source: draft-ietf-iri-3987bis/draft-ietf-iri-3987bis.xml @ 66

Last change on this file since 66 was 66, checked in by duerst@…, 9 years ago
  • added reason for punycode conversion
  • added an anchor attribute to a subsubsection
  • Property svn:executable set to *
File size: 123.7 KB
Line 
1<?xml version="1.0"?>
2<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
3<!ENTITY rfc1738 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.1738.xml">
4<!ENTITY rfc2045 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2045.xml">
5<!ENTITY rfc2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml">
6<!ENTITY rfc2130 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2130.xml">
7<!ENTITY rfc2141 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2141.xml">
8<!ENTITY rfc2192 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2192.xml">
9<!ENTITY rfc2277 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2277.xml">
10<!ENTITY rfc2368 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2368.xml">
11<!ENTITY rfc2384 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2384.xml">
12<!ENTITY rfc2396 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2396.xml">
13<!ENTITY rfc2397 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2397.xml">
14<!ENTITY rfc2616 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2616.xml">
15<!ENTITY rfc2640 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2640.xml">
16<!ENTITY rfc3490 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3490.xml">
17<!ENTITY rfc3491 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3491.xml">
18<!ENTITY rfc3629 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3629.xml">
19<!ENTITY rfc3986 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3986.xml">
20<!ENTITY rfc3987 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3987.xml">
21<!ENTITY rfc5890 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5890.xml">
22<!ENTITY rfc5891 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5891.xml">
23<!ENTITY rfc6055 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6055.xml">
24]>
25<?rfc strict='yes'?>
26
27<?xml-stylesheet type='text/css' href='rfc2629.css' ?>
28<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>
29<?rfc symrefs='yes'?>
30<?rfc sortrefs='yes'?>
31<?rfc iprnotified="no" ?>
32<?rfc toc='yes'?>
33<?rfc compact='yes'?>
34<?rfc subcompact='no'?>
35<rfc ipr="pre5378Trust200902" docName="draft-ietf-iri-3987bis-05" category="std" xml:lang="en" obsoletes="3987">
36<front>
37<title abbrev="IRIs">Internationalized Resource Identifiers (IRIs)</title>
38
39  <author initials="M.J." surname="Duerst" fullname='Martin Duerst'>
40    <!-- (Note: Please write "Duerst" with u-umlaut wherever
41      possible, for example as "D&#252;rst" in XML and HTML) -->
42  <organization abbrev="Aoyama Gakuin University">Aoyama Gakuin University</organization>
43  <address>
44  <postal>
45  <street>5-10-1 Fuchinobe</street>
46  <city>Sagamihara</city>
47  <region>Kanagawa</region>
48  <code>229-8558</code>
49  <country>Japan</country>
50  </postal>
51  <phone>+81 42 759 6329</phone>
52  <facsimile>+81 42 759 6495</facsimile>
53  <email>duerst@it.aoyama.ac.jp</email>
54  <uri>http://www.sw.it.aoyama.ac.jp/D%C3%BCrst/<!-- (Note: This is the percent-encoded form of an IRI)--></uri>
55  </address>
56</author>
57
58<author initials="M.L." surname="Suignard" fullname="Michel Suignard">
59   <organization>Unicode Consortium</organization>
60   <address>
61   <postal>
62   <street></street>
63   <street>P.O. Box 391476</street>
64   <city>Mountain View</city>
65   <region>CA</region>
66   <code>94039-1476</code>
67   <country>U.S.A.</country>
68   </postal>
69   <phone>+1-650-693-3921</phone>
70   <email>michel@unicode.org</email>
71   <uri>http://www.suignard.com</uri>
72   </address>
73</author>
74<author initials="L." surname="Masinter" fullname="Larry Masinter">
75   <organization>Adobe</organization>
76   <address>
77   <postal>
78   <street>345 Park Ave</street>
79   <city>San Jose</city>
80   <region>CA</region>
81   <code>95110</code>
82   <country>U.S.A.</country>
83   </postal>
84   <phone>+1-408-536-3024</phone>
85   <email>masinter@adobe.com</email>
86   <uri>http://larry.masinter.net</uri>
87   </address>
88</author>
89
90<date year="2011" month="March" day="29"/>
91<area>Applications</area>
92<workgroup>Internationalized Resource Identifiers (iri)</workgroup>
93<keyword>IRI</keyword>
94<keyword>Internationalized Resource Identifier</keyword>
95<keyword>UTF-8</keyword>
96<keyword>URI</keyword>
97<keyword>URL</keyword>
98<keyword>IDN</keyword>
99<keyword>LEIRI</keyword>
100
101<abstract>
102<t>This document defines the Internationalized Resource Identifier
103(IRI) protocol element, as an extension of the Uniform Resource
104Identifier (URI).  An IRI is a sequence of characters from the
105Universal Character Set (Unicode/ISO 10646). Grammar and processing
106rules are given for IRIs and related syntactic forms.</t>
107
108<t>In addition, this document provides named additional rule sets
109for processing otherwise invalid IRIs, in a way that supports
110other specifications that wish to mandate common behavior for
111'error' handling. In particular, rules used in some XML languages
112(LEIRI) and web applications are given.</t>
113
114<t>Defining IRI as new protocol element (rather than updating or
115extending the definition of URI) allows independent orderly
116transitions: other protocols and languages that use URIs must
117explicitly choose to allow IRIs.</t>
118
119<t>Guidelines are provided for the use and deployment of IRIs and
120related protocol elements when revising protocols, formats, and
121software components that currently deal only with URIs.</t>
122
123</abstract>
124  <note title='RFC Editor: Please remove the next paragraph before publication.'>
125    <t>This document is intended to update RFC 3987 and move towards IETF
126    Draft Standard.  For discussion and comments on this
127    draft, please join the IETF IRI WG by subscribing to the mailing
128    list public-iri@w3.org. For a list of open issues, please see
129    the issue tracker of the WG at http://trac.tools.ietf.org/wg/iri/trac/report/1.
130    For a list of individual edits, please see the change history at
131    http://trac.tools.ietf.org/wg/iri/trac/log/draft-ietf-iri-3987bis.</t>
132</note>
133</front>
134<middle>
135
136<section title="Introduction">
137
138<section title="Overview and Motivation" anchor="overview">
139
140<t>A Uniform Resource Identifier (URI) is defined in <xref
141target="RFC3986"/> as a sequence of characters chosen from a limited
142subset of the repertoire of US-ASCII <xref target="ASCII"/>
143characters.</t>
144
145<t>The characters in URIs are frequently used for representing words
146of natural languages.  This usage has many advantages: Such URIs are
147easier to memorize, easier to interpret, easier to transcribe, easier
148to create, and easier to guess. For most languages other than English,
149however, the natural script uses characters other than A - Z. For many
150people, handling Latin characters is as difficult as handling the
151characters of other scripts is for those who use only the Latin
152alphabet. Many languages with non-Latin scripts are transcribed with
153Latin letters. These transcriptions are now often used in URIs, but
154they introduce additional difficulties.</t>
155
156<t>The infrastructure for the appropriate handling of characters from
157additional scripts is now widely deployed in operating system and
158application software. Software that can handle a wide variety of
159scripts and languages at the same time is increasingly common. Also,
160an increasing number of protocols and formats can carry a wide range of
161characters.</t>
162
163<t>URIs are used both as a protocol element (for transmission and
164processing by software) and also a presentation element (for display
165and handling by people who read, interpret, coin, or guess them). The
166transition between these roles is more difficult and complex when
167dealing with the larger set of characters than allowed for URIs in
168<xref target="RFC3986"/>. </t>
169
170<t>This document defines the protocol element called Internationalized
171Resource Identifier (IRI), which allow applications of URIs to be
172extended to use resource identifiers that have a much wider repertoire
173of characters. It also provides corresponding "internationalized"
174versions of other constructs from <xref target="RFC3986"/>, such as
175URI references. The syntax of IRIs is defined in <xref
176target="syntax"/>.
177</t>
178
179<t>Using characters outside of A - Z in IRIs adds a number of
180difficulties. <xref target="Bidi"/> discusses the special case of
181bidirectional IRIs using characters from scripts written
182right-to-left.  <xref target="equivalence"/> discusses various forms
183of equivalence between IRIs. <xref target="IRIuse"/> discusses the use
184of IRIs in different situations.  <xref target="guidelines"/> gives
185additional informative guidelines.  <xref target="security"/>
186discusses IRI-specific security considerations.</t>
187
188  <t>When originally defining IRIs, several design alternatives were considered.
189    Historically interested readers can find an overview in Appendix A of <xref target="RFC3987"/>.
190  For some additional background on the design of URIs and IRIs, please also see
191    <xref target="Gettys"/>.</t>
192</section> <!-- overview -->
193
194<section title="Applicability" anchor="Applicability">
195
196<t>IRIs are designed to allow protocols and software that deal with
197URIs to be updated to handle IRIs. A "URI scheme" (as defined by <xref
198target="RFC3986"/> and registered through the IANA process defined in
199<xref target="RFC4395bis"/> also serves as an "IRI scheme". Processing of
200IRIs is accomplished by extending the URI syntax while retaining (and
201not expanding) the set of "reserved" characters, such that the syntax
202for any URI scheme may be uniformly extended to allow non-ASCII
203characters. In addition, following parsing of an IRI, it is possible
204to construct a corresponding URI by first encoding characters outside
205of the allowed URI range and then reassembling the components.
206</t>
207
208<t>Practical use of IRIs forms in place of URIs forms depends on the
209following conditions being met:</t>
210
211<t><list style="hanging">
212   
213<t hangText="a.">A protocol or format element MUST be explicitly designated to be
214  able to carry IRIs. The intent is to avoid introducing IRIs into
215  contexts that are not defined to accept them.  For example, XML
216  schema <xref target="XMLSchema"/> has an explicit type "anyURI" that
217  includes IRIs and IRI references. Therefore, IRIs and IRI references
218  can be in attributes and elements of type "anyURI".  On the other
219  hand, in the <xref target="RFC2616"/> definition of HTTP/1.1, the
220  Request URI is defined as a URI, which means that direct use of IRIs
221  is not allowed in HTTP requests.</t>
222
223<t hangText="b.">The protocol or format carrying the IRIs MUST have a
224  mechanism to represent the wide range of characters used in IRIs,
225  either natively or by some protocol- or format-specific escaping
226  mechanism (for example, numeric character references in <xref
227  target="XML1"/>).</t>
228
229<t hangText="c.">The URI scheme definition, if it explicitly allows a
230  percent sign ("%") in any syntactic component, SHOULD define the
231  interpretation of sequences of percent-encoded octets (using "%XX"
232  hex octets) as octet from sequences of UTF-8 encoded strings; this
233  is recommended in the guidelines for registering new schemes, <xref
234  target="RFC4395bis"/>.  For example, this is the practice for IMAP URLs
235  <xref target="RFC2192"/>, POP URLs <xref target="RFC2384"/> and the
236  URN syntax <xref target="RFC2141"/>). Note that use of
237  percent-encoding may also be restricted in some situations, for
238  example, URI schemes that disallow percent-encoding might still be
239  used with a fragment identifier which is percent-encoded (e.g.,
240  <xref target="XPointer"/>). See <xref target="UTF8use"/> for further
241  discussion.</t>
242</list></t>
243
244</section> <!-- applicability -->
245
246<section title="Definitions" anchor="sec-Definitions">
247 
248<t>The following definitions are used in this document; they follow the
249terms in <xref target="RFC2130"/>, <xref target="RFC2277"/>, and
250<xref target="ISO10646"/>.</t>
251<t><list style="hanging">
252   
253<t hangText="character:">A member of a set of elements used for the
254    organization, control, or representation of data. For example,
255    "LATIN CAPITAL LETTER A" names a character.</t>
256   
257<t hangText="octet:">An ordered sequence of eight bits considered as a
258    unit.</t>
259   
260<t hangText="character repertoire:">A set of characters (set in the
261    mathematical sense).</t>
262   
263<t hangText="sequence of characters:">A sequence of characters (one
264    after another).</t>
265   
266<t hangText="sequence of octets:">A sequence of octets (one after
267    another).</t>
268   
269<t hangText="character encoding:">A method of representing a sequence
270    of characters as a sequence of octets (maybe with variants). Also,
271    a method of (unambiguously) converting a sequence of octets into a
272    sequence of characters.</t>
273   
274<t hangText="charset:">The name of a parameter or attribute used to
275    identify a character encoding.</t>
276   
277<t hangText="UCS:">Universal Character Set. The coded character set
278    defined by ISO/IEC 10646 <xref target="ISO10646"/> and the Unicode
279    Standard <xref target="UNIV6"/>.</t>
280   
281<t hangText="IRI reference:">Denotes the common usage of an
282    Internationalized Resource Identifier. An IRI reference may be
283    absolute or relative.  However, the "IRI" that results from such a
284    reference only includes absolute IRIs; any relative IRI references
285    are resolved to their absolute form.  Note that in <xref
286    target="RFC2396"/> URIs did not include fragment identifiers, but
287    in <xref target="RFC3986"/> fragment identifiers are part of
288    URIs.</t>
289   
290<t hangText="URL:">The term "URL" was originally used <xref
291   target="RFC1738"/> for roughly what is now called a "URI".  Books,
292   software and documentation often refers to URIs and IRIs using the
293   "URL" term. Some usages restrict "URL" to those URIs which are not
294   URNs. Because of the ambiguity of the term using the term "URL" is
295   NOT RECOMMENDED in formal documents.</t>
296
297<t hangText="LEIRI (Legacy Extended IRI) processing:">  This term was used in
298   various XML specifications to refer
299   to strings that, although not valid IRIs, were acceptable input to
300   the processing rules in <xref target="LEIRIspec" />.</t>
301
302<t hangText="(Web Address, Hypertext Reference, HREF):"> These terms have been
303   added in this document for convenience, to allow other
304   specifications to refer to those strings that, although not valid
305   IRIs, are acceptable input to the processing rules in <xref
306   target="webaddress"/>. This usage corresponds to the parsing rules
307   of some popular web browsing applications.
