b></a>.)</p></div><div class="constraint"><p class="prefix"><a name="NoExternalRefs" id="NoExternalRefs" /><b>Well-formedness constraint: No External Entity References</b></p><p>Attribute
625values <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> contain direct or indirect entity references to external entities.</p></div><div class="constraint"><p class="prefix"><a name="CleanAttrVals" id="CleanAttrVals" /><b>Well-formedness constraint: No <code>&lt;</code> in Attribute Values</b></p><p>The <a title="Replacement Text" href="#dt-repltext">replacement text</a> of any entity
626referred to directly or indirectly in an attribute value <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> contain a <code>&lt;</code>.</p></div><p>An example of a start-tag:</p><div class="exampleInner"><pre>&lt;termdef id="dt-dog" term="dog"&gt;</pre></div><p>[<a name="dt-etag" id="dt-etag" title="End Tag">Definition</a>: The end of every element that begins
627with a start-tag <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be marked by an <b>end-tag</b> containing a name
628that echoes the element's type as given in the start-tag:]</p> <h5><a name="IDA3U3S" id="IDA3U3S" />End-tag</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-ETag" id="NT-ETag" />[42]   </td><td><code>ETag</code></td><td>   ::=   </td><td><code>'&lt;/' <a href="#NT-Name">Name</a> <a href="#NT-S">S</a>?
629'&gt;'</code></td></tr></tbody></table><p>An example of an end-tag:</p><div class="exampleInner"><pre>&lt;/termdef&gt;</pre></div><p>[<a name="dt-content" id="dt-content" title="Content">Definition</a>: The <a title="Text" href="#dt-text">text</a>
630between the start-tag and end-tag is called the element's <b>content</b>:]</p> <h5><a name="IDAKW3S" id="IDAKW3S" />Content of Elements</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-content" id="NT-content" />[43]   </td><td><code>content</code></td><td>   ::=   </td><td><code><a href="#NT-CharData">CharData</a>? ((<a href="#NT-element">element</a>
631| <a href="#NT-Reference">Reference</a> | <a href="#NT-CDSect">CDSect</a>
632| <a href="#NT-PI">PI</a> | <a href="#NT-Comment">Comment</a>) <a href="#NT-CharData">CharData</a>?)*</code></td></tr></tbody></table><p>[<a name="dt-empty" id="dt-empty" title="Empty">Definition</a>: An element
633with no <a href="#NT-content">content</a> is said to be <b>empty</b>.] The representation
634of an empty element is either a start-tag immediately followed by an end-tag,
635or an empty-element tag. [<a name="dt-eetag" id="dt-eetag" title="empty-element tag">Definition</a>: An <b>empty-element
636tag</b> takes a special form:]</p> <h5><a name="IDARY3S" id="IDARY3S" />Tags for Empty Elements</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-EmptyElemTag" id="NT-EmptyElemTag" />[44]   </td><td><code>EmptyElemTag</code></td><td>   ::=   </td><td><code>'&lt;' <a href="#NT-Name">Name</a> (<a href="#NT-S">S</a> <a href="#NT-Attribute">Attribute</a>)* <a href="#NT-S">S</a>? '/&gt;'</code></td><td><a href="#uniqattspec">[WFC: Unique Att Spec]</a></td></tr></tbody></table><p>Empty-element tags <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be used for any element which has no content, whether
637or not it is declared using the keyword <b>EMPTY</b>. <a title="For interoperability" href="#dt-interop">For
638interoperability</a>, the empty-element tag <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em>
639be used, and <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> only be used, for elements which are declared
640EMPTY.</p><p>Examples of empty elements:</p><div class="exampleInner"><pre>&lt;IMG align="left"
641src="http://www.w3.org/Icons/WWW/w3c_home" /&gt;
642&lt;br&gt;&lt;/br&gt;
643&lt;br/&gt;</pre></div></div><div class="div2"> <h3><a name="elemdecls" id="elemdecls" />3.2 Element Type Declarations</h3><p>The <a title="Element" href="#dt-element">element</a> structure of an <a title="XML Document" href="#dt-xml-doc">XML document</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>, for <a title="Validity" href="#dt-valid">validation</a>
644purposes, be constrained using element type and attribute-list declarations.
645An element type declaration constrains the element's <a title="Content" href="#dt-content">content</a>.</p><p>Element type declarations often constrain which element types can appear
646as <a title="Parent/Child" href="#dt-parentchild">children</a> of the element. At user
647option, an XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> issue a warning when a declaration mentions an
648element type for which no declaration is provided, but this is not an error.</p><p>[<a name="dt-eldecl" id="dt-eldecl" title="Element Type declaration">Definition</a>: An <b>element
649type declaration</b> takes the form:]</p> <h5><a name="IDAV13S" id="IDAV13S" />Element Type Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-elementdecl" id="NT-elementdecl" />[45]   </td><td><code>elementdecl</code></td><td>   ::=   </td><td><code>'&lt;!ELEMENT' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-S">S</a> <a href="#NT-contentspec">contentspec</a> <a href="#NT-S">S</a>?
650'&gt;'</code></td><td><a href="#EDUnique">[VC: Unique Element Type Declaration]</a></td></tr><tr valign="baseline"><td><a name="NT-contentspec" id="NT-contentspec" />[46]   </td><td><code>contentspec</code></td><td>   ::=   </td><td><code>'EMPTY' | 'ANY' | <a href="#NT-Mixed">Mixed</a>
651| <a href="#NT-children">children</a></code></td></tr></tbody></table><p>where the <a href="#NT-Name">Name</a> gives the element type being declared.</p><div class="constraint"><p class="prefix"><a name="EDUnique" id="EDUnique" /><b>Validity constraint: Unique Element Type Declaration</b></p><p><span class="mustard">An element
652type <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> be declared more than once.</p></div><p>Examples of element type declarations:</p><div class="exampleInner"><pre>&lt;!ELEMENT br EMPTY&gt;
653&lt;!ELEMENT p (#PCDATA|emph)* &gt;
654&lt;!ELEMENT %name.para; %content.para; &gt;
655&lt;!ELEMENT container ANY&gt;</pre></div><div class="div3"> <h4><a name="sec-element-content" id="sec-element-content" />3.2.1 Element Content</h4><p>[<a name="dt-elemcontent" id="dt-elemcontent" title="Element content">Definition</a>: An element <a title="Start-Tag" href="#dt-stag">type</a> has <b>element content</b> when elements
656of that type <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> contain only <a title="Parent/Child" href="#dt-parentchild">child</a>
657elements (no character data), optionally separated by white space (characters
658matching the nonterminal <a href="#NT-S">S</a>).] [<a name="dt-content-model" id="dt-content-model" title="Content model">Definition</a>: In this case, the constraint includes a <b>content
659model</b>, a simple grammar governing the allowed types of the
660child elements and the order in which they are allowed to appear.]
661The grammar is built on content particles (<a href="#NT-cp">cp</a>s), which
662consist of names, choice lists of content particles, or sequence lists of
663content particles:</p> <h5><a name="IDAP53S" id="IDAP53S" />Element-content Models</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-children" id="NT-children" />[47]   </td><td><code>children</code></td><td>   ::=   </td><td><code>(<a href="#NT-choice">choice</a> | <a href="#NT-seq">seq</a>)
664('?' | '*' | '+')?</code></td></tr><tr valign="baseline"><td><a name="NT-cp" id="NT-cp" />[48]   </td><td><code>cp</code></td><td>   ::=   </td><td><code>(<a href="#NT-Name">Name</a> | <a href="#NT-choice">choice</a>
665| <a href="#NT-seq">seq</a>) ('?' | '*' | '+')?</code></td></tr><tr valign="baseline"><td><a name="NT-choice" id="NT-choice" />[49]   </td><td><code>choice</code></td><td>   ::=   </td><td><code>'(' <a href="#NT-S">S</a>? <a href="#NT-cp">cp</a> ( <a href="#NT-S">S</a>? '|' <a href="#NT-S">S</a>? <a href="#NT-cp">cp</a> )+ <a href="#NT-S">S</a>? ')'</code></td><td><a href="#vc-PEinGroup">[VC: Proper Group/PE Nesting]</a></td></tr><tr valign="baseline"><td><a name="NT-seq" id="NT-seq" />[50]   </td><td><code>seq</code></td><td>   ::=   </td><td><code>'(' <a href="#NT-S">S</a>? <a href="#NT-cp">cp</a> ( <a href="#NT-S">S</a>? ',' <a href="#NT-S">S</a>? <a href="#NT-cp">cp</a> )* <a href="#NT-S">S</a>? ')'</code></td><td><a href="#vc-PEinGroup">[VC: Proper Group/PE Nesting]</a></td></tr></tbody></table><p>where each <a href="#NT-Name">Name</a> is the type of an element which
666<em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> appear as a <a title="Parent/Child" href="#dt-parentchild">child</a>. Any content
667particle in a choice list <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> appear in the <a title="Element content" href="#dt-elemcontent">element
668content</a> at the location where the choice list appears in the grammar;
669content particles occurring in a sequence list <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> each appear in the <a title="Element content" href="#dt-elemcontent">element content</a> in the order given in the list.
670The optional character following a name or list governs whether the element
671or the content particles in the list may occur one or more (<code>+</code>),
672zero or more (<code>*</code>), or zero or one times (<code>?</code>). The
673absence of such an operator means that the element or content particle <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
674appear exactly once. This syntax and meaning are identical to those used in
675the productions in this specification.</p><p>The content of an element matches a content model if and only if it is
676possible to trace out a path through the content model, obeying the sequence,
677choice, and repetition operators and matching each element in the content
678against an element type in the content model. <a title="For Compatibility" href="#dt-compat">For
679compatibility</a>, it is an error if <span>the content model
680allows an element to match more than one occurrence of an element type in the
681content model</span>. For more information, see <a href="#determinism"><b>D Deterministic Content Models</b></a>.</p><div class="constraint"><p class="prefix"><a name="vc-PEinGroup" id="vc-PEinGroup" /><b>Validity constraint: Proper Group/PE Nesting</b></p><p>Parameter-entity <a title="Replacement Text" href="#dt-repltext">replacement text</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be properly nested with parenthesized
682groups. That is to say, if either of the opening or closing parentheses in
683a <a href="#NT-choice">choice</a>, <a href="#NT-seq">seq</a>, or <a href="#NT-Mixed">Mixed</a>
684construct is contained in the replacement text for a <a title="Parameter-entity reference" href="#dt-PERef">parameter
685entity</a>, both <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be contained in the same replacement text.</p><p><a title="For interoperability" href="#dt-interop">For interoperability</a>, if a parameter-entity reference
686appears in a <a href="#NT-choice">choice</a>, <a href="#NT-seq">seq</a>, or <a href="#NT-Mixed">Mixed</a> construct, its replacement text <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> contain at
687least one non-blank character, and neither the first nor last non-blank character
688of the replacement text <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be a connector (<code>|</code> or <code>,</code>).</p></div><p>Examples of element-content models:</p><div class="exampleInner"><pre>&lt;!ELEMENT spec (front, body, back?)&gt;
689&lt;!ELEMENT div1 (head, (p | list | note)*, div2*)&gt;
690&lt;!ELEMENT dictionary-body (%div.mix; | %dict.mix;)*&gt;</pre></div></div><div class="div3"> <h4><a name="sec-mixed-content" id="sec-mixed-content" />3.2.2 Mixed Content</h4><p>[<a name="dt-mixed" id="dt-mixed" title="Mixed Content">Definition</a>: An element <a title="Start-Tag" href="#dt-stag">type</a>
691has <b>mixed content</b> when elements of that type <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> contain character
692data, optionally interspersed with <a title="Parent/Child" href="#dt-parentchild">child</a>
693elements.] In this case, the types of the child elements <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be constrained,
694but not their order or their number of occurrences:</p> <h5><a name="IDAUHCU" id="IDAUHCU" />Mixed-content Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-Mixed" id="NT-Mixed" />[51]   </td><td><code>Mixed</code></td><td>   ::=   </td><td><code>'(' <a href="#NT-S">S</a>? '#PCDATA' (<a href="#NT-S">S</a>?
695'|' <a href="#NT-S">S</a>? <a href="#NT-Name">Name</a>)* <a href="#NT-S">S</a>?
696')*' </code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| '(' <a href="#NT-S">S</a>? '#PCDATA' <a href="#NT-S">S</a>? ')' </code></td><td><a href="#vc-PEinGroup">[VC: Proper Group/PE Nesting]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#vc-MixedChildrenUnique">[VC: No Duplicate Types]</a></td></tr></tbody></table><p>where the <a href="#NT-Name">Name</a>s give the types of elements that
697may appear as children. The
698keyword <b>#PCDATA</b> derives historically from the term "parsed
699character data."</p><div class="constraint"><p class="prefix"><a name="vc-MixedChildrenUnique" id="vc-MixedChildrenUnique" /><b>Validity constraint: No Duplicate Types</b></p><p>The
700same name <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> appear more than once in a single mixed-content declaration.</p></div><p>Examples of mixed content declarations:</p><div class="exampleInner"><pre>&lt;!ELEMENT p (#PCDATA|a|ul|b|i|em)*&gt;
701&lt;!ELEMENT p (#PCDATA | %font; | %phrase; | %special; | %form;)* &gt;
702&lt;!ELEMENT b (#PCDATA)&gt;</pre></div></div></div><div class="div2"> <h3><a name="attdecls" id="attdecls" />3.3 Attribute-List Declarations</h3><p><a title="Attribute" href="#dt-attr">Attributes</a> are used to associate name-value
703pairs with <a title="Element" href="#dt-element">elements</a>. Attribute specifications
704<span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> appear outside of</span> <a title="Start-Tag" href="#dt-stag">start-tags</a> and <a title="empty-element tag" href="#dt-eetag">empty-element tags</a>; thus, the productions used to
705recognize them appear in <a href="#sec-starttags"><b>3.1 Start-Tags, End-Tags, and Empty-Element Tags</b></a>. Attribute-list declarations
706<em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be used:</p><ul><li><p>To define the set of attributes pertaining to a given element type.</p></li><li><p>To establish type constraints for these attributes.</p></li><li><p>To provide <a title="Attribute Default" href="#dt-default">default values</a> for
707attributes.</p></li></ul><p>[<a name="dt-attdecl" id="dt-attdecl" title="Attribute-List Declaration">Definition</a>: <b>Attribute-list
708declarations</b> specify the name, data type, and default value (if any)
709of each attribute associated with a given element type:]</p> <h5><a name="IDADMCU" id="IDADMCU" />Attribute-list Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-AttlistDecl" id="NT-AttlistDecl" />[52]   </td><td><code>AttlistDecl</code></td><td>   ::=   </td><td><code>'&lt;!ATTLIST' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-AttDef">AttDef</a>* <a href="#NT-S">S</a>? '&gt;'</code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-AttDef" id="NT-AttDef" />[53]   </td><td><code>AttDef</code></td><td>   ::=   </td><td><code><a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-S">S</a> <a href="#NT-AttType">AttType</a> <a href="#NT-S">S</a> <a href="#NT-DefaultDecl">DefaultDecl</a></code></td></tr></tbody></table><p>The <a href="#NT-Name">Name</a> in the <a href="#NT-AttlistDecl">AttlistDecl</a>
710rule is the type of an element. At user option, an XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> issue
711a warning if attributes are declared for an element type not itself declared,
712but this is not an error. The <a href="#NT-Name">Name</a> in the <a href="#NT-AttDef">AttDef</a>
713rule is the name of the attribute.</p><p>When more than one <a href="#NT-AttlistDecl">AttlistDecl</a> is provided
714for a given element type, the contents of all those provided are merged. When
715more than one definition is provided for the same attribute of a given element
716type, the first declaration is binding and later declarations are ignored. <a title="For interoperability" href="#dt-interop">For interoperability,</a> writers of DTDs <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> choose
717to provide at most one attribute-list declaration for a given element type,
718at most one attribute definition for a given attribute name in an attribute-list
719declaration, and at least one attribute definition in each attribute-list
720declaration. For interoperability, an XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> at user option
721issue a warning when more than one attribute-list declaration is provided
722for a given element type, or more than one attribute definition is provided
723for a given attribute, but this is not an error.</p><div class="div3"> <h4><a name="sec-attribute-types" id="sec-attribute-types" />3.3.1 Attribute Types</h4><p>XML attribute types are of three kinds: a string type, a set of tokenized
724types, and enumerated types. The string type may take any literal string as
725a value; the tokenized types have varying lexical and semantic constraints.
726The validity constraints noted in the grammar are applied after the attribute
727value has been normalized as described in <span><a href="#AVNormalize"><b>3.3.3 Attribute-Value Normalization</b></a></span>.</p> <h5><a name="IDAPPCU" id="IDAPPCU" />Attribute Types</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-AttType" id="NT-AttType" />[54]   </td><td><code>AttType</code></td><td>   ::=   </td><td><code><a href="#NT-StringType">StringType</a> | <a href="#NT-TokenizedType">TokenizedType</a>
728| <a href="#NT-EnumeratedType">EnumeratedType</a></code></td></tr><tr valign="baseline"><td><a name="NT-StringType" id="NT-StringType" />[55]   </td><td><code>StringType</code></td><td>   ::=   </td><td><code>'CDATA'</code></td></tr><tr valign="baseline"><td><a name="NT-TokenizedType" id="NT-TokenizedType" />[56]   </td><td><code>TokenizedType</code></td><td>   ::=   </td><td><code>'ID'</code></td><td><a href="#id">[VC: ID]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#one-id-per-el">[VC: One ID per Element Type]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#id-default">[VC: ID Attribute Default]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'IDREF'</code></td><td><a href="#idref">[VC: IDREF]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'IDREFS'</code></td><td><a href="#idref">[VC: IDREF]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'ENTITY'</code></td><td><a href="#entname">[VC: Entity Name]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'ENTITIES'</code></td><td><a href="#entname">[VC: Entity Name]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'NMTOKEN'</code></td><td><a href="#nmtok">[VC: Name Token]</a></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'NMTOKENS'</code></td><td><a href="#nmtok">[VC: Name Token]</a></td></tr></tbody></table><div class="constraint"><p class="prefix"><a name="id" id="id" /><b>Validity constraint: ID</b></p><p>Values of type <b>ID</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the <a href="#NT-Name">Name</a> production. A name <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> appear more than once
729in an XML document as a value of this type; i.e., ID values <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> uniquely
730identify the elements which bear them.</p></div><div class="constraint"><p class="prefix"><a name="one-id-per-el" id="one-id-per-el" /><b>Validity constraint: One ID per Element Type</b></p><p><span class="mustard">An element
731type <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> have more than one ID attribute specified.</p></div><div class="constraint"><p class="prefix"><a name="id-default" id="id-default" /><b>Validity constraint: ID Attribute Default</b></p><p>An ID attribute
732<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> have a declared default of <b>#IMPLIED</b> or <b>#REQUIRED</b>.</p></div><div class="constraint"><p class="prefix"><a name="idref" id="idref" /><b>Validity constraint: IDREF</b></p><p>Values of type <b>IDREF</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
733match the <a href="#NT-Name">Name</a> production, and values of type <b>IDREFS</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match <a href="#NT-Names">Names</a>; each <a href="#NT-Name">Name</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the value of an ID attribute on some element in the XML document;
734i.e. <b>IDREF</b> values <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the value of some ID attribute.</p></div><div class="constraint"><p class="prefix"><a name="entname" id="entname" /><b>Validity constraint: Entity Name</b></p><p>Values of type <b>ENTITY</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the <a href="#NT-Name">Name</a> production, values of type <b>ENTITIES</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match <a href="#NT-Names">Names</a>; each <a href="#NT-Name">Name</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the name of an <a title="Unparsed Entity" href="#dt-unparsed">unparsed entity</a>
735declared in the <a title="Document Type Declaration" href="#dt-doctype">DTD</a>.</p></div><div class="constraint"><p class="prefix"><a name="nmtok" id="nmtok" /><b>Validity constraint: Name Token</b></p><p>Values of type <b>NMTOKEN</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the <a href="#NT-Nmtoken">Nmtoken</a> production; values of type <b>NMTOKENS</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match <a href="#NT-Nmtokens">Nmtokens</a>.</p></div><p>[<a name="dt-enumerated" id="dt-enumerated" title="Enumerated Attribute&#xA;Values">Definition</a>: <b>Enumerated attributes</b> <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em></span> take one of a list of values
736provided in the declaration]. There are two kinds of enumerated types:</p> <h5><a name="IDAHXCU" id="IDAHXCU" />Enumerated Attribute Types</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-EnumeratedType" id="NT-EnumeratedType" />[57]   </td><td><code>EnumeratedType</code></td><td>   ::=   </td><td><code><a href="#NT-NotationType">NotationType</a>
737| <a href="#NT-Enumeration">Enumeration</a></code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-NotationType" id="NT-NotationType" />[58]   </td><td><code>NotationType</code></td><td>   ::=   </td><td><code>'NOTATION' <a href="#NT-S">S</a> '(' <a href="#NT-S">S</a>? <a href="#NT-Name">Name</a> (<a href="#NT-S">S</a>? '|' <a href="#NT-S">S</a>? <a href="#NT-Name">Name</a>)* <a href="#NT-S">S</a>? ')' </code></td><td><a href="#notatn">[VC: Notation Attributes]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#OneNotationPer">[VC: One Notation Per Element Type]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#NoNotationEmpty">[VC: No Notation on Empty Element]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#NoDuplicateTokens">[VC: No Duplicate
738Tokens]</a></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-Enumeration" id="NT-Enumeration" />[59]   </td><td><code>Enumeration</code></td><td>   ::=   </td><td><code>'(' <a href="#NT-S">S</a>? <a href="#NT-Nmtoken">Nmtoken</a>
739(<a href="#NT-S">S</a>? '|' <a href="#NT-S">S</a>? <a href="#NT-Nmtoken">Nmtoken</a>)* <a href="#NT-S">S</a>? ')'</code></td><td><a href="#enum">[VC: Enumeration]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#NoDuplicateTokens">[VC: No Duplicate
740Tokens]</a></td></tr></tbody></table><p>A <b>NOTATION</b> attribute identifies a <a title="Notation" href="#dt-notation">notation</a>,
741declared in the DTD with associated system and/or public identifiers, to be
742used in interpreting the element to which the attribute is attached.</p><div class="constraint"><p class="prefix"><a name="notatn" id="notatn" /><b>Validity constraint: Notation Attributes</b></p><p>Values of this type
743<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match one of the <a href="#Notations"><cite>notation</cite></a> names
744included in the declaration; all notation names in the declaration <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be
745declared.</p></div><div class="constraint"><p class="prefix"><a name="OneNotationPer" id="OneNotationPer" /><b>Validity constraint: One Notation Per Element Type</b></p><p><span class="mustard">An element type <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> have more than one <b>NOTATION</b>
746attribute specified.</p></div><div class="constraint"><p class="prefix"><a name="NoNotationEmpty" id="NoNotationEmpty" /><b>Validity constraint: No Notation on Empty Element</b></p><p><a title="For Compatibility" href="#dt-compat">For compatibility</a>,
747an attribute of type <b>NOTATION</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> be declared on an element
748declared <b>EMPTY</b>.</p></div><div class="constraint"><p class="prefix"><a name="NoDuplicateTokens" id="NoDuplicateTokens" /><b>Validity constraint: No Duplicate
749Tokens</b></p><p>The notation names in a single <a href="#NT-NotationType">NotationType</a>
750attribute declaration, as well as the <a href="#NT-Nmtoken">NmToken</a>s in a single
751<a href="#NT-Enumeration">Enumeration</a> attribute declaration, <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> all be distinct.</p></div><div class="constraint"><p class="prefix"><a name="enum" id="enum" /><b>Validity constraint: Enumeration</b></p><p>Values of this type <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match
752one of the <a href="#NT-Nmtoken">Nmtoken</a> tokens in the declaration.</p></div><p><a title="For interoperability" href="#dt-interop">For interoperability,</a> the same <a href="#NT-Nmtoken">Nmtoken</a> <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD NOT</em> occur more than once in the enumerated
753attribute types of a single element type.</p></div><div class="div3"> <h4><a name="sec-attr-defaults" id="sec-attr-defaults" />3.3.2 Attribute Defaults</h4><p>An <a title="Attribute-List Declaration" href="#dt-attdecl">attribute declaration</a> provides information
754on whether the attribute's presence is <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em>, and if not, how an XML processor
755<span>is
756to</span> react if a declared attribute is absent in a document.</p> <h5><a name="IDAR4CU" id="IDAR4CU" />Attribute Defaults</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-DefaultDecl" id="NT-DefaultDecl" />[60]   </td><td><code>DefaultDecl</code></td><td>   ::=   </td><td><code>'#REQUIRED' | '#IMPLIED' </code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| (('#FIXED' S)? <a href="#NT-AttValue">AttValue</a>)</code></td><td><a href="#RequiredAttr">[VC: Required Attribute]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#defattrvalid">[VC: Attribute
757Default Value Syntactically Correct]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#CleanAttrVals">[WFC: No &lt; in Attribute Values]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#FixedAttr">[VC: Fixed Attribute Default]</a></td></tr></tbody></table><p>In an attribute declaration, <b>#REQUIRED</b> means that the attribute
758<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> always be provided, <b>#IMPLIED</b> that no default value is provided.
759[<a name="dt-default" id="dt-default" title="Attribute Default">Definition</a>: If
760the declaration is neither <b>#REQUIRED</b> nor <b>#IMPLIED</b>, then
761the <a href="#NT-AttValue">AttValue</a> value contains the declared <b>default</b>
762value; the <b>#FIXED</b> keyword states that the attribute <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> always have
763the default value.
764When an XML processor encounters
765an <span>element
766without a specification for an attribute for which it has read a default
767value declaration, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> report the attribute with the declared default
768value to the application</span>.]</p><div class="constraint"><p class="prefix"><a name="RequiredAttr" id="RequiredAttr" /><b>Validity constraint: Required Attribute</b></p><p>If the default
769declaration is the keyword <b>#REQUIRED</b>, then the attribute <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be
770specified for all elements of the type in the attribute-list declaration.</p></div><div class="constraint"><p class="prefix"><a name="defattrvalid" id="defattrvalid" /><b>Validity constraint: <span>Attribute
771Default Value Syntactically Correct</span></b></p><p>The declared default value <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> meet the <span>syntactic</span>
772constraints of the declared attribute type.</p><p>Note that only the
773syntactic constraints of the type are required here; other constraints (e.g.
774that the value be the name of a declared unparsed entity, for an attribute of
775type ENTITY) may come into play if the declared default value is actually used
776(an element without a specification for this attribute occurs).</p></div><div class="constraint"><p class="prefix"><a name="FixedAttr" id="FixedAttr" /><b>Validity constraint: Fixed Attribute Default</b></p><p>If an attribute
777has a default value declared with the <b>#FIXED</b> keyword, instances of
778that attribute <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the default value.</p></div><p>Examples of attribute-list declarations:</p><div class="exampleInner"><pre>&lt;!ATTLIST termdef
779id ID #REQUIRED
780name CDATA #IMPLIED&gt;
781&lt;!ATTLIST list
782type (bullets|ordered|glossary) "ordered"&gt;
783&lt;!ATTLIST form
784method CDATA #FIXED "POST"&gt;</pre></div></div><div class="div3"> <h4><a name="AVNormalize" id="AVNormalize" />3.3.3 Attribute-Value Normalization</h4><p>Before the value of an attribute is passed to the application or checked
785for validity, the XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> normalize the attribute value by applying
786the algorithm below, or by using some other method such that the value passed
787to the application is the same as that produced by the algorithm.</p><ol type="1"><li><p>All line breaks <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> have been normalized on input to #xA as described
788in <a href="#sec-line-ends"><b>2.11 End-of-Line Handling</b></a>, so the rest of this algorithm operates
789on text normalized in this way.</p></li><li><p>Begin with a normalized value consisting of the empty string.</p></li><li><p>For each character, entity reference, or character reference in the
790unnormalized attribute value, beginning with the first and continuing to the
791last, do the following:</p><ul><li><p>For a character reference, append the referenced character to the
792normalized value.</p></li><li><p>For an entity reference, recursively apply step 3 of this algorithm
793to the replacement text of the entity.</p></li><li><p>For a white space character (#x20, #xD, #xA, #x9), append a space
794character (#x20) to the normalized value.</p></li><li><p>For another character, append the character to the normalized value.</p></li></ul></li></ol><p>If the attribute type is not CDATA, then the XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> further
795process the normalized attribute value by discarding any leading and trailing
796space (#x20) characters, and by replacing sequences of space (#x20) characters
797by a single space (#x20) character.</p><p>Note that if the unnormalized attribute value contains a character reference
798to a white space character other than space (#x20), the normalized value contains
799the referenced character itself (#xD, #xA or #x9). This contrasts with the
800case where the unnormalized value contains a white space character (not a
801reference), which is replaced with a space character (#x20) in the normalized
802value and also contrasts with the case where the unnormalized value contains
803an entity reference whose replacement text contains a white space character;
804being recursively processed, the white space character is replaced with a
805space character (#x20) in the normalized value.</p><p>All attributes for which no declaration has been read <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be treated
806by a non-validating processor as if declared <b>CDATA</b>.</p><p>It
807is an error if an
808<span><a title="Attribute Value" href="#dt-attrval">attribute
809value</a> contains a <a title="Entity Reference" href="#dt-entref">reference</a> to an
810entity for which no declaration has been read.</span></p><p>Following are examples of attribute normalization. Given the following
811declarations:</p><div class="exampleInner"><pre>&lt;!ENTITY d "&amp;#xD;"&gt;
812&lt;!ENTITY a "&amp;#xA;"&gt;
813&lt;!ENTITY da "&amp;#xD;&amp;#xA;"&gt;</pre></div><p>the attribute specifications in the left column below would be normalized
814to the character sequences of the middle column if the attribute <code>a</code>
815is declared <b>NMTOKENS</b> and to those of the right columns if <code>a</code>
816is declared <b>CDATA</b>.</p><table border="1" frame="border" summary="Attribute normalization summary"><thead><tr><th rowspan="1" colspan="1">Attribute specification</th><th rowspan="1" colspan="1">a is NMTOKENS</th><th rowspan="1" colspan="1">a is CDATA</th></tr></thead><tbody><tr><td rowspan="1" colspan="1"><div class="exampleInner"><pre>a="
817xyz"</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>x y z</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>#x20 #x20 x y z</pre></div></td></tr><tr><td rowspan="1" colspan="1"><div class="exampleInner"><pre>a="&amp;d;&amp;d;A&amp;a;<span>&amp;#x20;</span>&amp;a;B&amp;da;"</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>A #x20 B</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>#x20 #x20 A #x20 <span>#x20</span> #x20 B #x20 #x20</pre></div></td></tr><tr><td rowspan="1" colspan="1"><div class="exampleInner"><pre>a=
818"&amp;#xd;&amp;#xd;A&amp;#xa;&amp;#xa;B&amp;#xd;&amp;#xa;"</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>#xD #xD A #xA #xA B #xD #xA</pre></div></td><td rowspan="1" colspan="1"><div class="exampleInner"><pre>#xD #xD A #xA #xA B #xD #xA</pre></div></td></tr></tbody></table><p>Note that the last example is invalid (but well-formed) if <code>a</code>
819is declared to be of type <b>NMTOKENS</b>.</p></div></div><div class="div2"> <h3><a name="sec-condition-sect" id="sec-condition-sect" />3.4 Conditional Sections</h3><p>[<a name="dt-cond-section" id="dt-cond-section" title="conditional section">Definition</a>: <b>Conditional
820sections</b> are portions of the <a title="Document Type Declaration" href="#dt-doctype">document type
821declaration external subset</a> <span>or
822of external parameter entities </span>which are included in, or excluded from,
823the logical structure of the DTD based on the keyword which governs them.]</p> <h5><a name="IDAMHDU" id="IDAMHDU" />Conditional Section</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-conditionalSect" id="NT-conditionalSect" />[61]   </td><td><code>conditionalSect</code></td><td>   ::=   </td><td><code><a href="#NT-includeSect">includeSect</a> | <a href="#NT-ignoreSect">ignoreSect</a></code></td></tr><tr valign="baseline"><td><a name="NT-includeSect" id="NT-includeSect" />[62]   </td><td><code>includeSect</code></td><td>   ::=   </td><td><code>'&lt;![' S? 'INCLUDE' S? '[' <a href="#NT-extSubsetDecl">extSubsetDecl</a>
824']]&gt;' </code></td><td><a href="#condsec-nesting">[VC: Proper Conditional Section/PE Nesting]</a></td></tr><tr valign="baseline"><td><a name="NT-ignoreSect" id="NT-ignoreSect" />[63]   </td><td><code>ignoreSect</code></td><td>   ::=   </td><td><code>'&lt;![' S? 'IGNORE' S? '[' <a href="#NT-ignoreSectContents">ignoreSectContents</a>*
825']]&gt;'</code></td><td><a href="#condsec-nesting">[VC: Proper Conditional Section/PE Nesting]</a></td></tr><tr valign="baseline"><td><a name="NT-ignoreSectContents" id="NT-ignoreSectContents" />[64]   </td><td><code>ignoreSectContents</code></td><td>   ::=   </td><td><code><a href="#NT-Ignore">Ignore</a> ('&lt;![' <a href="#NT-ignoreSectContents">ignoreSectContents</a> ']]&gt;' <a href="#NT-Ignore">Ignore</a>)*</code></td></tr><tr valign="baseline"><td><a name="NT-Ignore" id="NT-Ignore" />[65]   </td><td><code>Ignore</code></td><td>   ::=   </td><td><code><a href="#NT-Char">Char</a>* - (<a href="#NT-Char">Char</a>*
826('&lt;![' | ']]&gt;') <a href="#NT-Char">Char</a>*) </code></td></tr></tbody></table><div class="constraint"><p class="prefix"><a name="condsec-nesting" id="condsec-nesting" /><b>Validity constraint: Proper Conditional Section/PE Nesting</b></p><p>If any of the "<code>&lt;![</code>",
827"<code>[</code>", or "<code>]]&gt;</code>" of a conditional section is contained
828in the replacement text for a parameter-entity reference, all of them <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
829be contained in the same replacement text.</p></div><p>Like the internal and external DTD subsets, a conditional section may contain
830one or more complete declarations, comments, processing instructions, or nested
831conditional sections, intermingled with white space.</p><p>If the keyword of the conditional section is <b>INCLUDE</b>, then the
832contents of the conditional section <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be considered</span> part of the DTD. If the keyword of
833the conditional section is <b>IGNORE</b>, then the contents of the conditional
834section <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be considered as</span> not logically part of the DTD.