308   ISSUE: Need to find a good name/abbreviation for these.</t>
309   
310<t hangText="running text:">Human text (paragraphs, sentences,
311   phrases) with syntax according to orthographic conventions of a
312   natural language, as opposed to syntax defined for ease of
313   processing by machines (e.g., markup, programming languages).</t>
314   
315<t hangText="protocol element:">Any portion of a message that affects
316    processing of that message by the protocol in question.</t>
317   
318<t hangText="presentation element:">A presentation form corresponding
319    to a protocol element; for example, using a wider range of
320    characters.</t>
321   
322<t hangText="create (a URI or IRI):">With respect to URIs and IRIs,
323     the term is used for the initial creation. This may be the
324     initial creation of a resource with a certain identifier, or the
325     initial exposition of a resource under a particular
326     identifier.</t>
327   
328<t hangText="generate (a URI or IRI):">With respect to URIs and IRIs,
329     the term is used when the identifier is generated by derivation
330     from other information.</t>
331
332<t hangText="parsed URI component:">When a URI processor parses a URI
333   (following the generic syntax or a scheme-specific syntax, the result
334   is a set of parsed URI components, each of which has a type
335   (corresponding to the syntactic definition) and a sequence of URI
336   characters.  </t>
337
338<t hangText="parsed IRI component:">When an IRI processor parses
339   an IRI directly, following the general syntax or a scheme-specific
340   syntax, the result is a set of parsed IRI components, each of
341   which has a type (corresponding to the syntactice definition)
342   and a sequence of IRI characters. (This definition is analogous
343   to "parsed URI component".)</t>
344
345<t hangText="IRI scheme:">A URI scheme may also be known as
346   an "IRI scheme" if the scheme's syntax has been extended to
347   allow non-US-ASCII characters according to the rules in this
348   document.</t>
349
350</list></t>
351</section> <!-- definitions -->
352<section title="Notation" anchor="sec-Notation">
353     
354<t>RFCs and Internet Drafts currently do not allow any characters
355outside the US-ASCII repertoire. Therefore, this document uses various
356special notations to denote such characters in examples.</t>
357     
358<t>In text, characters outside US-ASCII are sometimes referenced by
359using a prefix of 'U+', followed by four to six hexadecimal
360digits.</t>
361
362<t>To represent characters outside US-ASCII in examples, this document
363uses two notations: 'XML Notation' and 'Bidi Notation'.</t>
364
365<t>XML Notation uses a leading '&amp;#x', a trailing ';', and the
366hexadecimal number of the character in the UCS in between. For
367example, &amp;#x44F; stands for CYRILLIC CAPITAL LETTER YA. In this
368notation, an actual '&amp;' is denoted by '&amp;amp;'.</t>
369
370<t>Bidi Notation is used for bidirectional examples: Lower case
371letters stand for Latin letters or other letters that are written left
372to right, whereas upper case letters represent Arabic or Hebrew
373letters that are written right to left.</t>
374
375<t>To denote actual octets in examples (as opposed to percent-encoded
376octets), the two hex digits denoting the octet are enclosed in "&lt;"
377and "&gt;".  For example, the octet often denoted as 0xc9 is denoted
378here as &lt;c9&gt;.</t>
379
380<t> In this document, the key words "MUST", "MUST NOT", "REQUIRED",
381"SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY",
382and "OPTIONAL" are to be interpreted as described in <xref
383target="RFC2119"/>.</t>
384
385</section> <!-- notation -->
386</section> <!-- introduction -->
387
388<section title="IRI Syntax" anchor="syntax">
389<t>This section defines the syntax of Internationalized Resource
390Identifiers (IRIs).</t>
391
392<t>As with URIs, an IRI is defined as a sequence of characters, not as
393a sequence of octets. This definition accommodates the fact that IRIs
394may be written on paper or read over the radio as well as stored or
395transmitted digitally.  The same IRI might be represented as different
396sequences of octets in different protocols or documents if these
397protocols or documents use different character encodings (and/or
398transfer encodings).  Using the same character encoding as the
399containing protocol or document ensures that the characters in the IRI
400can be handled (e.g., searched, converted, displayed) in the same way
401as the rest of the protocol or document.</t>
402
403<section title="Summary of IRI Syntax" anchor="summary">
404
405<t>IRIs are defined by extending the URI syntax in <xref
406target="RFC3986"/>, but extending the class of unreserved characters
407by adding the characters of the UCS (Universal Character Set, <xref
408target="ISO10646"/>) beyond U+007F, subject to the limitations given
409in the syntax rules below and in <xref target="limitations"/>.</t>
410
411<t>The syntax and use of components and reserved characters is the
412same as that in <xref target="RFC3986"/>. Each "URI scheme" thus also
413functions as an "IRI scheme", in that scheme-specific parsing rules
414for URIs of a scheme are be extended to allow parsing of IRIs using
415the same parsing rules.</t>
416
417<t>All the operations defined in <xref target="RFC3986"/>, such as the
418resolution of relative references, can be applied to IRIs by
419IRI-processing software in exactly the same way as they are for URIs
420by URI-processing software.</t>
421
422<t>Characters outside the US-ASCII repertoire MUST NOT be reserved and
423therefore MUST NOT be used for syntactical purposes, such as to
424delimit components in newly defined schemes. For example, U+00A2, CENT
425SIGN, is not allowed as a delimiter in IRIs, because it is in the
426'iunreserved' category. This is similar to the fact that it is not
427possible to use '-' as a delimiter in URIs, because it is in the
428'unreserved' category.</t>
429
430</section> <!-- summary -->
431<section title="ABNF for IRI References and IRIs" anchor="abnf">
432
433<t>An ABNF definition for IRI references (which are the most general
434concept and the start of the grammar) and IRIs is given here. The
435syntax of this ABNF is described in <xref target="STD68"/>. Character
436numbers are taken from the UCS, without implying any actual binary
437encoding. Terminals in the ABNF are characters, not octets.</t>
438
439<t>The following grammar closely follows the URI grammar in <xref
440target="RFC3986"/>, except that the range of unreserved characters is
441expanded to include UCS characters, with the restriction that private
442UCS characters can occur only in query parts. The grammar is split
443into two parts: Rules that differ from <xref target="RFC3986"/>
444because of the above-mentioned expansion, and rules that are the same
445as those in <xref target="RFC3986"/>. For rules that are different
446than those in <xref target="RFC3986"/>, the names of the non-terminals
447have been changed as follows. If the non-terminal contains 'URI', this
448has been changed to 'IRI'. Otherwise, an 'i' has been prefixed.