835If a conditional section with a keyword of <b>INCLUDE</b> occurs within
836a larger conditional section with a keyword of <b>IGNORE</b>, both the outer
837and the inner conditional sections <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> ignored. The contents
838of an ignored conditional section <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> parsed by ignoring all characters after
839the "<code>[</code>" following the keyword, except conditional section starts
840"<code>&lt;![</code>" and ends "<code>]]&gt;</code>", until the matching conditional
841section end is found. Parameter entity references <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> be</span> recognized in this
842process.</p><p>If the keyword of the conditional section is a parameter-entity reference,
843the parameter entity <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be replaced by its content before the processor
844decides whether to include or ignore the conditional section.</p><p>An example:</p><div class="exampleInner"><pre>&lt;!ENTITY % draft 'INCLUDE' &gt;
845&lt;!ENTITY % final 'IGNORE' &gt;
846&lt;![%draft;[
847&lt;!ELEMENT book (comments*, title, body, supplements?)&gt;
848]]&gt;
849&lt;![%final;[
850&lt;!ELEMENT book (title, body, supplements?)&gt;
851]]&gt;</pre></div></div></div><div class="div1"> <h2><a name="sec-physical-struct" id="sec-physical-struct" />4 Physical Structures</h2><p>[<a name="dt-entity" id="dt-entity" title="Entity">Definition</a>: An XML document may consist of one
852or many storage units. These
853are called <b>entities</b>; they all have <b>content</b> and are
854all (except for the <a title="Document Entity" href="#dt-docent">document entity</a> and
855the <a title="Document Type Declaration" href="#dt-doctype">external DTD subset</a>) identified by
856entity <b>name</b>.] Each XML document has one entity
857called the <a title="Document Entity" href="#dt-docent">document entity</a>, which serves
858as the starting point for the <a title="XML Processor" href="#dt-xml-proc">XML processor</a>
859and may contain the whole document.</p><p>Entities may be either parsed or unparsed. [<a name="dt-parsedent" id="dt-parsedent" title="Text Entity">Definition</a>: The contents of a <b>parsed
860entity</b> are referred to as its <a title="Replacement Text" href="#dt-repltext">replacement
861text</a>; this <a title="Text" href="#dt-text">text</a> is considered an
862integral part of the document.]</p><p>[<a name="dt-unparsed" id="dt-unparsed" title="Unparsed Entity">Definition</a>: An <b>unparsed entity</b>
863is a resource whose contents may or may not be <a title="Text" href="#dt-text">text</a>,
864and if text, may
865be other than XML. Each unparsed entity has an associated <a title="Notation" href="#dt-notation">notation</a>, identified by name. Beyond a requirement
866that an XML processor make the identifiers for the entity and notation available
867to the application, XML places no constraints on the contents of unparsed
868entities.]</p><p>Parsed entities are invoked by name using entity references; unparsed entities
869by name, given in the value of <b>ENTITY</b> or <b>ENTITIES</b> attributes.</p><p>[<a name="gen-entity" id="gen-entity" title="general entity">Definition</a>: <b>General entities</b>
870are entities for use within the document content. In this specification, general
871entities are sometimes referred to with the unqualified term <em>entity</em>
872when this leads to no ambiguity.] [<a name="dt-PE" id="dt-PE" title="Parameter entity">Definition</a>: <b>Parameter
873entities</b> are parsed entities for use within the DTD.]
874These two types of entities use different forms of reference and are recognized
875in different contexts. Furthermore, they occupy different namespaces; a parameter
876entity and a general entity with the same name are two distinct entities.</p><div class="div2"> <h3><a name="sec-references" id="sec-references" />4.1 Character and Entity References</h3><p>[<a name="dt-charref" id="dt-charref" title="Character Reference">Definition</a>: A <b>character
877reference</b> refers to a specific character in the ISO/IEC 10646 character
878set, for example one not directly accessible from available input devices.]</p> <h5><a name="IDAFYDU" id="IDAFYDU" />Character Reference</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-CharRef" id="NT-CharRef" />[66]   </td><td><code>CharRef</code></td><td>   ::=   </td><td><code>'&amp;#' [0-9]+ ';' </code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| '&amp;#x' [0-9a-fA-F]+ ';'</code></td><td><a href="#wf-Legalchar">[WFC: Legal Character]</a></td></tr></tbody></table><div class="constraint"><p class="prefix"><a name="wf-Legalchar" id="wf-Legalchar" /><b>Well-formedness constraint: Legal Character</b></p><p>Characters referred
879to using character references <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the production for <a href="#NT-Char">Char</a>.</p></div><p>If the character reference begins with "<code>&amp;#x</code>",
880the digits and letters up to the terminating <code>;</code> provide a hexadecimal
881representation of the character's code point in ISO/IEC 10646. If it begins
882just with "<code>&amp;#</code>", the digits up to the terminating <code>;</code>
883provide a decimal representation of the character's code point.</p><p>[<a name="dt-entref" id="dt-entref" title="Entity Reference">Definition</a>: An <b>entity reference</b>
884refers to the content of a named entity.] [<a name="dt-GERef" id="dt-GERef" title="General Entity Reference">Definition</a>: References to parsed general entities use
885ampersand (<code>&amp;</code>) and semicolon (<code>;</code>) as delimiters.] [<a name="dt-PERef" id="dt-PERef" title="Parameter-entity reference">Definition</a>: <b>Parameter-entity references</b>
886use percent-sign (<code>%</code>) and semicolon (<code>;</code>) as delimiters.]</p> <h5><a name="IDAS0DU" id="IDAS0DU" />Entity Reference</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-Reference" id="NT-Reference" />[67]   </td><td><code>Reference</code></td><td>   ::=   </td><td><code><a href="#NT-EntityRef">EntityRef</a> | <a href="#NT-CharRef">CharRef</a></code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-EntityRef" id="NT-EntityRef" />[68]   </td><td><code>EntityRef</code></td><td>   ::=   </td><td><code>'&amp;' <a href="#NT-Name">Name</a> ';'</code></td><td><a href="#wf-entdeclared">[WFC: Entity Declared]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#vc-entdeclared">[VC: Entity Declared]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#textent">[WFC: Parsed Entity]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#norecursion">[WFC: No Recursion]</a></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-PEReference" id="NT-PEReference" />[69]   </td><td><code>PEReference</code></td><td>   ::=   </td><td><code>'%' <a href="#NT-Name">Name</a> ';'</code></td><td><a href="#vc-entdeclared">[VC: Entity Declared]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#norecursion">[WFC: No Recursion]</a></td></tr><tr valign="baseline"><td /><td /><td /><td /><td><a href="#indtd">[WFC: In DTD]</a></td></tr></tbody></table><div class="constraint"><p class="prefix"><a name="wf-entdeclared" id="wf-entdeclared" /><b>Well-formedness constraint: Entity Declared</b></p><p>In a document
887without any DTD, a document with only an internal DTD subset which contains
888no parameter entity references, or a document with "<code>standalone='yes'</code>", for
889an entity reference that does not occur within the external subset or a parameter
890entity, the <a href="#NT-Name">Name</a> given in the entity reference <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> <a title="match" href="#dt-match">match</a> that in an <a href="#sec-entity-decl"><cite>entity
891declaration</cite></a> that does not occur within the external subset or a
892parameter entity, except that well-formed documents need not declare
893any of the following entities: <code>amp</code>,
894<code>lt</code>,
895<code>gt</code>,
896<code>apos</code>,
897<code>quot</code>. The
898declaration of a general entity <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> precede any reference to it which appears
899in a default value in an attribute-list declaration.</p><p><span>Note
900that non-validating processors are <a href="#include-if-valid"><cite>not
901obligated to</cite></a> to read and process entity declarations occurring in parameter entities or in
902the external subset</span>; for such documents,
903the rule that an entity must be declared is a well-formedness constraint only
904if <a href="#sec-rmd"><cite>standalone='yes'</cite></a>.</p></div><div class="constraint"><p class="prefix"><a name="vc-entdeclared" id="vc-entdeclared" /><b>Validity constraint: Entity Declared</b></p><p>In a document with
905an external subset or external parameter entities with "<code>standalone='no'</code>",
906the <a href="#NT-Name">Name</a> given in the entity reference <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> <a title="match" href="#dt-match">match</a> that in an <a href="#sec-entity-decl"><cite>entity
907declaration</cite></a>. For interoperability, valid documents <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> declare
908the entities <code>amp</code>,
909<code>lt</code>,
910<code>gt</code>,
911<code>apos</code>,
912<code>quot</code>, in the form specified in <a href="#sec-predefined-ent"><b>4.6 Predefined Entities</b></a>.
913The declaration of a parameter entity <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> precede any reference to it. Similarly,
914the declaration of a general entity <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> precede any attribute-list
915declaration containing a default value with a direct or indirect reference
916to that general entity.</p></div><div class="constraint"><p class="prefix"><a name="textent" id="textent" /><b>Well-formedness constraint: Parsed Entity</b></p><p>An entity reference <em class="rfc2119" title="Keyword in RFC 2119 context">MUST
917NOT</em> contain the name of an <a title="Unparsed Entity" href="#dt-unparsed">unparsed entity</a>.
918Unparsed entities may be referred to only in <a title="Attribute Value" href="#dt-attrval">attribute
919values</a> declared to be of type <b>ENTITY</b> or <b>ENTITIES</b>.</p></div><div class="constraint"><p class="prefix"><a name="norecursion" id="norecursion" /><b>Well-formedness constraint: No Recursion</b></p><p>A parsed entity <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> contain a recursive reference to itself, either directly or indirectly.</p></div><div class="constraint"><p class="prefix"><a name="indtd" id="indtd" /><b>Well-formedness constraint: In DTD</b></p><p>Parameter-entity references <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> appear outside</span>
920the <a title="Document Type Declaration" href="#dt-doctype">DTD</a>.</p></div><p>Examples of character and entity references:</p><div class="exampleInner"><pre>Type &lt;key&gt;less-than&lt;/key&gt; (&amp;#x3C;) to save options.
921This document was prepared on &amp;docdate; and
922is classified &amp;security-level;.</pre></div><p>Example of a parameter-entity reference:</p><div class="exampleInner"><pre>&lt;!-- declare the parameter entity "ISOLat2"... --&gt;
923&lt;!ENTITY % ISOLat2
924SYSTEM "http://www.xml.com/iso/isolat2-xml.entities" &gt;
925&lt;!-- ... now reference it. --&gt;
926%ISOLat2;</pre></div></div><div class="div2"> <h3><a name="sec-entity-decl" id="sec-entity-decl" />4.2 Entity Declarations</h3><p>[<a name="dt-entdecl" id="dt-entdecl" title="entity declaration">Definition</a>: Entities are declared
927thus:]</p> <h5><a name="IDAECEU" id="IDAECEU" />Entity Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-EntityDecl" id="NT-EntityDecl" />[70]   </td><td><code>EntityDecl</code></td><td>   ::=   </td><td><code><a href="#NT-GEDecl">GEDecl</a> | <a href="#NT-PEDecl">PEDecl</a></code></td></tr><tr valign="baseline"><td><a name="NT-GEDecl" id="NT-GEDecl" />[71]   </td><td><code>GEDecl</code></td><td>   ::=   </td><td><code>'&lt;!ENTITY' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-S">S</a> <a href="#NT-EntityDef">EntityDef</a> <a href="#NT-S">S</a>?
928'&gt;'</code></td></tr><tr valign="baseline"><td><a name="NT-PEDecl" id="NT-PEDecl" />[72]   </td><td><code>PEDecl</code></td><td>   ::=   </td><td><code>'&lt;!ENTITY' <a href="#NT-S">S</a> '%' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-S">S</a> <a href="#NT-PEDef">PEDef</a> <a href="#NT-S">S</a>? '&gt;'</code></td></tr><tr valign="baseline"><td><a name="NT-EntityDef" id="NT-EntityDef" />[73]   </td><td><code>EntityDef</code></td><td>   ::=   </td><td><code><a href="#NT-EntityValue">EntityValue</a>| (<a href="#NT-ExternalID">ExternalID</a> <a href="#NT-NDataDecl">NDataDecl</a>?)</code></td></tr><tr valign="baseline"><td><a name="NT-PEDef" id="NT-PEDef" />[74]   </td><td><code>PEDef</code></td><td>   ::=   </td><td><code><a href="#NT-EntityValue">EntityValue</a> | <a href="#NT-ExternalID">ExternalID</a></code></td></tr></tbody></table><p>The <a href="#NT-Name">Name</a> identifies the entity in an <a title="Entity Reference" href="#dt-entref">entity
929reference</a> or, in the case of an unparsed entity, in the value of
930an <b>ENTITY</b> or <b>ENTITIES</b> attribute. If the same entity is declared
931more than once, the first declaration encountered is binding; at user option,
932an XML processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> issue a warning if entities are declared multiple times.</p><div class="div3"> <h4><a name="sec-internal-ent" id="sec-internal-ent" />4.2.1 Internal Entities</h4><p>[<a name="dt-internent" id="dt-internent" title="Internal Entity Replacement Text">Definition</a>: If the
933entity definition is an <a href="#NT-EntityValue">EntityValue</a>, the defined
934entity is called an <b>internal entity</b>. There is no separate physical
935storage object, and the content of the entity is given in the declaration.]