449The rule &lt;pct-form> has been introduced in order to be able to reference it from other parts of
450the document.</t>
451
452<!--
453for line length measuring in artwork (max 72 chars, three chars at start):
454      1         2         3         4         5         6         7
455456789012345678901234567890123456789012345678901234567890123456789012
456-->
457<figure>
458<preamble>The following rules are different from those in <xref target="RFC3986"/>:</preamble>
459<artwork>
460IRI            = scheme ":" ihier-part [ "?" iquery ]
461                 [ "#" ifragment ]
462
463ihier-part     = "//" iauthority ipath-abempty
464               / ipath-absolute
465               / ipath-rootless
466               / ipath-empty
467
468IRI-reference  = IRI / irelative-ref
469
470absolute-IRI   = scheme ":" ihier-part [ "?" iquery ]
471
472irelative-ref  = irelative-part [ "?" iquery ] [ "#" ifragment ]
473
474irelative-part = "//" iauthority ipath-abempty
475               / ipath-absolute
476               / ipath-noscheme
477               / ipath-empty
478
479iauthority     = [ iuserinfo "@" ] ihost [ ":" port ]
480iuserinfo      = *( iunreserved / pct-form / sub-delims / ":" )
481ihost          = IP-literal / IPv4address / ireg-name
482
483pct-form       = pct-encoded
484
485ireg-name      = *( iunreserved / sub-delims )
486
487ipath          = ipath-abempty   ; begins with "/" or is empty
488               / ipath-absolute  ; begins with "/" but not "//"
489               / ipath-noscheme  ; begins with a non-colon segment
490               / ipath-rootless  ; begins with a segment
491               / ipath-empty     ; zero characters
492
493ipath-abempty  = *( path-sep isegment )
494ipath-absolute = path-sep [ isegment-nz *( path-sep isegment ) ]
495ipath-noscheme = isegment-nz-nc *( path-sep isegment )
496ipath-rootless = isegment-nz *( path-sep isegment )
497ipath-empty    = 0&lt;ipchar&gt;
498path-sep       = "/"
499
500isegment       = *ipchar
501isegment-nz    = 1*ipchar
502isegment-nz-nc = 1*( iunreserved / pct-form / sub-delims
503                     / "@" )
504               ; non-zero-length segment without any colon ":"                     
505
506ipchar         = iunreserved / pct-form / sub-delims / ":"
507               / "@"
508 
509iquery         = *( ipchar / iprivate / "/" / "?" )
510
511ifragment      = *( ipchar / "/" / "?" )
512
513iunreserved    = ALPHA / DIGIT / "-" / "." / "_" / "~" / ucschar
514
515ucschar        = %xA0-D7FF / %xF900-FDCF / %xFDF0-FFEF
516               / %x10000-1FFFD / %x20000-2FFFD / %x30000-3FFFD
517               / %x40000-4FFFD / %x50000-5FFFD / %x60000-6FFFD
518               / %x70000-7FFFD / %x80000-8FFFD / %x90000-9FFFD
519               / %xA0000-AFFFD / %xB0000-BFFFD / %xC0000-CFFFD
520               / %xD0000-DFFFD / %xE1000-EFFFD
521
522iprivate       = %xE000-F8FF / %xE0000-E0FFF / %xF0000-FFFFD
523               / %x100000-10FFFD
524</artwork>
525</figure>
526
527<t>Some productions are ambiguous. The "first-match-wins" (a.k.a. "greedy")
528algorithm applies. For details, see <xref target="RFC3986"/>.</t>
529
530<figure>
531<preamble>The following rules are the same as those in <xref target="RFC3986"/>:</preamble>
532<artwork>
533scheme         = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
534 
535port           = *DIGIT
536 
537IP-literal     = "[" ( IPv6address / IPvFuture  ) "]"
538 
539IPvFuture      = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" )
540 
541IPv6address    =                            6( h16 ":" ) ls32
542               /                       "::" 5( h16 ":" ) ls32
543               / [               h16 ] "::" 4( h16 ":" ) ls32
544               / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
545               / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
546               / [ *3( h16 ":" ) h16 ] "::"    h16 ":"   ls32
547               / [ *4( h16 ":" ) h16 ] "::"              ls32
548               / [ *5( h16 ":" ) h16 ] "::"              h16
549               / [ *6( h16 ":" ) h16 ] "::"
550               
551h16            = 1*4HEXDIG
552ls32           = ( h16 ":" h16 ) / IPv4address
553
554IPv4address    = dec-octet "." dec-octet "." dec-octet "." dec-octet
555
556dec-octet      = DIGIT                 ; 0-9
557               / %x31-39 DIGIT         ; 10-99
558               / "1" 2DIGIT            ; 100-199
559               / "2" %x30-34 DIGIT     ; 200-249
560               / "25" %x30-35          ; 250-255
561           
562pct-encoded    = "%" HEXDIG HEXDIG
563
564unreserved     = ALPHA / DIGIT / "-" / "." / "_" / "~"
565reserved       = gen-delims / sub-delims
566gen-delims     = ":" / "/" / "?" / "#" / "[" / "]" / "@"
567sub-delims     = "!" / "$" / "&amp;" / "'" / "(" / ")"
568               / "*" / "+" / "," / ";" / "="
569</artwork></figure>
570
571<t>This syntax does not support IPv6 scoped addressing zone identifiers.</t>
572
573</section> <!-- abnf -->
574
575</section> <!-- syntax -->
576
577<section title="Processing IRIs and related protocol elements" anchor="processing">
578
579<t>IRIs are meant to replace URIs in identifying resources within new
580versions of protocols, formats, and software components that use a
581UCS-based character repertoire.  Protocols and components may use and
582process IRIs directly. However, there are still numerous systems and
583protocols which only accept URIs or components of parsed URIs; that is,
584they only accept sequences of characters within the subset of US-ASCII
585characters allowed in URIs. </t>
586
587<t>This section defines specific processing steps for IRI consumers
588which establish the relationship between the string given and the
589interpreted derivatives. These
590processing steps apply to both IRIs and IRI references (i.e., absolute
591or relative forms); for IRIs, some steps are scheme specific. </t>
592
593<section title="Converting to UCS" anchor="ucsconv"> 
594 
595<t>Input that is already in a Unicode form (i.e., a sequence of Unicode
596 characters or an octet-stream representing a Unicode-based character
597 encoding such as UTF-8 or UTF-16) should be left as is and not
598 normalized (see <xref target="normalization"/>).</t>
599
600  <t>An IRI or IRI reference is a sequence of characters from the UCS.
601    For IRIs that are not already in a Unicode form
602    (as when written on paper, read aloud, or represented in a text stream
603    using a legacy character encoding), convert the IRI to Unicode.
604    Note that some character encodings or transcriptions can be converted
605    to or represented by more than one sequence of Unicode characters.
606    Ideally the resulting IRI would use a normalized form,
607    such as Unicode Normalization Form C <xref target="UTR15"/>
608    (see <xref target='ladder'/> Normalization and Comparison),
609    since that ensures a stable, consistent representation
610    that is most likely to produce the intended results.
611    Implementers and users are cautioned that, while denormalized character sequences are valid,
612    they might be difficult for other users or processes to reproduce
613    and might lead to unexpected results.
614  </t>
615
616<t> In other cases (written on paper, read aloud, or otherwise
617 represented independent of any character encoding) represent the IRI
618 as a sequence of characters from the UCS normalized according to
619 Unicode Normalization Form C (NFC, <xref target="UTR15"/>).</t>
620</section> <!-- ucsconv -->
621
622<section title="Parse the IRI into IRI components">
623
624<t>Parse the IRI, either as a relative reference (no scheme)
625or using scheme specific processing (according to the scheme
626given); the result is a set of parsed IRI components.</t>
627
628<t>NOTE: The result of parsing into components will correspond
629to subtrings of the IRI that may be accessible via an API.
630For example, in <xref target="HTML5"/>, the protocol
631components of interest are SCHEME (scheme), HOST (ireg-name), PORT
632(port), the PATH (ipath after the initial "/"), QUERY (iquery),
633FRAGMENT (ifragment), and AUTHORITY (iauthority).
634</t>
635
636<t>Subsequent processing rules are sometimes used to define other
637syntactic components. For example, <xref target="HTML5"/> defines APIs
638for IRI processing; in these APIs:
639
640<list style="hanging">
641<t hangText="HOSTSPECIFIC"> the substring that follows
642the substring matched by the iauthority production, or the whole
643string if the iauthority production wasn't matched.</t>
644<t hangText="HOSTPORT"> if there is a scheme component and a port
645component and the port given by the port component is different than
646the default port defined for the protocol given by the scheme
647component, then HOSTPORT is the substring that starts with the
648substring matched by the host production and ends with the substring
649matched by the port production, and includes the colon in between the
650two. Otherwise, it is the same as the host component.
651</t>
652</list>
653</t>
654</section> <!-- parse -->
655
656<section title="General percent-encoding of IRI components" anchor="compmapping">
657   
658<t>Except as noted in the following subsections, IRI components are mapped
659to the equivalent URI components by percent-encoding those characters
660not allowed in URIs. Previous processing steps will have removed
661some characters, and the interpretation of reserved characters will
662have already been done (with the syntactic reserved characters outside
663of the IRI component). This mapping is defined for all sequences
664of Unicode characters, whether or not they are valid for the component
665in question. </t>
666   
667<t>For each character which is not allowed anywhere in a valid URI, apply the following steps. </t>
668
669<t><list style="hanging">
670
671<t hangText="Convert to UTF-8">Convert the character to a sequence of
672  one or more octets using UTF-8 <xref target="RFC3629"/>.</t>
673
674<t hangText="Percent encode">Convert each octet of this sequence to %HH,
675   where HH is the hexadecimal notation of the octet value. The
676   hexadecimal notation SHOULD use uppercase letters. (This is the
677   general URI percent-encoding mechanism in Section 2.1 of <xref
678   target="RFC3986"/>.)</t>
679   
680</list></t>
681
682<t>Note that the mapping is an identity transformation for parsed URI
683components of valid URIs, and is idempotent: applying the mapping a
684second time will not change anything.</t>
685</section> <!-- general conversion -->
686
687<section title="Mapping ireg-name" anchor="dnsmapping">
688  <section title='Mapping using Percent-Encoding' anchor='dnspercent'>
689  <t>The ireg-name component SHOULD be converted
690    according to the general procedure for percent-encoding
691    of IRI components described in <xref target="compmapping"/>.</t>
692 
693  <t>For example, the IRI
694    <vspace/>"http://r&amp;#xE9;sum&amp;#xE9;.example.org"<vspace/> will be
695    converted to <vspace/>"http://r%C3%A9sum%C3%A9.example.org".</t>
696
697  <t>This conversion for ireg-name is in line with Section 3.2.2
698    of <xref target="RFC3986"/>, which does not mandate
699    a particular registered name lookup technology. For further background,
700    see <xref target="RFC6055"/> and <xref target="Gettys"/>.</t>
701</section>
702<section title="Mapping using Punycode" anchor='dnspunycode'>
703  <t>The ireg-name component MAY also be converted as follows:</t>
704
705<t>Replace the ireg-name part of the IRI by the part converted using
706the Domain Name Lookup procedure (Subsections 5.3 to 5.5) of <xref target="RFC5891"/>.
707  on each dot-separated label, and by using U+002E
708(FULL STOP) as a label separator.
709This procedure may fail, but this would mean that the IRI cannot be resolved.
710In such cases, if the domain name conversion fails, then the
711entire IRI conversion fails. Processors that have no mechanism for
712signalling a failure MAY instead substitute an otherwise
713invalid host name, although such processing SHOULD be avoided.</t>
714
715<t>For example, the IRI
716<vspace/>"http://r&amp;#xE9;sum&amp;#xE9;.example.org"<vspace/> MAY be
717converted to <vspace/>"http://xn--rsum-bad.example.org"<vspace/>.</t>
718  <t>This conversion for ireg-name will be better able to deal with legacy
719    infrastructure that cannot handle percent-encoding in domain names.</t>
720</section>
721  <section title="Additional Considerations">
722<t><list style="hanging"> 
723
724<t hangText="Note:">Domain Names may appear in parts of an IRI other
725than the ireg-name part.  It is the responsibility of scheme-specific
726implementations (if the Internationalized Domain Name is part of the
727scheme syntax) or of server-side implementations (if the
728Internationalized Domain Name is part of 'iquery') to apply the
729necessary conversions at the appropriate point. Example: Trying to
730validate the Web page at<vspace/>
731http://r&amp;#xE9;sum&amp;#xE9;.example.org would lead to an IRI of
732<vspace/>http://validator.w3.org/check?uri=http%3A%2F%2Fr&amp;#xE9;sum&amp;#xE9;.<vspace/>example.org,
733which would convert to a URI
734of<vspace/>http://validator.w3.org/check?uri=http%3A%2F%2Fr%C3%A9sum%C3%A9.<vspace/>example.org.
735The server-side implementation is responsible for making the
736necessary conversions to be able to retrieve the Web page.</t>
737
738<t hangText="Note:">In this process, characters allowed in URI
739references and existing percent-encoded sequences are not encoded further.
740(This mapping is similar to, but different from, the encoding applied
741when arbitrary content is included in some part of a URI.)
742
743For example, an IRI of
744<vspace/>"http://www.example.org/red%09ros&amp;#xE9;#red"
745(in XML notation) is converted to
746<vspace/>"http://www.example.org/red%09ros%C3%A9#red", not to
747something like
748<vspace/>"http%3A%2F%2Fwww.example.org%2Fred%2509ros%C3%A9%23red".
749((DESIGN QUESTION: What about e.g. http://r%C3%A9sum%C3%A9.example.org in an IRI? Will that get converted to punycode, or not?))
750
751</t>
752
753</list></t>
754</section>
755</section> <!-- dnsmapping -->
756
757<section title="Mapping query components" anchor="querymapping">
758
759<t>((NOTE: SEE ISSUES LIST))
760
761For compatibility with existing deployed HTTP infrastructure,
762the following special case applies for schemes "http" and "https"
763and IRIs whose origin has a document charset other than one which
764is UCS-based (e.g., UTF-8 or UTF-16). In such a case, the "query"
765component of an IRI is mapped into a URI by using the document
766charset rather than UTF-8 as the binary representation before
767pct-encoding. This mapping is not applied for any other scheme
768or component.</t>
769
770</section> <!-- querymapping -->
771
772<section title="Mapping IRIs to URIs" anchor="mapping">
773
774<t>The canonical mapping from a IRI to URI is defined by applying the
775mapping above (from IRI to URI components) and then reassembling a URI
776from the parsed URI components using the original punctuation that
777delimited the IRI components. </t>
778
779</section> <!-- mapping -->
780
781<section title="Converting URIs to IRIs" anchor="URItoIRI">
782
783<t>In some situations, for presentation and further processing,
784it is desirable to convert a URI into an equivalent IRI in which
785natural characters are represented directly rather than
786percent encoded. Of course, every URI is already an IRI in
787its own right without any conversion, and in general there
788This section gives one such procedure for this conversion.