936Note that some processing of entity and character references in the <a title="Literal Entity Value" href="#dt-litentval">literal entity value</a> may be required to produce
937the correct <a title="Replacement Text" href="#dt-repltext">replacement text</a>: see <a href="#intern-replacement"><b>4.5 Construction of Entity Replacement Text</b></a>.</p><p>An internal entity is a <a title="Text Entity" href="#dt-parsedent">parsed entity</a>.</p><p>Example of an internal entity declaration:</p><div class="exampleInner"><pre>&lt;!ENTITY Pub-Status "This is a pre-release of the
938specification."&gt;</pre></div></div><div class="div3"> <h4><a name="sec-external-ent" id="sec-external-ent" />4.2.2 External Entities</h4><p>[<a name="dt-extent" id="dt-extent" title="External Entity">Definition</a>: If the entity is not internal,
939it is an <b>external entity</b>, declared as follows:]</p> <h5><a name="IDAUIEU" id="IDAUIEU" />External Entity Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-ExternalID" id="NT-ExternalID" />[75]   </td><td><code>ExternalID</code></td><td>   ::=   </td><td><code>'SYSTEM' <a href="#NT-S">S</a> <a href="#NT-SystemLiteral">SystemLiteral</a></code></td></tr><tr valign="baseline"><td /><td /><td /><td><code>| 'PUBLIC' <a href="#NT-S">S</a> <a href="#NT-PubidLiteral">PubidLiteral</a> <a href="#NT-S">S</a> <a href="#NT-SystemLiteral">SystemLiteral</a></code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-NDataDecl" id="NT-NDataDecl" />[76]   </td><td><code>NDataDecl</code></td><td>   ::=   </td><td><code><a href="#NT-S">S</a> 'NDATA' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a></code></td><td><a href="#not-declared">[VC: Notation Declared]</a></td></tr></tbody></table><p>If the <a href="#NT-NDataDecl">NDataDecl</a> is present, this is a general <a title="Unparsed Entity" href="#dt-unparsed">unparsed entity</a>; otherwise it is a parsed entity.</p><div class="constraint"><p class="prefix"><a name="not-declared" id="not-declared" /><b>Validity constraint: Notation Declared</b></p><p>The <a href="#NT-Name">Name</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> match the declared name of a <a title="Notation" href="#dt-notation">notation</a>.</p></div><p>[<a name="dt-sysid" id="dt-sysid" title="System Identifier">Definition</a>: The <a href="#NT-SystemLiteral">SystemLiteral</a> is called the entity's <b>system
940identifier</b>. It is <span>meant to be
941converted to</span> a URI reference
942(as defined in <a href="#rfc2396">[IETF RFC 2396]</a>, updated by <a href="#rfc2732">[IETF RFC 2732]</a>),
943<span>as part of the
944process of dereferencing it</span> to obtain input for the XML processor to construct the
945entity's replacement text.] It is an error for a fragment identifier
946(beginning with a <code>#</code> character) to be part of a system identifier.
947Unless otherwise provided by information outside the scope of this specification
948(e.g. a special XML element type defined by a particular DTD, or a processing
949instruction defined by a particular application specification), relative URIs
950are relative to the location of the resource within which the entity declaration
951occurs. <span>This is defined to
952be the external entity containing the '&lt;' which starts the declaration, at the
953point when it is parsed as a declaration.</span>
954A URI might thus be relative to the <a title="Document Entity" href="#dt-docent">document
955entity</a>, to the entity containing the <a title="Document Type Declaration" href="#dt-doctype">external
956DTD subset</a>, or to some other <a title="External Entity" href="#dt-extent">external parameter
957entity</a>. <span>Attempts to
958retrieve the resource identified by a URI <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be redirected at the parser
959level (for example, in an entity resolver) or below (at the protocol level,
960for example, via an HTTP <code>Location:</code> header). In the absence of additional
961information outside the scope of this specification within the resource,
962the base URI of a resource is always the URI of the actual resource returned.
963In other words, it is the URI of the resource retrieved after all redirection
964has occurred.</span></p><p>System
965identifiers (and other XML strings meant to be used as URI references) <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> contain
966characters that, according to <a href="#rfc2396">[IETF RFC 2396]</a> and <a href="#rfc2732">[IETF RFC 2732]</a>,
967must be escaped before a URI can be used to retrieve the referenced resource. The
968characters to be escaped are the control characters #x0 to #x1F and #x7F (most of
969which cannot appear in XML), space #x20, the delimiters '&lt;' #x3C, '&gt;' #x3E and
970'"' #x22, the <em>unwise</em> characters '{' #x7B, '}' #x7D, '|' #x7C, '\' #x5C, '^' #x5E and
971'`' #x60, as well as all characters above #x7F. Since escaping is not always a fully
972reversible process, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be performed only when absolutely necessary and as late
973as possible in a processing chain. In particular, neither the process of converting
974a relative URI to an absolute one nor the process of passing a URI reference to a
975process or software component responsible for dereferencing it <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> trigger escaping.
976When escaping does occur, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be performed as follows:</p><ol type="1"><li><p>Each
977character <span>to be escaped</span>
978is <span>represented in</span>
979UTF-8 <span><a href="#Unicode">[Unicode]</a></span>
980as one or more bytes.</p></li><li><p><span>The resulting bytes</span>
981are escaped with
982the URI escaping mechanism (that is, converted to <code>%</code><var>HH</var>,
983where HH is the hexadecimal notation of the byte value).</p></li><li><p>The original character is replaced by the resulting character sequence.</p></li></ol><p>[<a name="dt-pubid" id="dt-pubid" title="Public identifier">Definition</a>: In addition to a system
984identifier, an external identifier <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> include a <b>public identifier</b>.]
985An XML processor attempting to retrieve the entity's content <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> use
986<span>any combination of
987the public and system identifiers as well as additional information outside the
988scope of this specification</span> to try to generate an alternative URI reference.
989If the processor is unable to do so, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> use the URI
990reference specified in the system literal. Before a match is attempted,
991all strings of white space in the public identifier <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be normalized to
992single space characters (#x20), and leading and trailing white space <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
993be removed.</p><p>Examples of external entity declarations:</p><div class="exampleInner"><pre>&lt;!ENTITY open-hatch
994SYSTEM "http://www.textuality.com/boilerplate/OpenHatch.xml"&gt;
995&lt;!ENTITY open-hatch
996PUBLIC "-//Textuality//TEXT Standard open-hatch boilerplate//EN"
997"http://www.textuality.com/boilerplate/OpenHatch.xml"&gt;
998&lt;!ENTITY hatch-pic
999SYSTEM "../grafix/OpenHatch.gif"
1000NDATA gif &gt;</pre></div></div></div><div class="div2"> <h3><a name="TextEntities" id="TextEntities" />4.3 Parsed Entities</h3><div class="div3"> <h4><a name="sec-TextDecl" id="sec-TextDecl" />4.3.1 The Text Declaration</h4><p>External parsed entities <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> each begin with a <b>text declaration</b>.</p> <h5><a name="IDAUPEU" id="IDAUPEU" />Text Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-TextDecl" id="NT-TextDecl" />[77]   </td><td><code>TextDecl</code></td><td>   ::=   </td><td><code>'&lt;?xml' <a href="#NT-VersionInfo">VersionInfo</a>? <a href="#NT-EncodingDecl">EncodingDecl</a> <a href="#NT-S">S</a>? '?&gt;'</code></td></tr></tbody></table><p>The text declaration <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be provided literally, not by reference
1001to a parsed entity. <span class="mustard">The</span> text declaration
1002<span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> appear at any
1003position other than the beginning of an external parsed entity. The text declaration
1004in an external parsed entity is not considered part of its <a title="Replacement Text" href="#dt-repltext">replacement text</a>.</p></div><div class="div3"> <h4><a name="wf-entities" id="wf-entities" />4.3.2 Well-Formed Parsed Entities</h4><p>The document entity is well-formed if it matches the production labeled <a href="#NT-document">document</a>. An external general parsed entity is well-formed
1005if it matches the production labeled <a href="#NT-extParsedEnt">extParsedEnt</a>. All
1006external parameter entities are well-formed by definition.</p> <h5><a name="IDA2REU" id="IDA2REU" />Well-Formed External Parsed Entity</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-extParsedEnt" id="NT-extParsedEnt" />[78]   </td><td><code>extParsedEnt</code></td><td>   ::=   </td><td><code><a href="#NT-TextDecl">TextDecl</a>? <a href="#NT-content">content</a> - <a href="#NT-Char">Char</a>* <a href="#NT-RestrictedChar">RestrictedChar</a> <a href="#NT-Char">Char</a>*</code></td></tr></tbody></table><p>An internal general parsed entity is well-formed if its replacement text
1007matches the production labeled <a href="#NT-content">content</a>. All internal
1008parameter entities are well-formed by definition.</p><p>A consequence of well-formedness in <span>general</span>
1009entities is that the logical and physical
1010structures in an XML document are properly nested; no <a title="Start-Tag" href="#dt-stag">start-tag</a>, <a title="End Tag" href="#dt-etag">end-tag</a>, <a title="Empty" href="#dt-empty">empty-element tag</a>, <a title="Element" href="#dt-element">element</a>, <a title="Comment" href="#dt-comment">comment</a>, <a title="Processing instruction" href="#dt-pi">processing instruction</a>, <a title="Character Reference" href="#dt-charref">character
1011reference</a>, or <a title="Entity Reference" href="#dt-entref">entity reference</a>
1012can begin in one entity and end in another.</p></div><div class="div3"> <h4><a name="charencoding" id="charencoding" />4.3.3 Character Encoding in Entities</h4><p>Each external parsed entity in an XML document <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> use a different encoding
1013for its characters. All XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be able to read entities in both
1014the UTF-8 and UTF-16 encodings. The terms "UTF-8"
1015and "UTF-16" in this specification do not apply to character
1016encodings with any other labels, even if the encodings or labels are very
1017similar to UTF-8 or UTF-16.</p><p>Entities encoded in UTF-16 <span><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em></span> <span>and entities
1018encoded in UTF-8 <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em></span> begin with the Byte Order Mark described in
1019ISO/IEC 10646 <a href="#ISO10646">[ISO/IEC 10646]</a> or Unicode <a href="#Unicode">[Unicode]</a>
1020(the ZERO WIDTH NO-BREAK SPACE character, #xFEFF). This is an encoding signature,
1021not part of either the markup or the character data of the XML document. XML
1022processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be able to use this character to differentiate between UTF-8
1023and UTF-16 encoded documents.</p><p>Although an XML processor is required to read only entities in the UTF-8
1024and UTF-16 encodings, it is recognized that other encodings are used around
1025the world, and it may be desired for XML processors to read entities that
1026use them. In
1027the absence of external character encoding information (such as MIME headers),
1028parsed entities which are stored in an encoding other than UTF-8 or UTF-16
1029<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> begin with a text declaration (see <a href="#sec-TextDecl"><b>4.3.1 The Text Declaration</b></a>) containing
1030an encoding declaration:</p> <h5><a name="IDARVEU" id="IDARVEU" />Encoding Declaration</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-EncodingDecl" id="NT-EncodingDecl" />[80]   </td><td><code>EncodingDecl</code></td><td>   ::=   </td><td><code><a href="#NT-S">S</a> 'encoding' <a href="#NT-Eq">Eq</a>
1031('"' <a href="#NT-EncName">EncName</a> '"' | "'" <a href="#NT-EncName">EncName</a>
1032"'" ) </code></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-EncName" id="NT-EncName" />[81]   </td><td><code>EncName</code></td><td>   ::=   </td><td><code>[A-Za-z] ([A-Za-z0-9._] | '-')*</code></td><td><i>/* Encoding
1033name contains only Latin characters */</i></td></tr></tbody></table><p>In the <a title="Document Entity" href="#dt-docent">document entity</a>, the encoding
1034declaration is part of the <a title="XML Declaration" href="#dt-xmldecl">XML declaration</a>.
1035The <a href="#NT-EncName">EncName</a> is the name of the encoding used.</p><p>In an encoding declaration, the values "<code>UTF-8</code>", "<code>UTF-16</code>",
1036"<code>ISO-10646-UCS-2</code>", and "<code>ISO-10646-UCS-4</code>" <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be used
1037for the various encodings and transformations of Unicode / ISO/IEC 10646,
1038the values "<code>ISO-8859-1</code>", "<code>ISO-8859-2</code>",
1039... "<code>ISO-8859-</code><var>n</var>" (where <var>n</var>
1040is the part number) <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be used for the parts of ISO 8859, and
1041the values "<code>ISO-2022-JP</code>", "<code>Shift_JIS</code>",
1042and "<code>EUC-JP</code>" <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> be used for the various encoded
1043forms of JIS X-0208-1997. It
1044is <em class="rfc2119" title="Keyword in RFC 2119 context">RECOMMENDED</em> that character encodings registered (as <em>charset</em>s)
1045with the Internet Assigned Numbers Authority <a href="#IANA">[IANA-CHARSETS]</a>,
1046other than those just listed, be referred to using their registered names;
1047other encodings <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> use names starting with an "x-" prefix.