789</t>
790
791<t>
792The conversion described in this section, if given a valid URI, will
793result in an IRI that maps back to the URI used as an input for the
794conversion (except for potential case differences in percent-encoding
795and for potential percent-encoded unreserved characters).
796
797However, the IRI resulting from this conversion may differ
798from the original IRI (if there ever was one).</t> 
799
800<t>URI-to-IRI conversion removes percent-encodings, but not all
801percent-encodings can be eliminated. There are several reasons for
802this:</t>
803
804<t><list style="hanging">
805
806<t hangText="1.">Some percent-encodings are necessary to distinguish
807    percent-encoded and unencoded uses of reserved characters.</t>
808
809<t hangText="2.">Some percent-encodings cannot be interpreted as sequences
810    of UTF-8 octets.<vspace blankLines="1"/>
811    (Note: The octet patterns of UTF-8 are highly regular.
812    Therefore, there is a very high probability, but no guarantee,
813    that percent-encodings that can be interpreted as sequences of UTF-8
814    octets actually originated from UTF-8. For a detailed discussion,
815    see <xref target="Duerst97"/>.)</t>
816
817<t hangText="3.">The conversion may result in a character that is not
818    appropriate in an IRI. See <xref target="abnf"/>, <xref target="visual"/>,
819      and <xref target="limitations"/> for further details.</t>
820
821<t hangText="4.">IRI to URI conversion has different rules for
822    dealing with domain names and query parameters.</t>
823
824</list></t>
825
826<t>Conversion from a URI to an IRI MAY be done by using the following
827steps:
828
829<list style="hanging">
830<t hangText="1.">Represent the URI as a sequence of octets in
831       US-ASCII.</t>
832
833<t hangText="2.">Convert all percent-encodings ("%" followed by two
834      hexadecimal digits) to the corresponding octets, except those
835      corresponding to "%", characters in "reserved", and characters
836      in US-ASCII not allowed in URIs.</t> 
837
838<t hangText="3.">Re-percent-encode any octet produced in step 2 that
839      is not part of a strictly legal UTF-8 octet sequence.</t>
840
841
842<t hangText="4.">Re-percent-encode all octets produced in step 3 that
843      in UTF-8 represent characters that are not appropriate according
844      to <xref target="abnf"/>, <xref target="visual"/>, and <xref
845      target="limitations"/>.</t> 
846
847<t hangText="5.">Interpret the resulting octet sequence as a sequence
848      of characters encoded in UTF-8.</t>
849
850<t hangText="6.">URIs known to contain domain names in the reg-name
851      component SHOULD convert punycode-encoded domain name labels to
852      the corresponding characters using the ToUnicode procedure. </t>
853</list></t>
854
855<t>This procedure will convert as many percent-encoded characters as
856possible to characters in an IRI. Because there are some choices when
857step 4 is applied (see <xref target="limitations"/>), results may
858vary.</t>
859
860<t>Conversions from URIs to IRIs MUST NOT use any character
861encoding other than UTF-8 in steps 3 and 4, even if it might be
862possible to guess from the context that another character encoding
863than UTF-8 was used in the URI.  For example, the URI
864"http://www.example.org/r%E9sum%E9.html" might with some guessing be
865interpreted to contain two e-acute characters encoded as
866iso-8859-1. It must not be converted to an IRI containing these
867e-acute characters. Otherwise, in the future the IRI will be mapped to
868"http://www.example.org/r%C3%A9sum%C3%A9.html", which is a different
869URI from "http://www.example.org/r%E9sum%E9.html".</t>
870
871<section title="Examples">
872
873<t>This section shows various examples of converting URIs to IRIs.
874Each example shows the result after each of the steps 1 through 6 is
875applied. XML Notation is used for the final result.  Octets are
876denoted by "&lt;" followed by two hexadecimal digits followed by
877"&gt;".</t>
878
879<t>The following example contains the sequence "%C3%BC", which is a
880strictly legal UTF-8 sequence, and which is converted into the actual
881character U+00FC, LATIN SMALL LETTER U WITH DIAERESIS (also known as
882u-umlaut).
883
884<list style="hanging">
885<t hangText="1.">http://www.example.org/D%C3%BCrst</t>
886<t hangText="2.">http://www.example.org/D&lt;c3&gt;&lt;bc&gt;rst</t>
887<t hangText="3.">http://www.example.org/D&lt;c3&gt;&lt;bc&gt;rst</t>
888<t hangText="4.">http://www.example.org/D&lt;c3&gt;&lt;bc&gt;rst</t>
889<t hangText="5.">http://www.example.org/D&amp;#xFC;rst</t>
890<t hangText="6.">http://www.example.org/D&amp;#xFC;rst</t>
891</list>
892</t>
893
894<t>The following example contains the sequence "%FC", which might
895represent U+00FC, LATIN SMALL LETTER U WITH DIAERESIS, in
896the<vspace/>iso-8859-1 character encoding.  (It might represent other
897characters in other character encodings. For example, the octet
898&lt;fc&gt; in iso-8859-5 represents U+045C, CYRILLIC SMALL LETTER
899KJE.)  Because &lt;fc&gt; is not part of a strictly legal UTF-8
900sequence, it is re-percent-encoded in step 3.
901
902
903<list style="hanging">
904<t hangText="1.">http://www.example.org/D%FCrst</t>
905<t hangText="2.">http://www.example.org/D&lt;fc&gt;rst</t>
906<t hangText="3.">http://www.example.org/D%FCrst</t>
907<t hangText="4.">http://www.example.org/D%FCrst</t>
908<t hangText="5.">http://www.example.org/D%FCrst</t>
909<t hangText="6.">http://www.example.org/D%FCrst</t>
910</list>
911</t>
912
913<t>The following example contains "%e2%80%ae", which is the percent-encoded<vspace/>UTF-8
914character encoding of U+202E, RIGHT-TO-LEFT OVERRIDE. <xref target="visual"/>
915forbids the direct use of this character in an IRI. Therefore, the
916corresponding octets are re-percent-encoded in step 4. This example shows
917that the case (upper- or lowercase) of letters used in percent-encodings may not be preserved.
918The example also contains a punycode-encoded domain name label (xn--99zt52a),
919which is not converted.
920
921<list style="hanging">
922<t hangText="1.">http://xn--99zt52a.example.org/%e2%80%ae</t>
923<t hangText="2.">http://xn--99zt52a.example.org/&lt;e2&gt;&lt;80&gt;&lt;ae&gt;</t>
924<t hangText="3.">http://xn--99zt52a.example.org/&lt;e2&gt;&lt;80&gt;&lt;ae&gt;</t>
925<t hangText="4.">http://xn--99zt52a.example.org/%E2%80%AE</t>
926<t hangText="5.">http://xn--99zt52a.example.org/%E2%80%AE</t>
927<t hangText="6.">http://&amp;#x7D0D;&amp;#x8C46;.example.org/%E2%80%AE</t>
928</list></t>
929
930<t>Note that the label "xn--99zt52a" is converted to U+7D0D U+8C46
931(Japanese Natto). ((EDITOR NOTE: There is some inconsistency in this note.))</t>
932
933</section> <!-- examples -->
934</section> <!-- URItoIRI -->
935</section> <!-- processing -->
936<section title="Bidirectional IRIs for Right-to-Left Languages" anchor="Bidi">
937
938<t>Some UCS characters, such as those used in the Arabic and Hebrew
939scripts, have an inherent right-to-left (rtl) writing direction. IRIs
940containing these characters (called bidirectional IRIs or Bidi IRIs)
941require additional attention because of the non-trivial relation
942between logical representation (used for digital representation and
943for reading/spelling) and visual representation (used for
944display/printing).</t>
945
946<t>Because of the complex interaction between the logical representation,
947the visual representation, and the syntax of a Bidi IRI, a balance is
948needed between various requirements.
949The main requirements are<list style="hanging">
950<t hangText="1.">user-predictable conversion between visual and
951    logical representation;</t>
952<t hangText="2.">the ability to include a wide range of characters
953    in various parts of the IRI; and</t>
954<t hangText="3.">minor or no changes or restrictions for
955      implementations.</t>
956</list></t>
957
958<section title="Logical Storage and Visual Presentation" anchor="visual">
959
960<t>When stored or transmitted in digital representation, bidirectional
961IRIs MUST be in full logical order and MUST conform to the IRI syntax
962rules (which includes the rules relevant to their scheme). This
963ensures that bidirectional IRIs can be processed in the same way as
964other IRIs.</t> <t>Bidirectional IRIs MUST be rendered by using the
965Unicode Bidirectional Algorithm <xref target="UNIV6"/>, <xref
966target="UNI9"/>.  Bidirectional IRIs MUST be rendered in the same way
967as they would be if they were in a left-to-right embedding; i.e., as
968if they were preceded by U+202A, LEFT-TO-RIGHT EMBEDDING (LRE), and
969followed by U+202C, POP DIRECTIONAL FORMATTING (PDF).  Setting the
970embedding direction can also be done in a higher-level protocol (e.g.,
971the dir='ltr' attribute in HTML).</t> 
972
973<t>There is no requirement to use the above embedding if the display
974is still the same without the embedding. For example, a bidirectional
975IRI in a text with left-to-right base directionality (such as used for
976English or Cyrillic) that is preceded and followed by whitespace and
977strong left-to-right characters does not need an embedding.  Also, a
978bidirectional relative IRI reference that only contains strong
979right-to-left characters and weak characters and that starts and ends
980with a strong right-to-left character and appears in a text with
981right-to-left base directionality (such as used for Arabic or Hebrew)
982and is preceded and followed by whitespace and strong characters does
983not need an embedding.</t>
984
985<t>In some other cases, using U+200E, LEFT-TO-RIGHT MARK (LRM), may be
986sufficient to force the correct display behavior.  However, the
987details of the Unicode Bidirectional algorithm are not always easy to
988understand. Implementers are strongly advised to err on the side of
989caution and to use embedding in all cases where they are not
990completely sure that the display behavior is unaffected without the
991embedding.</t>
992
993<t>The Unicode Bidirectional Algorithm (<xref target="UNI9"/>, section
9944.3) permits higher-level protocols to influence bidirectional
995rendering. Such changes by higher-level protocols MUST NOT be used if
996they change the rendering of IRIs.</t> 
997
998<t>The bidirectional formatting characters that may be used before or
999after the IRI to ensure correct display are not themselves part of the
1000IRI.  IRIs MUST NOT contain bidirectional formatting characters (LRM,
1001RLM, LRE, RLE, LRO, RLO, and PDF). They affect the visual rendering of
1002the IRI but do not appear themselves. It would therefore not be
1003possible to input an IRI with such characters correctly.</t>
1004
1005</section> <!-- visual -->
1006<section title="Bidi IRI Structure" anchor="bidi-structure">
1007
1008<t>The Unicode Bidirectional Algorithm is designed mainly for running
1009text.  To make sure that it does not affect the rendering of
1010bidirectional IRIs too much, some restrictions on bidirectional IRIs
1011are necessary. These restrictions are given in terms of delimiters
1012(structural characters, mostly punctuation such as "@", ".", ":",
1013and<vspace/>"/") and components (usually consisting mostly of letters
1014and digits).</t>
1015
1016<t>The following syntax rules from <xref target="abnf"/> correspond to
1017components for the purpose of Bidi behavior: iuserinfo, ireg-name,
1018isegment, isegment-nz, isegment-nz-nc, ireg-name, iquery, and
1019ifragment.</t>
1020
1021<t>Specifications that define the syntax of any of the above
1022components MAY divide them further and define smaller parts to be
1023components according to this document. As an example, the restrictions
1024of <xref target="RFC3490"/> on bidirectional domain names correspond
1025to treating each label of a domain name as a component for schemes
1026with ireg-name as a domain name.  Even where the components are not
1027defined formally, it may be helpful to think about some syntax in
1028terms of components and to apply the relevant restrictions.  For
1029example, for the usual name/value syntax in query parts, it is
1030convenient to treat each name and each value as a component. As
1031another example, the extensions in a resource name can be treated as
1032separate components.</t>
1033
1034<t>For each component, the following restrictions apply:</t>
1035<t>
1036<list style="hanging">
1037
1038<t hangText="1.">A component SHOULD NOT use both right-to-left and
1039  left-to-right characters.</t>
1040
1041<t hangText="2.">A component using right-to-left characters SHOULD
1042  start and end with right-to-left characters.</t>
1043
1044</list></t>
1045
1046<t>The above restrictions are given as "SHOULD"s, rather than as
1047"MUST"s.  For IRIs that are never presented visually, they are not
1048relevant.  However, for IRIs in general, they are very important to
1049ensure consistent conversion between visual presentation and logical
1050representation, in both directions.</t>
1051
1052<t><list style="hanging">
1053
1054<t hangText="Note:">In some components, the above restrictions may
1055  actually be strictly enforced.  For example, <xref
1056  target="RFC3490"></xref> requires that these restrictions apply to
1057  the labels of a host name for those schemes where ireg-name is a
1058  host name.  In some other components (for example, path components)
1059  following these restrictions may not be too difficult.  For other
1060  components, such as parts of the query part, it may be very
1061  difficult to enforce the restrictions because the values of query
1062  parameters may be arbitrary character sequences.</t>
1063
1064</list></t>
1065
1066<t>If the above restrictions cannot be satisfied otherwise, the
1067affected component can always be mapped to URI notation as described
1068in <xref target="compmapping"/>. Please note that the whole component
1069has to be mapped (see also Example 9 below).</t>
1070
1071</section> <!-- bidi-structure -->
1072
1073<section title="Input of Bidi IRIs" anchor="bidiInput">
1074
1075<t>Bidi input methods MUST generate Bidi IRIs in logical order while
1076rendering them according to <xref target="visual"/>.  During input,
1077rendering SHOULD be updated after every new character is input to
1078avoid end-user confusion.</t>
1079
1080</section> <!-- bidiInput -->
1081
1082<section title="Examples">
1083
1084<t>This section gives examples of bidirectional IRIs, in Bidi
1085Notation.  It shows legal IRIs with the relationship between logical
1086and visual representation and explains how certain phenomena in this
1087relationship may look strange to somebody not familiar with
1088bidirectional behavior, but familiar to users of Arabic and Hebrew. It
1089also shows what happens if the restrictions given in <xref
1090target="bidi-structure"/> are not followed. The examples below can be
1091seen at <xref target="BidiEx"/>, in Arabic, Hebrew, and Bidi Notation
1092variants.</t>
1093
1094<t>To read the bidi text in the examples, read the visual
1095representation from left to right until you encounter a block of rtl
1096text. Read the rtl block (including slashes and other special
1097characters) from right to left, then continue at the next unread ltr
1098character.</t>
1099
1100<t>Example 1: A single component with rtl characters is inverted:
1101<vspace/>Logical representation:
1102"http://ab.CDEFGH.ij/kl/mn/op.html"<vspace/>Visual representation:
1103"http://ab.HGFEDC.ij/kl/mn/op.html"<vspace/> Components can be read
1104one by one, and each component can be read in its natural
1105direction.</t>
1106
1107<t>Example 2: More than one consecutive component with rtl characters
1108is inverted as a whole: <vspace/>Logical representation:
1109"http://ab.CDE.FGH/ij/kl/mn/op.html"<vspace/>Visual representation:
1110"http://ab.HGF.EDC/ij/kl/mn/op.html"<vspace/> A sequence of rtl
1111components is read rtl, in the same way as a sequence of rtl words is
1112read rtl in a bidi text.</t>
1113
1114<t>Example 3: All components of an IRI (except for the scheme) are
1115rtl.  All rtl components are inverted overall: <vspace/>Logical
1116representation:
1117"http://AB.CD.EF/GH/IJ/KL?MN=OP;QR=ST#UV"<vspace/>Visual
1118representation: "http://VU#TS=RQ;PO=NM?LK/JI/HG/FE.DC.BA"<vspace/> The
1119whole IRI (except the scheme) is read rtl. Delimiters between rtl
1120components stay between the respective components; delimiters between
1121ltr and rtl components don't move.</t>
1122
1123<t>Example 4: Each of several sequences of rtl components is inverted
1124on its own: <vspace/>Logical representation:
1125"http://AB.CD.ef/gh/IJ/KL.html"<vspace/>Visual representation:
1126"http://DC.BA.ef/gh/LK/JI.html"<vspace/> Each sequence of rtl
1127components is read rtl, in the same way as each sequence of rtl words
1128in an ltr text is read rtl.</t>
1129
1130<t>Example 5: Example 2, applied to components of different kinds:
1131<vspace/>Logical representation: "http://ab.cd.EF/GH/ij/kl.html"
1132<vspace/>Visual representation:
1133"http://ab.cd.HG/FE/ij/kl.html"<vspace/> The inversion of the domain
1134name label and the path component may be unexpected, but it is
1135consistent with other bidi behavior.  For reassurance that the domain
1136component really is "ab.cd.EF", it may be helpful to read aloud the
1137visual representation following the bidi algorithm. After
1138"http://ab.cd." one reads the RTL block "E-F-slash-G-H", which
1139corresponds to the logical representation.
1140</t>
1141
1142<t>Example 6: Same as Example 5, with more rtl components:
1143<vspace/>Logical representation:
1144"http://ab.CD.EF/GH/IJ/kl.html"<vspace/>Visual representation:
1145"http://ab.JI/HG/FE.DC/kl.html"<vspace/> The inversion of the domain
1146name labels and the path components may be easier to identify because
1147the delimiters also move.</t>
1148
1149<t>Example 7: A single rtl component includes digits: <vspace/>Logical
1150representation: "http://ab.CDE123FGH.ij/kl/mn/op.html"<vspace/>Visual
1151representation: "http://ab.HGF123EDC.ij/kl/mn/op.html"<vspace/>
1152Numbers are written ltr in all cases but are treated as an additional
1153embedding inside a run of rtl characters. This is completely
1154consistent with usual bidirectional text.</t>
1155
1156<t>Example 8 (not allowed): Numbers are at the start or end of an rtl
1157component:<vspace/>Logical representation:
1158"http://ab.cd.ef/GH1/2IJ/KL.html"<vspace/>Visual representation:
1159"http://ab.cd.ef/LK/JI1/2HG.html"<vspace/> The sequence "1/2" is
1160interpreted by the bidi algorithm as a fraction, fragmenting the
1161components and leading to confusion. There are other characters that
1162are interpreted in a special way close to numbers; in particular, "+",
1163"-", "#", "$", "%", ",", ".", and ":".</t>
1164
1165<t>Example 9 (not allowed): The numbers in the previous example are
1166percent-encoded: <vspace/>Logical representation:
1167"http://ab.cd.ef/GH%31/%32IJ/KL.html",<vspace/>Visual representation:
1168"http://ab.cd.ef/LK/JI%32/%31HG.html"</t>
1169
1170<t>Example 10 (allowed but not recommended): <vspace/>Logical
1171representation: "http://ab.CDEFGH.123/kl/mn/op.html"<vspace/>Visual
1172representation: "http://ab.123.HGFEDC/kl/mn/op.html"<vspace/>
1173Components consisting of only numbers are allowed (it would be rather
1174difficult to prohibit them), but these may interact with adjacent RTL
1175components in ways that are not easy to predict.</t>
1176
1177<t>Example 11 (allowed but not recommended): <vspace/>Logical
1178representation: "http://ab.CDEFGH.123ij/kl/mn/op.html"<vspace/>Visual
1179representation: "http://ab.123.HGFEDCij/kl/mn/op.html"<vspace/>
1180Components consisting of numbers and left-to-right characters are
1181allowed, but these may interact with adjacent RTL components in ways
1182that are not easy to predict.</t>
1183</section><!-- examples -->
1184</section><!-- bidi -->
1185
1186<section title="Normalization and Comparison" anchor="equivalence">
1187
1188<t><list style="hanging"><t hangText="Note:">The structure and much of
1189  the material for this section is taken from section 6 of <xref
1190  target="RFC3986"></xref>; the differences are due to the specifics
1191  of IRIs.</t></list></t>
1192
1193<t>One of the most common operations on IRIs is simple comparison:
1194Determining whether two IRIs are equivalent, without using the IRIs to
1195access their respective resource(s). A comparison is performed
1196whenever a response cache is accessed, a browser checks its history to
1197color a link, or an XML parser processes tags within a
1198namespace. Extensive normalization prior to comparison of IRIs may be
1199used by spiders and indexing engines to prune a search space or reduce
1200duplication of request actions and response storage.</t>
1201
1202<t>IRI comparison is performed for some particular purpose. Protocols
1203or implementations that compare IRIs for different purposes will often
1204be subject to differing design trade-offs in regards to how much
1205effort should be spent in reducing aliased identifiers. This section
1206describes various methods that may be used to compare IRIs, the
1207trade-offs between them, and the types of applications that might use
1208them.</t>
1209
1210<section title="Equivalence">
1211
1212<t>Because IRIs exist to identify resources, presumably they should be
1213considered equivalent when they identify the same resource. However,
1214this definition of equivalence is not of much practical use, as there
1215is no way for an implementation to compare two resources to determine
1216if they are "the same" unless it has full knowledge or control of
1217them. For this reason, determination of equivalence or difference of
1218IRIs is based on string comparison, perhaps augmented by reference to
1219additional rules provided by URI scheme definitions.  We use the terms
1220"different" and "equivalent" to describe the possible outcomes of such
1221comparisons, but there are many application-dependent versions of
1222equivalence.</t>
1223
1224<t>Even when it is possible to determine that two IRIs are equivalent,
1225IRI comparison is not sufficient to determine whether two IRIs
1226identify different resources. For example, an owner of two different
1227domain names could decide to serve the same resource from both,
1228resulting in two different IRIs. Therefore, comparison methods are
1229designed to minimize false negatives while strictly avoiding false
1230positives.</t>
1231
1232<t>In testing for equivalence, applications should not directly
1233compare relative references; the references should be converted to
1234their respective target IRIs before comparison. When IRIs are compared
1235to select (or avoid) a network action, such as retrieval of a
1236representation, fragment components (if any) should be excluded from
1237the comparison.</t>
1238
1239<t>Applications using IRIs as identity tokens with no relationship to
1240a protocol MUST use the Simple String Comparison (see <xref
1241target="stringcomp"></xref>).  All other applications MUST select one
1242of the comparison practices from the Comparison Ladder (see <xref
1243target="ladder"></xref>.</t>
1244</section> <!-- equivalence -->
1245
1246
1247<section title="Preparation for Comparison">
1248<t>Any kind of IRI comparison REQUIRES that any additional contextual
1249processing is first performed, including undoing higher-level
1250escapings or encodings in the protocol or format that carries an
1251IRI. This preprocessing is usually done when the protocol or format is
1252parsed.</t>
1253
1254<t>Examples of contextual preprocessing steps are described in <xref
1255target="LEIRIHREF"/>. </t>
1256
1257<t>Examples of such escapings or encodings are entities and
1258numeric character references in <xref target="HTML4"></xref> and <xref
1259target="XML1"></xref>. As an example,
1260"http://example.org/ros&amp;eacute;" (in HTML),
1261"http://example.org/ros&amp;#233;" (in HTML or XML), and
1262<vspace/>"http://example.org/ros&amp;#xE9;" (in HTML or XML) are all
1263resolved into what is denoted in this document (see <xref
1264target="sec-Notation"></xref>) as "http://example.org/ros&amp;#xE9;"
1265(the "&amp;#xE9;" here standing for the actual e-acute character, to
1266compensate for the fact that this document cannot contain non-ASCII
1267characters).</t>
1268
1269<t>Similar considerations apply to encodings such as Transfer Codings
1270in HTTP (see <xref target="RFC2616"></xref>) and Content Transfer
1271Encodings in MIME (<xref target="RFC2045"></xref>), although in these
1272cases, the encoding is based not on characters but on octets, and
1273additional care is required to make sure that characters, and not just
1274arbitrary octets, are compared (see <xref
1275target="stringcomp"></xref>).</t>
1276
1277</section> <!-- preparation -->
1278
1279<section title="Comparison Ladder" anchor="ladder">
1280
1281<t>In practice, a variety of methods are used to test IRI
1282equivalence. These methods fall into a range distinguished by the
1283amount of processing required and the degree to which the probability
1284of false negatives is reduced. As noted above, false negatives cannot
1285be eliminated. In practice, their probability can be reduced, but this
1286reduction requires more processing and is not cost-effective for all
1287applications.</t>
1288
1289
1290<t>If this range of comparison practices is considered as a ladder,
1291the following discussion will climb the ladder, starting with
1292practices that are cheap but have a relatively higher chance of
1293producing false negatives, and proceeding to those that have higher
1294computational cost and lower risk of false negatives.</t>
1295
1296<section title="Simple String Comparison" anchor="stringcomp">
1297
1298<t>If two IRIs, when considered as character strings, are identical,
1299then it is safe to conclude that they are equivalent.  This type of
1300equivalence test has very low computational cost and is in wide use in
1301a variety of applications, particularly in the domain of parsing. It
1302is also used when a definitive answer to the question of IRI
1303equivalence is needed that is independent of the scheme used and that
1304can be calculated quickly and without accessing a network. An example
1305of such a case is XML Namespaces (<xref
1306target="XMLNamespace"></xref>).</t>
1307
1308
1309<t>Testing strings for equivalence requires some basic precautions.