1048XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> match character encoding names in a case-insensitive
1049way and <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> either interpret an IANA-registered name as the encoding registered
1050at IANA for that name or treat it as unknown (processors are, of course, not
1051required to support all IANA-registered encodings).</p><p>In the absence of information provided by an external transport protocol
1052(e.g. HTTP or MIME), it is a <a title="Fatal Error" href="#dt-fatal">fatal error</a> for
1053an entity including an encoding declaration to be presented to the XML processor
1054in an encoding other than that named in the declaration, or for an entity which
1055begins with neither a Byte Order Mark
1056nor an encoding declaration to use an encoding other than UTF-8. Note that
1057since ASCII is a subset of UTF-8, ordinary ASCII entities do not strictly
1058need an encoding declaration.</p><p>It is a <a title="Fatal Error" href="#dt-fatal">fatal error</a> for a <a href="#NT-TextDecl">TextDecl</a> to occur other
1059than at the beginning of an external entity.</p><p>It is a <a title="Fatal Error" href="#dt-fatal">fatal error</a> when an XML processor
1060encounters an entity with an encoding that it is unable to process. It
1061is a <a title="Fatal Error" href="#dt-fatal">fatal error</a> if an XML entity is determined (via default, encoding declaration,
1062or higher-level protocol) to be in a certain encoding but contains <span>byte</span>
1063sequences that are not legal in that encoding. <span>Specifically, it is a
1064fatal error if an entity encoded in UTF-8 contains any irregular code unit sequences,
1065as defined in Unicode <a href="#Unicode">[Unicode]</a>.</span> <span>Unless an encoding
1066is determined by a higher-level protocol, </span>it is also a <a title="Fatal Error" href="#dt-fatal">fatal error</a> if an XML entity
1067contains no encoding declaration and its content is not legal UTF-8 or UTF-16.</p><p>Examples of text declarations containing encoding declarations:</p><div class="exampleInner"><pre>&lt;?xml encoding='UTF-8'?&gt;
1068&lt;?xml encoding='EUC-JP'?&gt;</pre></div></div><div class="div3"> <h4><a name="sec-version-info" id="sec-version-info" />4.3.4 Version Information in Entities</h4><p>Each entity, including the <a title="Document Entity" href="#dt-docent">document entity</a>,
1069can be separately
1070declared as XML 1.0 or XML 1.1. The version declaration appearing
1071in the document entity determines the version of the document as a
1072whole. An XML 1.1 document may invoke XML 1.0 external entities, so
1073that otherwise duplicated versions of external entities,
1074particularly DTD external subsets, need not be maintained. However,
1075in such a case the rules of XML 1.1 are applied to the entire
1076document.</p><p> If an entity (including the document entity) is not labeled
1077with a version number, it is treated as if labeled as version
10781.0.</p></div></div><div class="div2"> <h3><a name="entproc" id="entproc" />4.4 XML Processor Treatment of Entities and References</h3><p>The table below summarizes the contexts in which character references,
1079entity references, and invocations of unparsed entities might appear and the
1080<em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em> behavior of an <a title="XML Processor" href="#dt-xml-proc">XML processor</a>
1081in each case. The labels in the leftmost column describe the recognition context: </p><dl><dt class="label">Reference in Content</dt><dd><p>as a reference anywhere after the <a title="Start-Tag" href="#dt-stag">start-tag</a>
1082and before the <a title="End Tag" href="#dt-etag">end-tag</a> of an element; corresponds
1083to the nonterminal <a href="#NT-content">content</a>.</p></dd><dt class="label">Reference in Attribute Value</dt><dd><p>as a reference within either the value of an attribute in a <a title="Start-Tag" href="#dt-stag">start-tag</a>,
1084or a default value in an <a title="Attribute-List Declaration" href="#dt-attdecl">attribute declaration</a>;
1085corresponds to the nonterminal <a href="#NT-AttValue">AttValue</a>.</p></dd><dt class="label">Occurs as Attribute Value</dt><dd><p>as a <a href="#NT-Name">Name</a>, not a reference, appearing either as
1086the value of an attribute which has been declared as type <b>ENTITY</b>,
1087or as one of the space-separated tokens in the value of an attribute which
1088has been declared as type <b>ENTITIES</b>.</p></dd><dt class="label">Reference in Entity Value</dt><dd><p>as a reference within a parameter or internal entity's <a title="Literal Entity Value" href="#dt-litentval">literal
1089entity value</a> in the entity's declaration; corresponds to the nonterminal <a href="#NT-EntityValue">EntityValue</a>.</p></dd><dt class="label">Reference in DTD</dt><dd><p>as a reference within either the internal or external subsets of the <a title="Document Type Declaration" href="#dt-doctype">DTD</a>, but outside of an <a href="#NT-EntityValue">EntityValue</a>, <a href="#NT-AttValue">AttValue</a>, <a href="#NT-PI">PI</a>, <a href="#NT-Comment">Comment</a>, <a href="#NT-SystemLiteral">SystemLiteral</a>, <a href="#NT-PubidLiteral">PubidLiteral</a>,
1090or the contents of an ignored conditional section (see <a href="#sec-condition-sect"><b>3.4 Conditional Sections</b></a>).</p><p>.</p></dd></dl><p></p><table border="1" frame="border" cellpadding="7" summary="Entity type/reference matrix"><tbody align="center"><tr><td rowspan="2" colspan="1"></td><td colspan="4" align="center" valign="bottom" rowspan="1">Entity
1091Type</td><td rowspan="2" align="center" colspan="1">Character</td></tr><tr align="center" valign="bottom"><td rowspan="1" colspan="1">Parameter</td><td rowspan="1" colspan="1">Internal General</td><td rowspan="1" colspan="1">External Parsed
1092General</td><td rowspan="1" colspan="1">Unparsed</td></tr><tr align="center" valign="middle"><td align="right" rowspan="1" colspan="1">Reference
1093in Content</td><td rowspan="1" colspan="1"><a href="#not-recognized"><cite>Not recognized</cite></a></td><td rowspan="1" colspan="1"><a href="#included"><cite>Included</cite></a></td><td rowspan="1" colspan="1"><a href="#include-if-valid"><cite>Included
1094if validating</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#included"><cite>Included</cite></a></td></tr><tr align="center" valign="middle"><td align="right" rowspan="1" colspan="1">Reference in Attribute Value</td><td rowspan="1" colspan="1"><a href="#not-recognized"><cite>Not recognized</cite></a></td><td rowspan="1" colspan="1"><a href="#inliteral"><cite>Included
1095in literal</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#included"><cite>Included</cite></a></td></tr><tr align="center" valign="middle"><td align="right" rowspan="1" colspan="1">Occurs as Attribute
1096Value</td><td rowspan="1" colspan="1"><a href="#not-recognized"><cite>Not recognized</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#notify"><cite>Notify</cite></a></td><td rowspan="1" colspan="1"><a href="#not-recognized"><cite>Not recognized</cite></a></td></tr><tr align="center" valign="middle"><td align="right" rowspan="1" colspan="1">Reference in EntityValue</td><td rowspan="1" colspan="1"><a href="#inliteral"><cite>Included in literal</cite></a></td><td rowspan="1" colspan="1"><a href="#bypass"><cite>Bypassed</cite></a></td><td rowspan="1" colspan="1"><a href="#bypass"><cite>Bypassed</cite></a></td><td rowspan="1" colspan="1"><a href="#error"><cite><span>Error</span></cite></a></td><td rowspan="1" colspan="1"><a href="#included"><cite>Included</cite></a></td></tr><tr align="center" valign="middle"><td align="right" rowspan="1" colspan="1">Reference in DTD</td><td rowspan="1" colspan="1"><a href="#as-PE"><cite>Included as PE</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td><td rowspan="1" colspan="1"><a href="#forbidden"><cite>Forbidden</cite></a></td></tr></tbody></table><div class="div3"> <h4><a name="not-recognized" id="not-recognized" />4.4.1 Not Recognized</h4><p>Outside the DTD, the <code>%</code> character has no special significance;
1097thus, what would be parameter entity references in the DTD are not recognized
1098as markup in <a href="#NT-content">content</a>. Similarly, the names of unparsed
1099entities are not recognized except when they appear in the value of an appropriately
1100declared attribute.</p></div><div class="div3"> <h4><a name="included" id="included" />4.4.2 Included</h4><p>[<a name="dt-include" id="dt-include" title="Include">Definition</a>: An entity is <b>included</b>
1101when its <a title="Replacement Text" href="#dt-repltext">replacement text</a> is retrieved
1102and processed, in place of the reference itself, as though it were part of
1103the document at the location the reference was recognized.] The replacement
1104text <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> contain both <a title="Character Data" href="#dt-chardata">character data</a>
1105and (except for parameter entities) <a title="Markup" href="#dt-markup">markup</a>,
1106which <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be recognized in the usual way. (The string "<code>AT&amp;amp;T;</code>"
1107expands to "<code>AT&amp;T;</code>" and the remaining ampersand
1108is not recognized as an entity-reference delimiter.) A character reference
1109is <b>included</b> when the indicated character is processed in place
1110of the reference itself. </p></div><div class="div3"> <h4><a name="include-if-valid" id="include-if-valid" />4.4.3 Included If Validating</h4><p>When an XML processor recognizes a reference to a parsed entity, in order
1111to <a title="Validity" href="#dt-valid">validate</a> the document, the processor
1112<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> <a title="Include" href="#dt-include">include</a> its replacement text. If
1113the entity is external, and the processor is not attempting to validate the
1114XML document, the processor <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>, but need
1115not, include the entity's replacement text. If a non-validating processor
1116does not include the replacement text, it <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> inform the application that
1117it recognized, but did not read, the entity.</p><p>This rule is based on the recognition that the automatic inclusion provided
1118by the SGML and XML entity mechanism, primarily designed to support modularity
1119in authoring, is not necessarily appropriate for other applications, in particular
1120document browsing. Browsers, for example, when encountering an external parsed
1121entity reference, might choose to provide a visual indication of the entity's
1122presence and retrieve it for display only on demand.</p></div><div class="div3"> <h4><a name="forbidden" id="forbidden" />4.4.4 Forbidden</h4><p>The following are forbidden, and constitute <a title="Fatal Error" href="#dt-fatal">fatal
1123errors</a>:</p><ul><li><p>the appearance of a reference to an <a title="Unparsed Entity" href="#dt-unparsed">unparsed
1124entity</a><span>, except in the
1125<a href="#NT-EntityValue">EntityValue</a> in an entity declaration</span>.</p></li><li><p>the appearance of any character or general-entity reference in the
1126DTD except within an <a href="#NT-EntityValue">EntityValue</a> or <a href="#NT-AttValue">AttValue</a>.</p></li><li><p>a reference to an external entity in an attribute value.</p></li></ul></div><div class="div3"> <h4><a name="inliteral" id="inliteral" />4.4.5 Included in Literal</h4><p>When an <a title="Entity Reference" href="#dt-entref">entity reference</a> appears in
1127an attribute value, or a parameter entity reference appears in a literal entity
1128value, its <a title="Replacement Text" href="#dt-repltext">replacement text</a> <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> processed
1129in place of the reference itself as though it were part of the document at
1130the location the reference was recognized, except that a single or double
1131quote character in the replacement text <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> always be</span> treated as a normal data
1132character and <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> terminate the literal. For example, this is well-formed:</p><div class="exampleInner"><pre>&lt;!ENTITY % YN '"Yes"' &gt;
1133&lt;!ENTITY WhatHeSaid "He said %YN;" &gt;</pre></div><p>while this is not:</p><div class="exampleInner"><pre>&lt;!ENTITY EndAttr "27'" &gt;
1134&lt;element attribute='a-&amp;EndAttr;&gt;</pre></div></div><div class="div3"> <h4><a name="notify" id="notify" />4.4.6 Notify</h4><p>When the name of an <a title="Unparsed Entity" href="#dt-unparsed">unparsed entity</a>
1135appears as a token in the value of an attribute of declared type <b>ENTITY</b>
1136or <b>ENTITIES</b>, a validating processor <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> inform the application of
1137the <a title="System Identifier" href="#dt-sysid">system</a> and <a title="Public identifier" href="#dt-pubid">public</a>
1138(if any) identifiers for both the entity and its associated <a title="Notation" href="#dt-notation">notation</a>.</p></div><div class="div3"> <h4><a name="bypass" id="bypass" />4.4.7 Bypassed</h4><p>When a general entity reference appears in the <a href="#NT-EntityValue">EntityValue</a>
1139in an entity declaration, it <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> bypassed and left as is.</p></div><div class="div3"> <h4><a name="as-PE" id="as-PE" />4.4.8 Included as PE</h4><p>Just as with external parsed entities, parameter entities need only be <a href="#include-if-valid"><cite>included if validating</cite></a>. When a parameter-entity
1140reference is recognized in the DTD and included, its <a title="Replacement Text" href="#dt-repltext">replacement
1141text</a> <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be</span> enlarged by the attachment of one leading and one following
1142space (#x20) character; the intent is to constrain the replacement text of
1143parameter entities to contain an integral number of grammatical tokens in
1144the DTD. This
1145behavior <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em></span> apply to parameter entity references within entity values;
1146these are described in <a href="#inliteral"><b>4.4.5 Included in Literal</b></a>.</p></div><div class="div3"> <h4><a name="error" id="error" />4.4.9 Error</h4><p>It is an <a title="Error" href="#dt-error">error</a> for a reference to
1147an unparsed entity to appear in the <a href="#NT-EntityValue">EntityValue</a> in an
1148entity declaration.</p></div></div><div class="div2"> <h3><a name="intern-replacement" id="intern-replacement" />4.5 Construction of Entity Replacement Text</h3><p>In discussing the treatment of entities, it is useful to distinguish
1149two forms of the entity's value.
1150[<a name="dt-litentval" id="dt-litentval" title="Literal Entity Value">Definition</a>: <span>For an
1151internal entity, </span>the <b>literal
1152entity value</b> is the quoted string actually present in the entity declaration,
1153corresponding to the non-terminal <a href="#NT-EntityValue">EntityValue</a>.] [<a name="dt-extlitentval" id="dt-extlitentval" title="Literal Entity Value">Definition</a>: For an external entity, the <b>literal
1154entity value</b> is the exact text contained in the entity.] [<a name="dt-repltext" id="dt-repltext" title="Replacement Text">Definition</a>: <span>For an
1155internal entity, </span>the <b>replacement text</b>
1156is the content of the entity, after replacement of character references and
1157parameter-entity references.] [<a name="dt-extrepltext" id="dt-extrepltext" title="Replacement Text">Definition</a>: For
1158an external entity, the <b>replacement text</b> is the content of the entity,
1159after stripping the text declaration (leaving any surrounding white space) if there
1160is one but without any replacement of character references or parameter-entity
1161references.]</p><p>The literal entity value as given in an internal entity declaration (<a href="#NT-EntityValue">EntityValue</a>) <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> contain character, parameter-entity,
1162and general-entity references. Such references <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be contained entirely
1163within the literal entity value. The actual replacement text that is <a title="Include" href="#dt-include">included</a><span> (or <a title="" href="#inliteral">included in literal</a>)</span> as described above
1164<em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> contain the <em>replacement
1165text</em> of any parameter entities referred to, and <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> contain the character
1166referred to, in place of any character references in the literal entity value;
1167however, general-entity references <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be left as-is, unexpanded. For example,
1168given the following declarations:</p><div class="exampleInner"><pre>&lt;!ENTITY % pub "&amp;#xc9;ditions Gallimard" &gt;
1169&lt;!ENTITY rights "All rights reserved" &gt;
1170&lt;!ENTITY book "La Peste: Albert Camus,
1171&amp;#xA9; 1947 %pub;. &amp;rights;" &gt;</pre></div><p>then the replacement text for the entity "<code>book</code>"
1172is:</p><div class="exampleInner"><pre>La Peste: Albert Camus,
1173© 1947 Éditions Gallimard. &amp;rights;</pre></div><p>The general-entity reference "<code>&amp;rights;</code>" would
1174be expanded should the reference "<code>&amp;book;</code>" appear
1175in the document's content or an attribute value.</p><p>These simple rules may have complex interactions; for a detailed discussion
1176of a difficult example, see <a href="#sec-entexpand"><b>C Expansion of Entity and Character References</b></a>.</p></div><div class="div2"> <h3><a name="sec-predefined-ent" id="sec-predefined-ent" />4.6 Predefined Entities</h3><p>[<a name="dt-escape" id="dt-escape" title="escape">Definition</a>: Entity and character references <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em>
1177both be used to <b>escape</b> the left angle bracket, ampersand, and
1178other delimiters. A set of general entities (<code>amp</code>,
1179<code>lt</code>,
1180<code>gt</code>,
1181<code>apos</code>,
1182<code>quot</code>) is specified for
1183this purpose. Numeric character references <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> also be used; they are expanded
1184immediately when recognized and <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be treated as character data, so the
1185numeric character references "<code>&amp;#60;</code>" and "<code>&amp;#38;</code>" <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> be used to escape <code>&lt;</code> and <code>&amp;</code> when they occur
1186in character data.]</p><p>All XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> recognize these entities whether they are declared
1187or not. <a title="For interoperability" href="#dt-interop">For interoperability</a>, valid XML
1188documents <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> declare these entities, like any others, before using them. If
1189the entities <code>lt</code> or <code>amp</code> are declared, they <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be
1190declared as internal entities whose replacement text is a character reference
1191to the respective
1192character (less-than sign or ampersand) being escaped; the double
1193escaping is <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em> for these entities so that references to them produce
1194a well-formed result. If the entities <code>gt</code>, <code>apos</code>,
1195or <code>quot</code> are declared, they <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be declared as internal entities
1196whose replacement text is the single character being escaped (or a character
1197reference to that character; the double escaping here is <span class="mustard"><em class="rfc2119" title="Keyword in RFC 2119 context">OPTIONAL</em></span> but harmless).
1198For example:</p><div class="exampleInner"><pre>&lt;!ENTITY lt "&amp;#38;#60;"&gt;
1199&lt;!ENTITY gt "&amp;#62;"&gt;
1200&lt;!ENTITY amp "&amp;#38;#38;"&gt;
1201&lt;!ENTITY apos "&amp;#39;"&gt;
1202&lt;!ENTITY quot "&amp;#34;"&gt;</pre></div></div><div class="div2"> <h3><a name="Notations" id="Notations" />4.7 Notation Declarations</h3><p>[<a name="dt-notation" id="dt-notation" title="Notation">Definition</a>: <b>Notations</b> identify
1203by name the format of <a title="External Entity" href="#dt-extent">unparsed entities</a>,
1204the format of elements which bear a notation attribute, or the application
1205to which a <a title="Processing instruction" href="#dt-pi">processing instruction</a> is addressed.]</p><p>[<a name="dt-notdecl" id="dt-notdecl" title="Notation Declaration">Definition</a>: <b>Notation declarations</b>
1206provide a name for the notation, for use in entity and attribute-list declarations
1207and in attribute specifications, and an external identifier for the notation
1208which may allow an XML processor or its client application to locate a helper
1209application capable of processing data in the given notation.]</p> <h5><a name="IDAYTFU" id="IDAYTFU" />Notation Declarations</h5><table class="scrap" summary="Scrap"><tbody><tr valign="baseline"><td><a name="NT-NotationDecl" id="NT-NotationDecl" />[82]   </td><td><code>NotationDecl</code></td><td>   ::=   </td><td><code>'&lt;!NOTATION' <a href="#NT-S">S</a> <a href="#NT-Name">Name</a> <a href="#NT-S">S</a> (<a href="#NT-ExternalID">ExternalID</a> | <a href="#NT-PublicID">PublicID</a>) <a href="#NT-S">S</a>? '&gt;'</code></td><td><a href="#UniqueNotationName">[VC: Unique Notation Name]</a></td></tr></tbody><tbody><tr valign="baseline"><td><a name="NT-PublicID" id="NT-PublicID" />[83]   </td><td><code>PublicID</code></td><td>   ::=   </td><td><code>'PUBLIC' <a href="#NT-S">S</a> <a href="#NT-PubidLiteral">PubidLiteral</a></code></td></tr></tbody></table><div class="constraint"><p class="prefix"><a name="UniqueNotationName" id="UniqueNotationName" /><b>Validity constraint: Unique Notation Name</b></p><p><span class="mustard">A given <a href="#NT-Name">Name</a> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> be declared in more than one notation declaration.</span></p></div><p>XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> provide applications with the name and external identifier(s)
1210of any notation declared and referred to in an attribute value, attribute
1211definition, or entity declaration. They <em class="rfc2119" title="Keyword in RFC 2119 context">MAY</em> additionally resolve the external
1212identifier into the <a title="System Identifier" href="#dt-sysid">system identifier</a>, file
1213name, or other information needed to allow the application to call a processor
1214for data in the notation described. (It is not an error, however, for XML
1215documents to declare and refer to notations for which notation-specific applications
1216are not available on the system where the XML processor or application is
1217running.)</p></div><div class="div2"> <h3><a name="sec-doc-entity" id="sec-doc-entity" />4.8 Document Entity</h3><p>[<a name="dt-docent" id="dt-docent" title="Document Entity">Definition</a>: The <b>document entity</b>
1218serves as the root of the entity tree and a starting-point for an <a title="XML Processor" href="#dt-xml-proc">XML processor</a>.] This specification does
1219not specify how the document entity is to be located by an XML processor;
1220unlike other entities, the document entity has no name and might well appear
1221on a processor input stream without any identification at all.</p></div></div><div class="div1"> <h2><a name="sec-conformance" id="sec-conformance" />5 Conformance</h2><div class="div2"> <h3><a name="proc-types" id="proc-types" />5.1 Validating and Non-Validating Processors</h3><p>Conforming <a title="XML Processor" href="#dt-xml-proc">XML processors</a> fall into
1222two classes: validating and non-validating.</p><p>Validating and non-validating processors alike <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> report violations of
1223this specification's well-formedness constraints in the content of the <a title="Document Entity" href="#dt-docent">document entity</a> and any other <a title="Text Entity" href="#dt-parsedent">parsed
1224entities</a> that they read.</p><p>[<a name="dt-validating" id="dt-validating" title="Validating Processor">Definition</a>: <b>Validating
1225processors</b> <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>,
1226at user option, report violations of the constraints expressed by
1227the declarations in the <a title="Document Type Declaration" href="#dt-doctype">DTD</a>, and failures
1228to fulfill the validity constraints given in this specification.]
1229To accomplish this, validating XML processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> read and process the entire
1230DTD and all external parsed entities referenced in the document.</p><p>Non-validating processors are <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em> to check only the <a title="Document Entity" href="#dt-docent">document
1231entity</a>, including the entire internal DTD subset, for well-formedness. [<a name="dt-use-mdecl" id="dt-use-mdecl" title="Process Declarations">Definition</a>: While they are not required
1232to check the document for validity, they are <em class="rfc2119" title="Keyword in RFC 2119 context">REQUIRED</em> to <b>process</b>
1233all the declarations they read in the internal DTD subset and in any parameter
1234entity that they read, up to the first reference to a parameter entity that
1235they do <em>not</em> read; that is to say, they <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> use the information
1236in those declarations to <a href="#AVNormalize"><cite>normalize</cite></a>
1237attribute values, <a href="#included"><cite>include</cite></a> the replacement
1238text of internal entities, and supply <a href="#sec-attr-defaults"><cite>default
1239attribute values</cite></a>.] Except when <code>standalone="yes"</code>, they
1240<em class="rfc2119" title="Keyword in RFC 2119 context">MUST NOT</em> <a title="Process Declarations" href="#dt-use-mdecl">process</a> <a title="entity declaration" href="#dt-entdecl">entity
1241declarations</a> or <a title="Attribute-List Declaration" href="#dt-attdecl">attribute-list declarations</a>
1242encountered after a reference to a parameter entity that is not read, since
1243the entity may have contained overriding declarations<span>; when <code>standalone="yes"</code>, processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em>
1244process these declarations</span>.</p><p>Note
1245that when processing invalid documents with a non-validating
1246processor the application may not be presented with consistent
1247information. For example, several requirements for uniqueness
1248within the document may not be met, including more than one element
1249with the same id, duplicate declarations of elements or notations
1250with the same name, etc. In these cases the behavior of the parser
1251with respect to reporting such information to the application is
1252undefined.</p><p>XML 1.1 processors <em class="rfc2119" title="Keyword in RFC 2119 context">MUST</em> be able to process both XML 1.0
1253and XML 1.1 documents. Programs which generate XML <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em>
1254generate XML 1.0, unless one of the specific features of XML 1.1 is required.</p></div><div class="div2"> <h3><a name="safe-behavior" id="safe-behavior" />5.2 Using XML Processors</h3><p>The behavior of a validating XML processor is highly predictable; it must
1255read every piece of a document and report all well-formedness and validity
1256violations. Less is required of a non-validating processor; it need not read
1257any part of the document other than the document entity. This has two effects
1258that may be important to users of XML processors:</p><ul><li><p>Certain well-formedness errors, specifically those that require reading
1259external entities, <span>may fail to</span> be detected by a non-validating processor. Examples
1260include the constraints entitled <a href="#wf-entdeclared"><cite>Entity Declared</cite></a>, <a href="#textent"><cite>Parsed Entity</cite></a>, and <a href="#norecursion"><cite>No
1261Recursion</cite></a>, as well as some of the cases described as <a href="#forbidden"><cite>forbidden</cite></a> in <a href="#entproc"><b>4.4 XML Processor Treatment of Entities and References</b></a>.</p></li><li><p>The information passed from the processor to the application may
1262vary, depending on whether the processor reads parameter and external entities.
1263For example, a non-validating processor <span>may fail to</span> <a href="#AVNormalize"><cite>normalize</cite></a>
1264attribute values, <a href="#included"><cite>include</cite></a> the replacement
1265text of internal entities, or supply <a href="#sec-attr-defaults"><cite>default
1266attribute values</cite></a>, where doing so depends on having read declarations
1267in external or parameter entities.</p></li></ul><p>For maximum reliability in interoperating between different XML processors,
1268applications which use non-validating processors <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD NOT</em> rely on any behaviors
1269not required of such processors. Applications which require DTD facilities
1270not related to validation (such
1271as the declaration of default attributes and internal entities that are
1272or may be specified in
1273external entities <em class="rfc2119" title="Keyword in RFC 2119 context">SHOULD</em> use validating XML processors.</p></div></div><div class="div1"> <h2><a name="sec-notation" id="sec-notation" />6 Notation</h2><p>The formal grammar of XML is given in this specification using a simple
1274Extended Backus-Naur Form (EBNF) notation. Each rule in the grammar defines
1275one symbol, in the form</p><div class="exampleInner"><pre>symbol ::= expression</pre></div><p>Symbols are written with an initial capital letter if they are the
1276start symbol of a regular language, otherwise with an initial lowercase
1277letter. Literal strings are quoted.</p><p>Within the expression on the right-hand side of a rule, the following expressions
1278are used to match strings of one or more characters: </p><dl><dt class="label"><code>#xN</code></dt><dd><p>where <code>N</code> is a hexadecimal integer, the expression matches the character
1279<span>whose</span><span> number
1280(code point) in</span> ISO/IEC 10646 <span>is <code>N</code></span>. The number of leading zeros in the <code>#xN</code>
1281form is insignificant.</p></dd><dt class="label"><code>[a-zA-Z]</code>, <code>[#xN-#xN]</code></dt><dd><p>matches any <a href="#NT-Char">Char</a> with a value in the range(s) indicated (inclusive).</p></dd><dt class="label"><code>[abc]</code>, <code>[#xN#xN#xN]</code></dt><dd><p>matches any <a href="#NT-Char">Char</a> with a value among the characters
1282enumerated. Enumerations and ranges can be mixed in one set of brackets.</p></dd><dt class="label"><code>[^a-z]</code>, <code>[^#xN-#xN]</code></dt><dd><p>matches any <a href="#NT-Char">Char</a> with a value <em>outside</em> the range
1283indicated.</p></dd><dt class="label"><code>[^abc]</code>, <code>[^#xN#xN#xN]</code></dt><dd><p>matches any <a href="#NT-Char">Char</a> with a value not among the characters given. Enumerations
1284and ranges of forbidden values can be mixed in one set of brackets.</p></dd><dt class="label"><code>"string"</code></dt><dd><p>matches a literal string <a title="match" href="#dt-match">matching</a> that
1285given inside the double quotes.</p></dd><dt class="label"><code>'string'</code></dt><dd><p>matches a literal string <a title="match" href="#dt-match">matching</a> that
1286given inside the single quotes.</p></dd></dl><p> These symbols may be combined to match more complex patterns as follows,
1287where <code>A</code> and <code>B</code> represent simple expressions: </p><dl><dt class="label">(<code>expression</code>)</dt><dd><p><code>expression</code> is treated as a unit and may be combined as described
1288in this list.</p></dd><dt class="label"><code>A?</code></dt><dd><p>matches <code>A</code> or nothing; optional <code>A</code>.</p></dd><dt class="label"><code>A B</code></dt><dd><p>matches <code>A</code> followed by <code>B</code>. This
1289operator has higher precedence than alternation; thus <code>A B | C D</code>
1290is identical to <code>(A B) | (C D)</code>.</p></dd><dt class="label"><code>A | B</code></dt><dd><p>matches <code>A</code> or <code>B</code>.</p></dd><dt class="label"><code>A - B</code></dt><dd><p>matches any string that matches <code>A</code> but does not match <code>B</code>.</p></dd><dt class="label"><code>A+</code></dt><dd><p>matches one or more occurrences of <code>A</code>. Concatenation
1291has higher precedence than alternation; thus <code>A+ | B+</code> is identical
1292to <code>(A+) | (B+)</code>.</p></dd><dt class="label"><code>A*</code></dt><dd><p>matches zero or more occurrences of <code>A</code>. Concatenation
1293has higher precedence than alternation; thus <code>A* | B*</code> is identical
1294to <code>(A*) | (B*)</code>.</p></dd></dl><p> Other notations used in the productions are: </p><dl><dt class="label"><code>/* ... */</code></dt><dd><p>comment.</p></dd><dt class="label"><code>[ wfc: ... ]</code></dt><dd><p>well-formedness constraint; this identifies by name a constraint on <a title="Well-Formed" href="#dt-wellformed">well-formed</a> documents associated with a production.</p></dd><dt class="label"><code>[ vc: ... ]</code></dt><dd><p>validity constraint; this identifies by name a constraint on <a title="Validity" href="#dt-valid">valid</a>
1295documents associated with a production.</p></dd></dl><p></p></div></div><div class="back"><div class="div1"> <h2><a name="sec-bibliography" id="sec-bibliography" />A References</h2><div class="div2"> <h3><a name="sec-existing-stds" id="sec-existing-stds" />A.1 Normative References</h3><dl><dt class="label"><a name="IANA" id="IANA" />IANA-CHARSETS</dt><dd>(Internet
1296Assigned Numbers Authority) <a href="http://www.iana.org/assignments/character-sets"><cite>Official Names for Character Sets</cite></a>,
1297ed. Keld Simonsen et al. (See http://www.iana.org/assignments/character-sets.)</dd><dt class="label"><a name="rfc2119" id="rfc2119" />IETF RFC 2119</dt><dd>IETF
1298(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc2119.txt"><cite>RFC 2119: Key words for use in RFCs to Indicate Requirement Levels</cite></a>.
1299Scott Bradner, 1997. (See http://www.ietf.org/rfc/rfc2119.txt.)</dd><dt class="label"><a name="rfc2396" id="rfc2396" />IETF RFC 2396</dt><dd>IETF
1300(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc2396.txt"><cite>RFC 2396: Uniform Resource Identifiers
1301(URI): Generic Syntax</cite></a>. T. Berners-Lee, R. Fielding, L. Masinter.
13021998. (See http://www.ietf.org/rfc/rfc2396.txt.)</dd><dt class="label"><a name="rfc2732" id="rfc2732" />IETF RFC 2732</dt><dd>IETF
1303(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc2732.txt"><cite>RFC 2732: Format for Literal
1304IPv6 Addresses in URL's</cite></a>. R. Hinden, B. Carpenter, L. Masinter.
13051999. (See http://www.ietf.org/rfc/rfc2732.txt.)</dd><dt class="label"><a name="RFC1766" id="RFC1766" />IETF RFC 3066</dt><dd>IETF
1306(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc3066.txt"><cite>RFC 3066: Tags for the Identification
1307of Languages</cite></a>, ed. H. Alvestrand. 2001. (See http://www.ietf.org/rfc/rfc3066.txt.)</dd><dt class="label"><a name="ISO10646" id="ISO10646" />ISO/IEC 10646</dt><dd><span>ISO (International
1308Organization for Standardization). <cite>ISO/IEC 10646-1:2000. Information
1309technology — Universal Multiple-Octet Coded Character Set (UCS) —
1310Part 1: Architecture and Basic Multilingual Plane</cite> and <cite>ISO/IEC 10646-2:2001.
1311Information technology — Universal Multiple-Octet Coded Character Set (UCS) — Part 2:
1312Supplementary Planes</cite>, as, from time to time, amended, replaced by a new edition or
1313expanded by the addition of new parts. [Geneva]: International Organization for Standardization.