1310This procedure is often referred to as "bit-for-bit" or
1311"byte-for-byte" comparison, which is potentially misleading. Testing
1312strings for equality is normally based on pair comparison of the
1313characters that make up the strings, starting from the first and
1314proceeding until both strings are exhausted and all characters are
1315found to be equal, until a pair of characters compares unequal, or
1316until one of the strings is exhausted before the other.</t>
1317
1318<t>This character comparison requires that each pair of characters be
1319put in comparable encoding form. For example, should one IRI be stored
1320in a byte array in UTF-8 encoding form and the second in a UTF-16
1321encoding form, bit-for-bit comparisons applied naively will produce
1322errors. It is better to speak of equality on a character-for-character
1323rather than on a byte-for-byte or bit-for-bit basis.  In practical
1324terms, character-by-character comparisons should be done codepoint by
1325codepoint after conversion to a common character encoding form.
1326
1327When comparing character by character, the comparison function MUST
1328NOT map IRIs to URIs, because such a mapping would create additional
1329spurious equivalences. It follows that an IRI SHOULD NOT be modified
1330when being transported if there is any chance that this IRI might be
1331used in a context that uses Simple String Comparison.</t>
1332
1333
1334<t>False negatives are caused by the production and use of IRI
1335aliases. Unnecessary aliases can be reduced, regardless of the
1336comparison method, by consistently providing IRI references in an
1337already normalized form (i.e., a form identical to what would be
1338produced after normalization is applied, as described below).
1339Protocols and data formats often limit some IRI comparisons to simple
1340string comparison, based on the theory that people and implementations
1341will, in their own best interest, be consistent in providing IRI
1342references, or at least be consistent enough to negate any efficiency
1343that might be obtained from further normalization.</t>
1344</section> <!-- stringcomp -->
1345
1346<section title="Syntax-Based Normalization">
1347
1348<figure><preamble>Implementations may use logic based on the
1349definitions provided by this specification to reduce the probability
1350of false negatives. This processing is moderately higher in cost than
1351character-for-character string comparison. For example, an application
1352using this approach could reasonably consider the following two IRIs
1353equivalent:</preamble>
1354
1355<artwork>
1356   example://a/b/c/%7Bfoo%7D/ros&amp;#xE9;
1357   eXAMPLE://a/./b/../b/%63/%7bfoo%7d/ros%C3%A9
1358</artwork></figure>
1359
1360<t>Web user agents, such as browsers, typically apply this type of IRI
1361normalization when determining whether a cached response is
1362available. Syntax-based normalization includes such techniques as case
1363normalization, character normalization, percent-encoding
1364normalization, and removal of dot-segments.</t>
1365
1366<section title="Case Normalization">
1367
1368<t>For all IRIs, the hexadecimal digits within a percent-encoding
1369triplet (e.g., "%3a" versus "%3A") are case-insensitive and therefore
1370should be normalized to use uppercase letters for the digits A-F.</t>
1371
1372<t>When an IRI uses components of the generic syntax, the component
1373syntax equivalence rules always apply; namely, that the scheme and
1374US-ASCII only host are case insensitive and therefore should be
1375normalized to lowercase. For example, the URI
1376"HTTP://www.EXAMPLE.com/" is equivalent to
1377"http://www.example.com/". Case equivalence for non-ASCII characters
1378in IRI components that are IDNs are discussed in <xref
1379target="schemecomp"></xref>.  The other generic syntax components are
1380assumed to be case sensitive unless specifically defined otherwise by
1381the scheme.</t>
1382
1383<t>Creating schemes that allow case-insensitive syntax components
1384containing non-ASCII characters should be avoided. Case normalization
1385of non-ASCII characters can be culturally dependent and is always a
1386complex operation. The only exception concerns non-ASCII host names
1387for which the character normalization includes a mapping step derived
1388from case folding.</t>
1389
1390</section> <!-- casenorm -->
1391
1392<section title="Character Normalization" anchor="normalization">
1393
1394<t>The Unicode Standard <xref target="UNIV6"></xref> defines various
1395equivalences between sequences of characters for various
1396purposes. Unicode Standard Annex #15 <xref target="UTR15"></xref>
1397defines various Normalization Forms for these equivalences, in
1398particular Normalization Form C (NFC, Canonical Decomposition,
1399followed by Canonical Composition) and Normalization Form KC (NFKC,
1400Compatibility Decomposition, followed by Canonical Composition).</t>
1401
1402<t> IRIs already in Unicode MUST NOT be normalized before parsing or
1403interpreting. In many non-Unicode character encodings, some text
1404cannot be represented directly. For example, the word "Vietnam" is
1405natively written "Vi&amp;#x1EC7;t Nam" (containing a LATIN SMALL
1406LETTER E WITH CIRCUMFLEX AND DOT BELOW) in NFC, but a direct
1407transcoding from the windows-1258 character encoding leads to
1408"Vi&amp;#xEA;&amp;#x323;t Nam" (containing a LATIN SMALL LETTER E WITH
1409CIRCUMFLEX followed by a COMBINING DOT BELOW). Direct transcoding of
1410other 8-bit encodings of Vietnamese may lead to other
1411representations.</t>
1412
1413<t>Equivalence of IRIs MUST rely on the assumption that IRIs are
1414appropriately pre-character-normalized rather than apply character
1415normalization when comparing two IRIs. The exceptions are conversion
1416from a non-digital form, and conversion from a non-UCS-based character
1417encoding to a UCS-based character encoding. In these cases, NFC or a
1418normalizing transcoder using NFC MUST be used for interoperability. To
1419avoid false negatives and problems with transcoding, IRIs SHOULD be
1420created by using NFC. Using NFKC may avoid even more problems; for
1421example, by choosing half-width Latin letters instead of full-width
1422ones, and full-width instead of half-width Katakana.</t>
1423
1424
1425<t>As an example,
1426"http://www.example.org/r&amp;#xE9;sum&amp;#xE9;.html" (in XML
1427Notation) is in NFC. On the other hand,
1428"http://www.example.org/re&amp;#x301;sume&amp;#x301;.html" is not in
1429NFC.</t>
1430
1431<t>The former uses precombined e-acute characters, and the latter uses
1432"e" characters followed by combining acute accents. Both usages are
1433defined as canonically equivalent in <xref target="UNIV6"></xref>.</t>
1434
1435<t><list style="hanging">
1436
1437<t hangText="Note:">
1438Because it is unknown how a particular sequence of characters is being
1439treated with respect to character normalization, it would be
1440inappropriate to allow third parties to normalize an IRI
1441arbitrarily. This does not contradict the recommendation that when a
1442resource is created, its IRI should be as character normalized as
1443possible (i.e., NFC or even NFKC). This is similar to the
1444uppercase/lowercase problems.  Some parts of a URI are case
1445insensitive (for example, the domain name). For others, it is unclear
1446whether they are case sensitive, case insensitive, or something in
1447between (e.g., case sensitive, but with a multiple choice selection if
1448the wrong case is used, instead of a direct negative result).  The
1449best recipe is that the creator use a reasonable capitalization and,
1450when transferring the URI, capitalization never be
1451changed.</t></list></t>
1452
1453<t>Various IRI schemes may allow the usage of Internationalized Domain
1454Names (IDN) <xref target="RFC5890"/> either in the ireg-name
1455part or elsewhere. Character Normalization also applies to IDNs, as
1456discussed in <xref target="schemecomp"/>.</t>
1457</section> <!-- charnorm -->
1458
1459<section title="Percent-Encoding Normalization">
1460
1461<t>The percent-encoding mechanism (Section 2.1 of <xref
1462target="RFC3986"></xref>) is a frequent source of variance among
1463otherwise identical IRIs. In addition to the case normalization issue
1464noted above, some IRI producers percent-encode octets that do not
1465require percent-encoding, resulting in IRIs that are equivalent to
1466their nonencoded counterparts. These IRIs should be normalized by
1467decoding any percent-encoded octet sequence that corresponds to an
1468unreserved character, as described in section 2.3 of <xref
1469target="RFC3986"></xref>.</t>
1470
1471<t>For actual resolution, differences in percent-encoding (except for
1472the percent-encoding of reserved characters) MUST always result in the
1473same resource.  For example, "http://example.org/~user",
1474"http://example.org/%7euser", and "http://example.org/%7Euser", must
1475resolve to the same resource.</t>
1476
1477<t>If this kind of equivalence is to be tested, the percent-encoding
1478of both IRIs to be compared has to be aligned; for example, by
1479converting both IRIs to URIs (see Section 3.1), eliminating escape
1480differences in the resulting URIs, and making sure that the case of
1481the hexadecimal characters in the percent-encoding is always the same
1482(preferably upper case). If the IRI is to be passed to another
1483application or used further in some other way, its original form MUST
1484be preserved.  The conversion described here should be performed only
1485for local comparison.</t>
1486
1487</section> <!-- pctnorm -->
1488
1489<section title="Path Segment Normalization">
1490
1491<t>The complete path segments "." and ".." are intended only for use
1492within relative references (Section 4.1 of <xref
1493target="RFC3986"></xref>) and are removed as part of the reference
1494resolution process (Section 5.2 of <xref target="RFC3986"></xref>).