1314(See <a href="http://www.iso.ch">http://www.iso.ch</a> for the latest version.)</span></dd><dt class="label"><a name="Unicode" id="Unicode" />Unicode</dt><dd>The Unicode Consortium. <em>The Unicode
1315Standard, Version 4.0.</em> Reading, Mass.: Addison-Wesley,
13162003,
1317as updated from time to time by the publication of new versions. (See
1318<a href="http://www.unicode.org/unicode/standard/versions">
1319http://www.unicode.org/unicode/standard/versions</a> for the latest version
1320and additional information on versions of the standard and of the Unicode
1321Character Database).</dd><dt class="label"><a name="XML1.0" />XML-1.0</dt><dd>W3C. <a href="http://www.w3.org/TR/REC-xml"><cite>Extensible Markup Language (XML) 1.0 (Third
1322Edition)</cite></a>. Tim Bray, Jean Paoli, C.M. Sperberg-McQueen, Eve Maler, François Yergeau
1323(editors) (See http://www.w3.org/TR/REC-xml.)</dd></dl></div><div class="div2"> <h3><a name="null" id="null" />A.2 Other References</h3><dl><dt class="label"><a name="Aho" id="Aho" />Aho/Ullman</dt><dd>Aho, Alfred V., Ravi Sethi, and Jeffrey D.
1324Ullman. <cite>Compilers: Principles, Techniques, and Tools</cite>.
1325Reading: Addison-Wesley, 1986, rpt. corr. 1988.</dd><dt class="label"><a name="ABK" id="ABK" />Brüggemann-Klein</dt><dd>Brüggemann-Klein,
1326Anne. <a href="ftp://ftp.informatik.uni-freiburg.de/documents/papers/brueggem/habil.ps"><cite>Formal Models in Document Processing</cite></a>. Habilitationsschrift. Faculty
1327of Mathematics at the University of Freiburg, 1993. (See ftp://ftp.informatik.uni-freiburg.de/documents/papers/brueggem/habil.ps.)</dd><dt class="label"><a name="ABKDW" id="ABKDW" />Brüggemann-Klein and Wood</dt><dd>Brüggemann-Klein,
1328Anne, and Derick Wood. <cite>Deterministic Regular Languages</cite>.
1329Universität Freiburg, Institut für Informatik, Bericht 38, Oktober 1991. Extended
1330abstract in A. Finkel, M. Jantzen, Hrsg., STACS 1992, S. 173-184. Springer-Verlag,
1331Berlin 1992. Lecture Notes in Computer Science 577. Full version titled <cite>One-Unambiguous
1332Regular Languages</cite> in Information and Computation 140 (2): 229-253,
1333February 1998.</dd><dt class="label"><a name="Charmod" />Charmod</dt><dd>W3C Working Draft.
1334
1335<a href="http://www.w3.org/TR/2003/WD-charmod-20030822/"><cite>Character Model for the World Wide Web 1.0</cite></a>.
1336
1337Martin J. Dürst, François Yergeau, Richard Ishida, Misha Wolf, Tex Texin. (See http://www.w3.org/TR/2003/WD-charmod-20030822/.)</dd><dt class="label"><a name="Clark" id="Clark" />Clark</dt><dd>James Clark.
1338<a href="http://www.w3.org/TR/NOTE-sgml-xml-971215"><cite>Comparison of SGML and XML</cite></a>. (See http://www.w3.org/TR/NOTE-sgml-xml-971215.)</dd><dt class="label"><a name="IANA-LANGCODES" id="IANA-LANGCODES" />IANA-LANGCODES</dt><dd>(Internet
1339Assigned Numbers Authority) <a href="http://www.iana.org/assignments/language-tags"><cite>Registry of Language Tags</cite></a>,
1340ed. Keld Simonsen et al. (See http://www.iana.org/assignments/language-tags.)</dd><dt class="label"><a name="RFC2141" id="RFC2141" />IETF RFC 2141</dt><dd>IETF
1341(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc2141.txt"><cite>RFC 2141: URN Syntax</cite></a>, ed.
1342R. Moats. 1997. (See http://www.ietf.org/rfc/rfc2141.txt.)</dd><dt class="label"><a name="rfc2376" id="rfc2376" />IETF RFC 3023</dt><dd>IETF
1343(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc3023.txt"><cite>RFC 3023: XML Media Types</cite></a>.
1344eds. M. Murata, S. St.Laurent, D. Kohn. 2001. (See http://www.ietf.org/rfc/rfc3023.txt.)</dd><dt class="label"><a name="rfc2781" id="rfc2781" />IETF RFC 2781</dt><dd>IETF
1345(Internet Engineering Task Force). <a href="http://www.ietf.org/rfc/rfc2781.txt"><cite>RFC 2781: UTF-16, an encoding
1346of ISO 10646</cite></a>, ed. P. Hoffman, F. Yergeau. 2000. (See http://www.ietf.org/rfc/rfc2781.txt.)</dd><dt class="label"><a name="ISO639" id="ISO639" />ISO 639</dt><dd>(International Organization for Standardization).
1347<cite>ISO 639:1988 (E).
1348Code for the representation of names of languages.</cite> [Geneva]: International
1349Organization for Standardization, 1988.</dd><dt class="label"><a name="ISO3166" id="ISO3166" />ISO 3166</dt><dd>(International Organization for Standardization).
1350<cite>ISO 3166-1:1997
1351(E). Codes for the representation of names of countries and their subdivisions —
1352Part 1: Country codes</cite> [Geneva]: International Organization for
1353Standardization, 1997.</dd><dt class="label"><a name="ISO8879" id="ISO8879" />ISO 8879</dt><dd>ISO (International Organization for Standardization). <cite>ISO
13548879:1986(E). Information processing — Text and Office Systems —
1355Standard Generalized Markup Language (SGML).</cite> First edition —
13561986-10-15. [Geneva]: International Organization for Standardization, 1986. </dd><dt class="label"><a name="ISO10744" id="ISO10744" />ISO/IEC 10744</dt><dd>ISO (International Organization for
1357Standardization). <cite>ISO/IEC 10744-1992 (E). Information technology —
1358Hypermedia/Time-based Structuring Language (HyTime). </cite> [Geneva]:
1359International Organization for Standardization, 1992. <em>Extended Facilities
1360Annexe.</em> [Geneva]: International Organization for Standardization, 1996. </dd><dt class="label"><a name="websgml" id="websgml" />WEBSGML</dt><dd>ISO
1361(International Organization for Standardization). <a href="http://www.sgmlsource.com/8879/n0029.htm"><cite>ISO 8879:1986
1362TC2. Information technology — Document Description and Processing Languages</cite></a>.
1363[Geneva]: International Organization for Standardization, 1998. (See http://www.sgmlsource.com/8879/n0029.htm.)</dd><dt class="label"><a name="xml-names" id="xml-names" />XML Names</dt><dd>Tim Bray,
1364Dave Hollander, and Andrew Layman, editors. <a href="http://www.w3.org/TR/REC-xml-names/"><cite>Namespaces in XML</cite></a>.
1365Textuality, Hewlett-Packard, and Microsoft. World Wide Web Consortium, 1999. (See http://www.w3.org/TR/REC-xml-names/.)</dd></dl></div></div><div class="div1"> <h2><a name="sec-CharNorm" id="sec-CharNorm" />B Definitions for Character Normalization</h2><p>This appendix contains the necessary definitions for character normalization.
1366For additional background information and examples, see <a href="#Charmod">[Charmod]</a>.</p><p>
1367[<a name="dt-Uni-encform" id="dt-Uni-encform" title="Unicode encoding form">Definition</a>: Text is said to be
1368in a <b>Unicode encoding form</b> if it is encoded in
1369UTF-8, UTF-16 or UTF-32.]</p><p>
1370[<a name="dt-legacyenc" id="dt-legacyenc" title="legacy encoding">Definition</a>: <b>Legacy encoding</b>
1371is taken to mean any character encoding not based on Unicode.]</p><p>
1372[<a name="dt-normtransc" id="dt-normtransc" title="normalizing transcoder">Definition</a>: A
1373<b>normalizing transcoder</b> is a transcoder that converts from a
1374<a title="legacy encoding" href="#dt-legacyenc">legacy encoding</a> to a
1375<a title="Unicode encoding form" href="#dt-Uni-encform">Unicode encoding form</a> and
1376ensures that the result is in Unicode Normalization Form C
1377(see UAX #15 <a href="#Unicode">[Unicode]</a>).]</p><p>[<a name="dt-charesc" id="dt-charesc" title="character escape">Definition</a>: A <b>character escape</b>
1378is a syntactic device defined in a markup or programming language that allows
1379one or more of:]</p><ol type="1"><li><p>expressing syntax-significant characters while disregarding
1380their significance in the syntax of the language, or</p></li><li><p>expressing characters not representable in the character encoding
1381chosen for an instance of the language, or</p></li><li><p>expressing characters in general, without use of the corresponding
1382character codes.</p></li></ol><p>
1383[<a name="dt-certified" id="dt-certified" title="certified">Definition</a>: <b>Certified</b> text
1384is text which satisfies at least one of the following conditions:]</p><ol type="1"><li><p>it has been confirmed through inspection that the text
1385is in normalized form</p></li><li><p>the source text-processing component is identified
1386and is known to produce only normalized text.</p></li></ol><p>
1387[<a name="dt-uninorm" id="dt-uninorm" title="Unicode-normalized">Definition</a>: Text is, for the purposes of
1388this specification, <b>Unicode-normalized</b> if it is in a
1389<a title="Unicode encoding form" href="#dt-Uni-encform">Unicode encoding form</a> and is in
1390Unicode Normalization Form C, according to a version of Unicode Standard Annex #15:
1391Unicode Normalization Forms <a href="#Unicode">[Unicode]</a> at least as recent as the
1392oldest version of the Unicode Standard that contains all the characters
1393actually present in the text, but no earlier
1394than version 3.2.]</p><p>
1395[<a name="dt-inclnorm" id="dt-inclnorm" title="include-normalized">Definition</a>: Text is
1396<b>include-normalized</b> if:]</p><ol type="1"><li><p>the text is <a title="Unicode-normalized" href="#dt-uninorm">Unicode-normalized</a>
1397and does not contain any <a title="character escape" href="#dt-charesc">character escapes</a>
1398or <a title="Include" href="#dt-include">includes</a> whose expansion would
1399cause the text to become no longer <a title="Unicode-normalized" href="#dt-uninorm">Unicode-normalized</a>;
1400or</p></li><li><p>the text is in a <a title="legacy encoding" href="#dt-legacyenc">legacy encoding</a> and, if it were transcoded
1401to a <a title="Unicode encoding form" href="#dt-Uni-encform">Unicode encoding form</a> by a
1402<a title="normalizing transcoder" href="#dt-normtransc">normalizing transcoder</a>, the resulting
1403text would satisfy clause 1 above.</p></li></ol><p>
1404[<a name="dt-compchar" id="dt-compchar" title="composing character">Definition</a>: A <b>composing character</b>
1405is a character that is one or both of the following:]</p><ol type="1"><li><p>the second character in the canonical decomposition mapping of
1406some primary composite (as defined in D3 of UAX #15 <a href="#Unicode">[Unicode]</a>), or</p></li><li><p>of non-zero canonical combining class (as defined in Unicode
1407<a href="#Unicode">[Unicode]</a>).</p></li></ol><p>
1408[<a name="dt-fullnorm" id="dt-fullnorm" title="fully normalized">Definition</a>: Text is
1409<b>fully-normalized</b> if:]</p><ol type="1"><li><p>the text is in a <a title="Unicode encoding form" href="#dt-Uni-encform">Unicode encoding
1410form</a>, is <a title="include-normalized" href="#dt-inclnorm">include-normalized</a> and
1411none of the <a title="" href="#dt-relconst"><span>relevant</span>
1412constructs</a> comprising the text begin with a
1413<a title="composing character" href="#dt-compchar">composing character</a> or a
1414character escape representing a
1415<a title="composing character" href="#dt-compchar">composing character</a>; or</p></li><li><p>the text is in a <a title="legacy encoding" href="#dt-legacyenc">legacy encoding</a> and,
1416if it were transcoded to a <a title="Unicode encoding form" href="#dt-Uni-encform">Unicode encoding form</a>
1417by a <a title="normalizing transcoder" href="#dt-normtransc">normalizing transcoder</a>, the resulting text
1418would satisfy clause 1 above.</p></li></ol></div><div class="div1"> <h2><a name="sec-entexpand" id="sec-entexpand" />C Expansion of Entity and Character References (Non-Normative)</h2><p>This appendix contains some examples illustrating the sequence of entity-
1419and character-reference recognition and expansion, as specified in <a href="#entproc"><b>4.4 XML Processor Treatment of Entities and References</b></a>.</p><p>If the DTD contains the declaration</p><div class="exampleInner"><pre>&lt;!ENTITY example "&lt;p&gt;An ampersand (&amp;#38;#38;) may be escaped
1420numerically (&amp;#38;#38;#38;) or with a general entity
1421(&amp;amp;amp;).&lt;/p&gt;" &gt;</pre></div><p>then the XML processor will recognize the character references when it
1422parses the entity declaration, and resolve them before storing the following
1423string as the value of the entity "<code>example</code>":</p><div class="exampleInner"><pre>&lt;p&gt;An ampersand (&amp;#38;) may be escaped
1424numerically (&amp;#38;#38;) or with a general entity
1425(&amp;amp;amp;).&lt;/p&gt;</pre></div><p>A reference in the document to "<code>&amp;example;</code>"
1426will cause the text to be reparsed, at which time the start- and end-tags
1427of the <code>p</code> element will be recognized and the three references will
1428be recognized and expanded, resulting in a <code>p</code> element with the following
1429content (all data, no delimiters or markup):</p><div class="exampleInner"><pre>An ampersand (&amp;) may be escaped
1430numerically (&amp;#38;) or with a general entity
1431(&amp;amp;).</pre></div><p>A more complex example will illustrate the rules and their effects fully.
1432In the following example, the line numbers are solely for reference.</p><div class="exampleInner"><pre>1 &lt;?xml version='1.0'?&gt;
14332 &lt;!DOCTYPE test [
14343 &lt;!ELEMENT test (#PCDATA) &gt;
14354 &lt;!ENTITY % xx '&amp;#37;zz;'&gt;
14365 &lt;!ENTITY % zz '&amp;#60;!ENTITY tricky "error-prone" &gt;' &gt;
14376 %xx;
14387 ]&gt;
14398 &lt;test&gt;This sample shows a &amp;tricky; method.&lt;/test&gt;</pre></div><p>This produces the following:</p><ul><li><p>in line 4, the reference to character 37 is expanded immediately,
1440and the parameter entity "<code>xx</code>" is stored in the symbol
1441table with the value "<code>%zz;</code>". Since the replacement
1442text is not rescanned, the reference to parameter entity "<code>zz</code>"
1443is not recognized. (And it would be an error if it were, since "<code>zz</code>"
1444is not yet declared.)</p></li><li><p>in line 5, the character reference "<code>&amp;#60;</code>"
1445is expanded immediately and the parameter entity "<code>zz</code>"
1446is stored with the replacement text "<code>&lt;!ENTITY tricky "error-prone"
1447&gt;</code>", which is a well-formed entity declaration.</p></li><li><p>in line 6, the reference to "<code>xx</code>" is recognized,
1448and the replacement text of "<code>xx</code>" (namely "<code>%zz;</code>")
1449is parsed. The reference to "<code>zz</code>" is recognized in
1450its turn, and its replacement text ("<code>&lt;!ENTITY tricky "error-prone"
1451&gt;</code>") is parsed. The general entity "<code>tricky</code>"
1452has now been declared, with the replacement text "<code>error-prone</code>".</p></li><li><p>in line 8, the reference to the general entity "<code>tricky</code>"
1453is recognized, and it is expanded, so the full content of the <code>test</code>
1454element is the self-describing (and ungrammatical) string <em>This sample
1455shows a error-prone method.</em></p></li></ul></div><div class="div1"> <h2><a name="determinism" id="determinism" />D Deterministic Content Models (Non-Normative)</h2><p>As
1456noted in <a href="#sec-element-content"><b>3.2.1 Element Content</b></a>, it is required that content
1457models in element type declarations be deterministic. This requirement is <a title="For Compatibility" href="#dt-compat">for compatibility</a> with SGML (which calls deterministic
1458content models "unambiguous"); XML processors built
1459using SGML systems may flag non-deterministic content models as errors.</p><p>For example, the content model <code>((b, c) | (b, d))</code> is non-deterministic,
1460because given an initial <code>b</code> the XML processor
1461cannot know which <code>b</code> in the model is being matched without looking
1462ahead to see which element follows the <code>b</code>. In this case, the two references
1463to <code>b</code> can be collapsed into a single reference, making the model read <code>(b,
1464(c | d))</code>. An initial <code>b</code> now clearly matches only a single name
1465in the content model. The processor doesn't need to look ahead to see what follows; either <code>c</code> or <code>d</code>
1466would be accepted.</p><p>More formally: a finite state automaton may be constructed from the content
1467model using the standard algorithms, e.g. algorithm 3.5 in section 3.9 of
1468Aho, Sethi, and Ullman <a href="#Aho">[Aho/Ullman]</a>. In many such algorithms, a follow
1469set is constructed for each position in the regular expression (i.e., each
1470leaf node in the syntax tree for the regular expression); if any position
1471has a follow set in which more than one following position is labeled with
1472the same element type name, then the content model is in error and may be
1473reported as an error.</p><p>Algorithms exist which allow many but not all non-deterministic content
1474models to be reduced automatically to equivalent deterministic models; see
1475Brüggemann-Klein 1991 <a href="#ABK">[Brüggemann-Klein]</a>.</p></div><div class="div1"> <h2><a name="sec-guessing" id="sec-guessing" />E Autodetection of Character Encodings (Non-Normative)</h2><p>The XML encoding declaration functions as an internal label on each entity,
1476indicating which character encoding is in use. Before an XML processor can
1477read the internal label, however, it apparently has to know what character
1478encoding is in use — which is what the internal label is trying to indicate.