1495However, some implementations may incorrectly assume that reference
1496resolution is not necessary when the reference is already an IRI, and
1497thus fail to remove dot-segments when they occur in non-relative
1498paths.  IRI normalizers should remove dot-segments by applying the
1499remove_dot_segments algorithm to the path, as described in Section
15005.2.4 of <xref target="RFC3986"></xref>.</t>
1501
1502</section> <!-- pathnorm -->
1503</section> <!-- ladder -->
1504
1505<section title="Scheme-Based Normalization" anchor="schemecomp">
1506
1507<t>The syntax and semantics of IRIs vary from scheme to scheme, as
1508described by the defining specification for each
1509scheme. Implementations may use scheme-specific rules, at further
1510processing cost, to reduce the probability of false negatives. For
1511example, because the "http" scheme makes use of an authority
1512component, has a default port of "80", and defines an empty path to be
1513equivalent to "/", the following four IRIs are equivalent:</t>
1514
1515<figure><artwork>
1516   http://example.com
1517   http://example.com/
1518   http://example.com:/
1519   http://example.com:80/</artwork></figure>
1520
1521<t>In general, an IRI that uses the generic syntax for authority with
1522an empty path should be normalized to a path of "/". Likewise, an
1523explicit ":port", for which the port is empty or the default for the
1524scheme, is equivalent to one where the port and its ":" delimiter are
1525elided and thus should be removed by scheme-based normalization. For
1526example, the second IRI above is the normal form for the "http"
1527scheme.</t>
1528
1529<t>Another case where normalization varies by scheme is in the
1530handling of an empty authority component or empty host
1531subcomponent. For many scheme specifications, an empty authority or
1532host is considered an error; for others, it is considered equivalent
1533to "localhost" or the end-user's host. When a scheme defines a default
1534for authority and an IRI reference to that default is desired, the
1535reference should be normalized to an empty authority for the sake of
1536uniformity, brevity, and internationalization. If, however, either the
1537userinfo or port subcomponents are non-empty, then the host should be
1538given explicitly even if it matches the default.</t>
1539
1540<t>Normalization should not remove delimiters when their associated
1541component is empty unless it is licensed to do so by the scheme
1542specification. For example, the IRI "http://example.com/?" cannot be
1543assumed to be equivalent to any of the examples above. Likewise, the
1544presence or absence of delimiters within a userinfo subcomponent is
1545usually significant to its interpretation.  The fragment component is
1546not subject to any scheme-based normalization; thus, two IRIs that
1547differ only by the suffix "#" are considered different regardless of
1548the scheme.</t>
1549 
1550<t>Some IRI schemes allow the usage of Internationalized Domain
1551Names (IDN) <xref target='RFC5890'></xref> either in their ireg-name
1552part or elswhere. When in use in IRIs, those names SHOULD
1553conform to the definition of U-Label in <xref
1554target='RFC5890'></xref>. An IRI containing an invalid IDN cannot
1555successfully be resolved. For legibility purposes, they
1556SHOULD NOT be converted into ASCII Compatible Encoding (ACE).</t>
1557
1558<t>Scheme-based normalization may also consider IDN
1559components and their conversions to punycode as equivalent. As an
1560example, "http://r&amp;#xE9;sum&amp;#xE9;.example.org" may be
1561considered equivalent to
1562"http://xn--rsum-bpad.example.org".</t><t>Other scheme-specific
1563normalizations are possible.</t>
1564
1565</section> <!-- schemenorm -->
1566
1567<section title="Protocol-Based Normalization">
1568
1569<t>Substantial effort to reduce the incidence of false negatives is
1570often cost-effective for web spiders. Consequently, they implement
1571even more aggressive techniques in IRI comparison. For example, if
1572they observe that an IRI such as</t>
1573
1574<figure><artwork>
1575   http://example.com/data</artwork></figure>
1576<t>redirects to an IRI differing only in the trailing slash</t>
1577<figure><artwork>
1578   http://example.com/data/</artwork></figure>
1579
1580<t>they will likely regard the two as equivalent in the future.  This
1581kind of technique is only appropriate when equivalence is clearly
1582indicated by both the result of accessing the resources and the common
1583conventions of their scheme's dereference algorithm (in this case, use
1584of redirection by HTTP origin servers to avoid problems with relative
1585references).</t>
1586
1587</section> <!-- protonorm -->
1588</section> <!-- equivalence -->
1589</section> 
1590
1591<section title="Use of IRIs" anchor="IRIuse">
1592
1593<section title="Limitations on UCS Characters Allowed in IRIs" anchor="limitations">
1594
1595<t>This section discusses limitations on characters and character
1596sequences usable for IRIs beyond those given in <xref target="abnf"/>
1597and <xref target="visual"/>. The considerations in this section are
1598relevant when IRIs are created and when URIs are converted to
1599IRIs.</t>
1600
1601<t>
1602
1603<list style="hanging"><t hangText="a.">The repertoire of characters allowed
1604    in each IRI component is limited by the definition of that component.
1605    For example, the definition of the scheme component does not allow
1606    characters beyond US-ASCII.
1607    <vspace blankLines="1"/>
1608    (Note: In accordance with URI practice, generic IRI
1609    software cannot and should not check for such limitations.)</t>
1610
1611<t hangText="b.">The UCS contains many areas of characters for which
1612    there are strong visual look-alikes. Because of the likelihood of
1613    transcription errors, these also should be avoided. This includes
1614    the full-width equivalents of Latin characters, half-width
1615    Katakana characters for Japanese, and many others. It also
1616    includes many look-alikes of "space", "delims", and "unwise",
1617    characters excluded in <xref target="RFC3491"/>.</t>
1618   
1619</list>
1620</t>
1621
1622<t>Additional information is available from <xref target="UNIXML"/>.
1623    <xref target="UNIXML"/> is written in the context of running text
1624    rather than in that of identifiers. Nevertheless, it discusses
1625    many of the categories of characters not appropriate for IRIs.</t>
1626</section> <!-- limitations -->
1627
1628<section title="Software Interfaces and Protocols">
1629
1630<t>Although an IRI is defined as a sequence of characters, software
1631interfaces for URIs typically function on sequences of octets or other
1632kinds of code units. Thus, software interfaces and protocols MUST
1633define which character encoding is used.</t>
1634
1635<t>Intermediate software interfaces between IRI-capable components and
1636URI-only components MUST map the IRIs per <xref target="mapping"/>,
1637when transferring from IRI-capable to URI-only components.
1638
1639This mapping SHOULD be applied as late as possible. It SHOULD NOT be
1640applied between components that are known to be able to handle IRIs.</t>
1641</section> <!-- software -->
1642
1643<section title="Format of URIs and IRIs in Documents and Protocols">
1644
1645<t>Document formats that transport URIs may have to be upgraded to allow
1646the transport of IRIs. In cases where the document as a whole
1647has a native character encoding, IRIs MUST also be encoded in this
1648character encoding and converted accordingly by a parser or interpreter.
1649
1650IRI characters not expressible in the native character encoding SHOULD
1651be escaped by using the escaping conventions of the document format if
1652such conventions are available. Alternatively, they MAY be
1653percent-encoded according to <xref target="mapping"/>. For example, in
1654HTML or XML, numeric character references SHOULD be used. If a
1655document as a whole has a native character encoding and that character
1656encoding is not UTF-8, then IRIs MUST NOT be placed into the document
1657in the UTF-8 character encoding.</t>
1658
1659<t>((UPDATE THIS NOTE)) Note: Some formats already accommodate IRIs,
1660although they use different terminology. HTML 4.0 <xref
1661target="HTML4"/> defines the conversion from IRIs to URIs as
1662error-avoiding behavior. XML 1.0 <xref target="XML1"/>, XLink <xref
1663target="XLink"/>, XML Schema <xref target="XMLSchema"/>, and
1664specifications based upon them allow IRIs. Also, it is expected that
1665all relevant new W3C formats and protocols will be required to handle
1666IRIs <xref target="CharMod"/>.</t>
1667
1668</section> <!-- format -->
1669
1670<section title="Use of UTF-8 for Encoding Original Characters" anchor="UTF8use">
1671
1672<t>This section discusses details and gives examples for point c) in
1673<xref target="Applicability"/>. To be able to use IRIs, the URI
1674corresponding to the IRI in question has to encode original characters
1675into octets by using UTF-8.  This can be specified for all URIs of a
1676URI scheme or can apply to individual URIs for schemes that do not
1677specify how to encode original characters.  It can apply to the whole
1678URI, or only to some part. For background information on encoding
1679characters into URIs, see also Section 2.5 of <xref
1680target="RFC3986"/>.</t>
1681
1682<t>For new URI schemes, using UTF-8 is recommended in <xref
1683target="RFC4395bis"/>.  Examples where UTF-8 is already used are the URN
1684syntax <xref target="RFC2141"/>, IMAP URLs <xref target="RFC2192"/>,
1685and POP URLs <xref target="RFC2384"/>.  On the other hand, because the
1686HTTP URI scheme does not specify how to encode original characters,
1687only some HTTP URLs can have corresponding but different IRIs.</t>
1688
1689<t>For example, for a document with a URI
1690of<vspace/>"http://www.example.org/r%C3%A9sum%C3%A9.html", it is
1691possible to construct a corresponding IRI (in XML notation, see <xref
1692target="sec-Notation"/>):
1693"http://www.example.org/r&amp;#xE9;sum&amp;#xE9;.html" ("&amp;#xE9;"
1694stands for the e-acute character, and "%C3%A9" is the UTF-8 encoded
1695and percent-encoded representation of that character). On the other
1696hand, for a document with a URI of
1697"http://www.example.org/r%E9sum%E9.html", the percent-encoding octets
1698cannot be converted to actual characters in an IRI, as the
1699percent-encoding is not based on UTF-8.</t>
1700
1701<t>For most URI schemes, there is no need to upgrade their scheme
1702definition in order for them to work with IRIs.  The main case where
1703upgrading makes sense is when a scheme definition, or a particular
1704component of a scheme, is strictly limited to the use of US-ASCII
1705characters with no provision to include non-ASCII characters/octets
1706via percent-encoding, or if a scheme definition currently uses highly
1707scheme-specific provisions for the encoding of non-ASCII characters.
1708An example of this is the mailto: scheme <xref target="RFC2368"/>.</t>
1709
1710<t>This specification updates the IANA registry of URI schemes to note
1711their applicability to IRIs, see <xref target="iana"/>.  All IRIs use
1712URI schemes, and all URIs with URI schemes can be used as IRIs, even
1713though in some cases only by using URIs directly as IRIs, without any
1714conversion.</t>
1715
1716<t>Scheme definitions can impose restrictions on the syntax of
1717scheme-specific URIs; i.e., URIs that are admissible under the generic
1718URI syntax <xref target="RFC3986"/> may not be admissible due to
1719narrower syntactic constraints imposed by a URI scheme
1720specification. URI scheme definitions cannot broaden the syntactic
1721restrictions of the generic URI syntax; otherwise, it would be
1722possible to generate URIs that satisfied the scheme-specific syntactic
1723constraints without satisfying the syntactic constraints of the
1724generic URI syntax. However, additional syntactic constraints imposed
1725by URI scheme specifications are applicable to IRI, as the
1726corresponding URI resulting from the mapping defined in <xref
1727target="mapping"/> MUST be a valid URI under the syntactic
1728restrictions of generic URI syntax and any narrower restrictions
1729imposed by the corresponding URI scheme specification.</t>
1730
1731<t>The requirement for the use of UTF-8 generally applies to all parts
1732of a URI.  However, it is possible that the capability of IRIs to
1733represent a wide range of characters directly is used just in some
1734parts of the IRI (or IRI reference). The other parts of the IRI may
1735only contain US-ASCII characters, or they may not be based on
1736UTF-8. They may be based on another character encoding, or they may
1737directly encode raw binary data (see also <xref
1738target="RFC2397"/>). </t>
1739
1740<t>For example, it is possible to have a URI reference
1741of<vspace/>"http://www.example.org/r%E9sum%E9.xml#r%C3%A9sum%C3%A9",
1742where the document name is encoded in iso-8859-1 based on server
1743settings, but where the fragment identifier is encoded in UTF-8 according
1744to <xref target="XPointer"/>. The IRI corresponding to the above
1745URI would be (in XML notation)<vspace/>"http://www.example.org/r%E9sum%E9.xml#r&amp;#xE9;sum&amp;#xE9;".</t>
1746
1747<t>Similar considerations apply to query parts. The functionality
1748of IRIs (namely, to be able to include non-ASCII characters) can
1749only be used if the query part is encoded in UTF-8.</t>
1750
1751</section> <!-- utf8 -->
1752
1753<section title="Relative IRI References">
1754<t>Processing of relative IRI references against a base is handled
1755straightforwardly; the algorithms of <xref target="RFC3986"/> can
1756be applied directly, treating the characters additionally allowed
1757in IRI references in the same way that unreserved characters are in URI
1758references.</t>
1759
1760</section> <!-- relative -->
1761</section> <!-- IRIuse -->
1762
1763<section title="Liberal Handling of Otherwise Invalid IRIs" anchor="LEIRIHREF">
1764
1765<t>(EDITOR NOTE: This Section may move to an appendix.)