1479In the general case, this is a hopeless situation. It is not entirely hopeless
1480in XML, however, because XML limits the general case in two ways: each implementation
1481is assumed to support only a finite set of character encodings, and the XML
1482encoding declaration is restricted in position and content in order to make
1483it feasible to autodetect the character encoding in use in each entity in
1484normal cases. Also, in many cases other sources of information are available
1485in addition to the XML data stream itself. Two cases may be distinguished,
1486depending on whether the XML entity is presented to the processor without,
1487or with, any accompanying (external) information. We consider the first case
1488first.</p><div class="div2"> <h3><a name="sec-guessing-no-ext-info" id="sec-guessing-no-ext-info" />E.1 Detection Without External Encoding Information</h3><p>Because each XML entity not accompanied by external
1489encoding information and not in UTF-8 or UTF-16 encoding must
1490begin with an XML encoding declaration, in which the first characters must
1491be '<code>&lt;?xml</code>', any conforming processor can detect, after two
1492to four octets of input, which of the following cases apply. In reading this
1493list, it may help to know that in UCS-4, '&lt;' is "<code>#x0000003C</code>"
1494and '?' is "<code>#x0000003F</code>", and the Byte Order Mark
1495required of UTF-16 data streams is "<code>#xFEFF</code>". The notation
1496<var>##</var> is used to denote any byte value except that two consecutive
1497<var>##</var>s cannot be both 00.</p><p>With a Byte Order Mark:</p><table border="1" frame="border" summary="Encoding detection summary"><tbody><tr><td rowspan="1" colspan="1"><code>00 00 FE
1498FF</code></td><td rowspan="1" colspan="1">UCS-4, big-endian machine (1234 order)</td></tr><tr><td rowspan="1" colspan="1"><code>FF
1499FE 00 00</code></td><td rowspan="1" colspan="1">UCS-4, little-endian machine (4321 order)</td></tr><tr><td rowspan="1" colspan="1"><code>00 00 FF FE</code></td><td rowspan="1" colspan="1">UCS-4, unusual octet order (2143)</td></tr><tr><td rowspan="1" colspan="1"><code>FE FF 00 00</code></td><td rowspan="1" colspan="1">UCS-4, unusual octet order (3412)</td></tr><tr><td rowspan="1" colspan="1"><code>FE FF ## ##</code></td><td rowspan="1" colspan="1">UTF-16, big-endian</td></tr><tr><td rowspan="1" colspan="1"><code>FF FE ## ##</code></td><td rowspan="1" colspan="1">UTF-16, little-endian</td></tr><tr><td rowspan="1" colspan="1"><code>EF BB BF</code></td><td rowspan="1" colspan="1">UTF-8</td></tr></tbody></table><p>Without a Byte Order Mark:</p><table border="1" frame="border" summary="Encoding detection summary"><tbody><tr><td rowspan="1" colspan="1"><code>00 00 00 3C</code></td><td rowspan="4" colspan="1">UCS-4 or other encoding with a 32-bit code unit and ASCII
1500characters encoded as ASCII values, in respectively big-endian (1234), little-endian
1501(4321) and two unusual byte orders (2143 and 3412). The encoding declaration
1502must be read to determine which of UCS-4 or other supported 32-bit encodings
1503applies.</td></tr><tr><td rowspan="1" colspan="1"><code>3C 00 00 00</code></td></tr><tr><td rowspan="1" colspan="1"><code>00 00 3C 00</code></td></tr><tr><td rowspan="1" colspan="1"><code>00 3C 00 00</code></td></tr><tr><td rowspan="1" colspan="1"><code>00 3C 00 3F</code></td><td rowspan="1" colspan="1">UTF-16BE or big-endian ISO-10646-UCS-2
1504or other encoding with a 16-bit code unit in big-endian order and ASCII characters
1505encoded as ASCII values (the encoding declaration must be read to determine
1506which)</td></tr><tr><td rowspan="1" colspan="1"><code>3C 00 3F 00</code></td><td rowspan="1" colspan="1">UTF-16LE or little-endian
1507ISO-10646-UCS-2 or other encoding with a 16-bit code unit in little-endian
1508order and ASCII characters encoded as ASCII values (the encoding declaration
1509must be read to determine which)</td></tr><tr><td rowspan="1" colspan="1"><code>3C 3F 78 6D</code></td><td rowspan="1" colspan="1">UTF-8, ISO 646, ASCII, some part of ISO 8859, Shift-JIS, EUC, or any other
15107-bit, 8-bit, or mixed-width encoding which ensures that the characters of
1511ASCII have their normal positions, width, and values; the actual encoding
1512declaration must be read to detect which of these applies, but since all of
1513these encodings use the same bit patterns for the relevant ASCII characters,
1514the encoding declaration itself may be read reliably</td></tr><tr><td rowspan="1" colspan="1"><code>4C
15156F A7 94</code></td><td rowspan="1" colspan="1">EBCDIC (in some flavor; the full encoding declaration
1516must be read to tell which code page is in use)</td></tr><tr><td rowspan="1" colspan="1">Other</td><td rowspan="1" colspan="1">UTF-8 without an encoding declaration, or else the data stream is mislabeled
1517(lacking a required encoding declaration), corrupt, fragmentary, or enclosed
1518in a wrapper of some kind</td></tr></tbody></table><div class="note"><p class="prefix"><b>Note:</b></p><p>In cases above which do not require reading the encoding declaration to
1519determine the encoding, section 4.3.3 still requires that the encoding declaration,
1520if present, be read and that the encoding name be checked to match the actual
1521encoding of the entity. Also, it is possible that new character encodings
1522will be invented that will make it necessary to use the encoding declaration
1523to determine the encoding, in cases where this is not required at present.</p></div><p>This level of autodetection is enough to read the XML encoding declaration
1524and parse the character-encoding identifier, which is still necessary to distinguish
1525the individual members of each family of encodings (e.g. to tell UTF-8 from
15268859, and the parts of 8859 from each other, or to distinguish the specific
1527EBCDIC code page in use, and so on).</p><p>Because the contents of the encoding declaration are restricted to characters
1528from the ASCII repertoire (however encoded),
1529a processor can reliably read the entire encoding declaration as soon as it
1530has detected which family of encodings is in use. Since in practice, all widely
1531used character encodings fall into one of the categories above, the XML encoding
1532declaration allows reasonably reliable in-band labeling of character encodings,
1533even when external sources of information at the operating-system or transport-protocol
1534level are unreliable. Character encodings such as UTF-7
1535that make overloaded usage of ASCII-valued bytes may fail to be reliably detected.</p><p>Once the processor has detected the character encoding in use, it can act
1536appropriately, whether by invoking a separate input routine for each case,
1537or by calling the proper conversion function on each character of input.</p><p>Like any self-labeling system, the XML encoding declaration will not work
1538if any software changes the entity's character set or encoding without updating
1539the encoding declaration. Implementors of character-encoding routines should
1540be careful to ensure the accuracy of the internal and external information
1541used to label the entity.</p></div><div class="div2"> <h3><a name="sec-guessing-with-ext-info" id="sec-guessing-with-ext-info" />E.2 Priorities in the Presence of External Encoding Information</h3><p>The second possible case occurs when the XML entity is accompanied by encoding
1542information, as in some file systems and some network protocols. When multiple
1543sources of information are available, their relative priority and the preferred
1544method of handling conflict should be specified as part of the higher-level
1545protocol used to deliver XML. In particular, please refer
1546to <a href="#rfc2376">[IETF RFC 3023]</a> or its successor, which defines the <code>text/xml</code>
1547and <code>application/xml</code> MIME types and provides some useful guidance.
1548In the interests of interoperability, however, the following rule is recommended.</p><ul><li><p>If an XML entity is in a file, the Byte-Order Mark and encoding declaration are used
1549(if present) to determine the character encoding.</p></li></ul></div></div><div class="div1"> <h2><a name="sec-xml-wg" id="sec-xml-wg" />F W3C XML Working Group (Non-Normative)</h2><p>This specification was prepared and approved for publication by the W3C
1550XML Working Group (WG). WG approval of this specification does not necessarily
1551imply that all WG participants voted for its approval. The current and former members
1552in the XML WG are:</p><ul><li>Jon Bosak, Sun (<i>Chair</i>) </li><li>James Clark (<i>Technical Lead</i>) </li><li>Tim Bray, Textuality and Netscape (<i>XML Co-editor</i>) </li><li>Jean Paoli, Microsoft (<i>XML
1553Co-editor</i>) </li><li>C. M. Sperberg-McQueen, U. of Ill. (<i>XML Co-editor</i>) </li><li>Dan Connolly, W3C (<i>W3C Liaison</i>) </li><li>Paula Angerstein, Texcel</li><li>Steve DeRose, INSO</li><li>Dave Hollander, HP</li><li>Eliot Kimber, ISOGEN</li><li>Eve Maler, ArborText</li><li>Tom Magliery, NCSA</li><li>Murray Maloney, SoftQuad, Grif
1554SA, Muzmo and Veo Systems</li><li>MURATA Makoto (FAMILY Given), Fuji
1555Xerox Information Systems</li><li>Joel Nava, Adobe</li><li>Conleth O'Connell, Vignette</li><li>Peter Sharpe, SoftQuad</li><li>John Tigue, DataChannel</li></ul></div><div class="div1"> <h2><a name="sec-core-wg" id="sec-core-wg" />G W3C XML Core <span>Working</span> Group (Non-Normative)</h2><p>The present edition of this specification was prepared by the W3C XML Core
1556Working Group (WG). The participants in the WG at the time of publication of this
1557edition were:</p><ul><li>Leonid Arbouzov, Sun Microsystems</li><li>Mary Brady</li><li>John Cowan (<i>XML 1.1 First Edition Editor</i>) </li><li>John Evdemon, Microsoft</li><li>Andrew Fang, Arbortext</li><li>Paul Grosso, Arbortext (<i>Co-Chair</i>) </li><li>Arnaud Le Hors, IBM</li><li>Dmitry Lenkov, Oracle</li><li>Anjana Manian, Oracle</li><li>Glenn Marcy, IBM</li><li>Jonathan Marsh, Microsoft</li><li>Sandra Martinez, NIST</li><li>Liam Quin, W3C (<i>Staff Contact</i>) </li><li>Lew Shannon</li><li>Richard Tobin, University of Edinburgh</li><li>Daniel Veillard</li><li>Norman Walsh, Sun Microsystems (<i>Co-Chair</i>) </li><li>François Yergeau</li></ul></div><div class="div1"> <h2><a name="prod-notes" id="prod-notes" />H Production Notes (Non-Normative)</h2><p>This edition was encoded in a
1558slightly modified version of the
1559<a href="http://www.w3.org/2002/xmlspec/dtd/2.5/xmlspec.dtd">XMLspec DTD, 2.5</a>.
1560The XHTML versions were produced with a combination of the
1561<a href="http://www.w3.org/2002/xmlspec/xhtml/1.9/xmlspec.xsl">xmlspec.xsl</a>,
1562<a href="http://www.w3.org/2002/xmlspec/xhtml/1.9/diffspec.xsl">diffspec.xsl</a>,
1563and <a href="REC-xml-3e.xsl">REC-xml-3e.xsl</a>
1564XSLT stylesheets.</p></div><div class="div1"> <h2><a name="sec-suggested-names" id="sec-suggested-names" />I Suggestions for XML Names (Non-Normative)</h2><p>The following suggestions define what is believed to be best
1565practice in the construction of XML names used as element names,
1566attribute names, processing instruction targets, entity names,
1567notation names, and the values of attributes of type ID, and are
1568intended as guidance for document authors and schema designers.
1569All references to Unicode are understood with respect to
1570a particular version of the Unicode Standard greater than or equal
1571to 3.0; which version should be used is left to the discretion of
1572the document author or schema designer.</p><p>The first two suggestions are directly derived from the rules
1573given for identifiers in the Unicode Standard, version 3.0, and
1574exclude all control characters, enclosing nonspacing marks,
1575non-decimal numbers, private-use characters, punctuation characters
1576(with the noted exceptions), symbol characters, unassigned
1577codepoints, and white space characters. The other suggestions
1578are mostly derived from <a href="#XML1.0">[XML-1.0]</a> Appendix B.</p><ol type="1"><li><p>The first character of any name should have a Unicode General
1579Category of Ll, Lu, Lo, Lm, Lt, or Nl, or else be '_' #x5F.</p></li><li><p>Characters other than the first should have a Unicode General
1580Category of Ll, Lu, Lo, Lm, Lt, Mc, Mn, Nl, Nd, Pc, or Cf, or else
1581be one of the following: '-' #x2D, '.' #x2E, ':' #x3A or
1582'·' #xB7 (middle dot). Since Cf characters are not
1583directly visible, they should be employed with caution and only
1584when necessary, to avoid creating names which are distinct to XML
1585processors but look the same to human beings.</p></li><li><p>Ideographic characters which have a canonical decomposition
1586(including those in the ranges [#xF900-#xFAFF] and
1587[#x2F800-#x2FFFD], with 12 exceptions) should not be used in names.
1588</p></li><li><p>Characters which have a compatibility decomposition (those with
1589a "compatibility formatting tag" in field 5 of the Unicode
1590Character Database -- marked by field 5 beginning with a "&lt;")
1591should not be used in names. This suggestion does not apply
1592to #x0E33 THAI CHARACTER SARA AM or #x0EB3 LAO CHARACTER AM, which
1593despite their compatibility decompositions are in regular use in
1594those scripts.</p></li><li><p>Combining characters meant for use with symbols only (including
1595those in the ranges [#x20D0-#x20EF] and [#x1D165-#x1D1AD]) should
1596not be used in names.</p></li><li><p>The interlinear annotation characters ([#xFFF9-#xFFFB) should
1597not be used in names.</p></li><li><p>Variation selector characters should not be used in names.</p></li><li><p>Names which are nonsensical, unpronounceable, hard to read, or
1598easily confusable with other names should not be employed.</p></li></ol></div></div></body></html>