1766 
1767Some technical specifications and widely-deployed software have
1768allowed additional variations and extensions of IRIs to be used in
1769syntactic components. This section describes two widely-used
1770preprocessing agreements. Other technical specifications may wish to
1771reference a syntactic component which is "a valid IRI or a string that
1772will map to a valid IRI after this preprocessing algorithm". These two
1773variants are known as <xref target="LEIRI">Legacy Extended IRI or
1774LEIRI</xref>, and <xref target="HTML5">Web Address</xref>).
1775</t>
1776
1777<t>Future technical specifications SHOULD NOT allow conforming
1778producers to produce, or conforming content to contain, such forms,
1779as they are not interoperable with other IRI consuming software.</t>
1780
1781<section title="LEIRI Processing"  anchor="LEIRIspec">
1782  <t>This section defines Legacy Extended IRIs (LEIRIs).
1783    The syntax of Legacy Extended IRIs is the same as that for &lt;IRI-reference>,
1784    except that the ucschar production is replaced by the leiri-ucschar production:</t>
1785<figure>
1786
1787<artwork>
1788  leiri-ucschar  = " " / "&lt;" / "&gt;" / '"' / "{" / "}" / "|"
1789                   / "\" / "^" / "`" / %x0-1F / %x7F-D7FF
1790                   / %xE000-FFFD / %x10000-10FFFF
1791</artwork>
1792
1793<postamble>
1794  Among other extensions, processors based on this specification also
1795  did not enforce the restriction on bidirectional formatting
1796  characters in <xref target="visual"></xref>, and the iprivate
1797  production becomes redundant.</postamble>
1798</figure>
1799
1800<t>To convert a string allowed as a LEIRI to an IRI, each character
1801allowed in leiri-ucschar but not in ucschar must be percent-encoded
1802using <xref target="compmapping"/>.</t>
1803</section> <!-- leiriproc -->
1804
1805<section title="Web Address Processing" anchor="webaddress">
1806
1807<t>Many popular web browsers have taken the approach of being quite
1808liberal in what is accepted as a "URL" or its relative
1809forms. This section describes their behavior in terms of a preprocessor
1810which maps strings into the IRI space for subsequent parsing and
1811interpretation as an IRI.</t>
1812
1813<t>In some situations, it might be appropriate to describe the syntax
1814that a liberal consumer implementation might accept as a "Web
1815Address" or "Hypertext Reference" or "HREF". However,
1816technical specifications SHOULD restrict the syntactic form allowed by compliant producers
1817to the IRI or IRI reference syntax defined in this document
1818even if they want to mandate this processing.</t>
1819
1820<t>
1821Summary:
1822<list style="symbols">
1823   <t>Leading and trailing whitespace is removed.</t>
1824   <t>Some additional characters are removed.</t>
1825   <t>Some additional characters are allowed and escaped (as with LEIRI).</t>
1826   <t>If interpreting an IRI as a URI, the pct-encoding of the query
1827   component of the parsed URI component depends on operational
1828   context.</t>
1829</list>
1830</t>
1831
1832<t>Each string provided may have an associated charset (called
1833the HREF-charset here); this defaults to UTF-8.
1834For web browsers interpreting HTML, the document
1835charset of a string is determined:
1836
1837<list style="hanging">
1838<t hangText="If the string came from a script (e.g. as an argument to
1839 a method)">The HRef-charset is the script's charset.</t>
1840
1841<t hangText="If the string came from a DOM node (e.g. from an
1842  element)">The node has a Document, and the HRef-charset is the
1843  Document's character encoding.</t>
1844
1845<t hangText="If the string had a HRef-charset defined when the string was
1846created or defined">The HRef-charset is as defined.</t>
1847
1848</list></t>
1849
1850<t>If the resulting HRef-charset is a unicode based character encoding
1851(e.g., UTF-16), then use UTF-8 instead.</t>
1852
1853
1854<figure>
1855<preamble>The syntax for Web Addresses is obtained by replacing the 'ucschar',
1856  pct-form, path-sep, and ifragment rules with the href-ucschar, href-pct-form, href-path-sep,
1857  and href-ifragment
1858  rules below. In addition, some characters are stripped.</preamble>
1859
1860<artwork type='abnf'>
1861  href-ucschar   = " " / "&lt;" / "&gt;" / DQUOTE / "{" / "}" / "|"
1862                 / "\" / "^" / "`" / %x0-1F / %x7F-D7FF
1863                 / %xE000-FFFD / %x10000-10FFFF
1864  href-pct-form  = pct-encoded / "%"
1865  href-path-sep  = "/" / "\"
1866  href-ifragment = *( ipchar / "/" / "?" / "#" )  ; adding "#"
1867  href-strip     = &lt;to be done&gt;
1868</artwork>
1869
1870<postamble>
1871(NOTE: NEED TO FIX THESE SETS TO MATCH HTML5; NOT SURE ABOUT NEXT SENTENCE)
1872browsers did not enforce the restriction on bidirectional formatting
1873  characters in <xref target="visual"></xref>, and the iprivate
1874  production becomes redundant.</postamble>
1875</figure>
1876
1877<t>'Web Address processing' requires the following additional
1878preprocessing steps:
1879
1880<list style="numbers">
1881
1882<t>Leading and trailing instances of space (U+0020),
1883CR (U+000A), LF (U+000D), and TAB (U+0009) characters are removed.</t>
1884
1885<t>strip all characters in href-strip.</t>
1886  <t>Percent-encode all characters in href-ucschar not in ucschar.</t>
1887  <t>Replace occurrences of "%" not followed by two hexadecimal digits by "%25".</t>
1888  <t>Convert backslashes ('\') matching href-path-sep to forward slashes ('/').</t>
1889</list></t>
1890</section> <!-- webaddress -->
1891
1892<section title="Characters Not Allowed in IRIs" anchor="notAllowed">
1893
1894<t>This section provides a list of the groups of characters and code
1895points that are allowed by LEIRI or HREF but are not allowed in IRIs or are
1896allowed in IRIs only in the query part. For each group of characters,
1897advice on the usage of these characters is also given, concentrating
1898on the reasons for why they are excluded from IRI use.</t>
1899
1900<t>
1901
1902<list><t>Space (U+0020): Some formats and applications use space as a
1903delimiter, e.g. for items in a list. Appendix C of <xref
1904target="RFC3986"></xref> also mentions that white space may have to be
1905added when displaying or printing long URIs; the same applies to long
1906IRIs. This means that spaces can disappear, or can make the what is
1907intended as a single IRI or IRI reference to be treated as two or more
1908separate IRIs.</t>
1909
1910<t>Delimiters "&lt;" (U+003C), "&gt;" (U+003E), and '"' (U+0022):
1911Appendix C of <xref target="RFC3986"></xref> suggests the use of
1912double-quotes ("http://example.com/") and angle brackets
1913(&lt;http://example.com/&gt;) as delimiters for URIs in plain
1914text. These conventions are often used, and also apply to IRIs.  Using
1915these characters in strings intended to be IRIs would result in the
1916IRIs being cut off at the wrong place.</t>
1917
1918<t>Unwise characters "\" (U+005C), "^" (U+005E), "`"
1919(U+0060), "{" (U+007B), "|" (U+007C), and "}" (U+007D): These
1920characters originally have been excluded from URIs because the
1921respective codepoints are assigned to different graphic characters in
1922some 7-bit or 8-bit encoding. Despite the move to Unicode, some of
1923these characters are still occasionally displayed differently on some
1924systems, e.g. U+005C may appear as a Japanese Yen symbol on some
1925systems. Also, the fact that these characters are not used in URIs or
1926IRIs has encouraged their use outside URIs or IRIs in contexts that
1927may include URIs or IRIs. If a string with such a character were used
1928as an IRI in such a context, it would likely be interpreted
1929piecemeal.</t>
1930
1931<t>The controls (C0 controls, DEL, and C1 controls, #x0 - #x1F #x7F -
1932#x9F): There is generally no way to transmit these characters reliably
1933as text outside of a charset encoding.  Even when in encoded form,
1934many software components silently filter out some of these characters,
1935or may stop processing alltogether when encountering some of
1936them. These characters may affect text display in subtle, unnoticable
1937ways or in drastic, global, and irreversible ways depending on the
1938hardware and software involved. The use of some of these characters
1939would allow malicious users to manipulate the display of an IRI and
1940its context in many situations.</t>
1941
1942<t>Bidi formatting characters (U+200E, U+200F, U+202A-202E): These
1943characters affect the display ordering of characters. If IRIs were
1944allowed to contain these characters and the resulting visual display
1945transcribed. they could not be converted back to electronic form
1946(logical order) unambiguously. These characters, if allowed in IRIs,
1947might allow malicious users to manipulate the display of IRI and its
1948context.</t>
1949
1950<t>Specials (U+FFF0-FFFD): These code points provide functionality
1951beyond that useful in an IRI, for example byte order identification,
1952annotation, and replacements for unknown characters and objects. Their
1953use and interpretation in an IRI would serve no purpose and might lead
1954to confusing display variations.</t>
1955
1956<t>Private use code points (U+E000-F8FF, U+F0000-FFFFD,
1957U+100000-10FFFD): Display and interpretation of these code points is
1958by definition undefined without private agreement. Therefore, these
1959code points are not suited for use on the Internet. They are not
1960interoperable and may have unpredictable effects.</t>
1961
1962<t>Tags (U+E0000-E0FFF): These characters provide a way to language
1963tag in Unicode plain text. They are not appropriate for IRIs because
1964language information in identifiers cannot reliably be input,
1965transmitted (e.g. on a visual medium such as paper), or
1966recognized.</t>
1967
1968<t>Non-characters (U+FDD0-FDEF, U+1FFFE-1FFFF, U+2FFFE-2FFFF,
1969U+3FFFE-3FFFF, U+4FFFE-4FFFF, U+5FFFE-5FFFF, U+6FFFE-6FFFF,
1970U+7FFFE-7FFFF, U+8FFFE-8FFFF, U+9FFFE-9FFFF, U+AFFFE-AFFFF,
1971U+BFFFE-BFFFF, U+CFFFE-CFFFF, U+DFFFE-DFFFF, U+EFFFE-EFFFF,
1972U+FFFFE-FFFFF, U+10FFFE-10FFFF): These code points are defined as
1973non-characters. Applications may use some of them internally, but are
1974not prepared to interchange them.</t>
1975
1976</list></t>
1977
1978<t>LEIRI preprocessing disallowed some code points and
1979code units:
1980
1981<list><t>Surrogate code units (D800-DFFF): These do not represent
1982Unicode codepoints.</t></list></t>
1983</section> <!-- notallowed -->
1984</section> <!-- lieirihref -->
1985 
1986<section title="URI/IRI Processing Guidelines (Informative)" anchor="guidelines">
1987
1988<t>This informative section provides guidelines for supporting IRIs in
1989the same software components and operations that currently process
1990URIs: Software interfaces that handle URIs, software that allows users
1991to enter URIs, software that creates or generates URIs, software that
1992displays URIs, formats and protocols that transport URIs, and software
1993that interprets URIs. These may all require modification before
1994functioning properly with IRIs. The considerations in this section
1995also apply to URI references and IRI references.</t>
1996
1997<section title="URI/IRI Software Interfaces">
1998<t>Software interfaces that handle URIs, such as URI-handling APIs and
1999protocols transferring URIs, need interfaces and protocol elements
2000that are designed to carry IRIs.</t>
2001
2002<t>In case the current handling in an API or protocol is based on
2003US-ASCII, UTF-8 is recommended as the character encoding for IRIs, as
2004it is compatible with US-ASCII, is in accordance with the
2005recommendations of <xref target="RFC2277"/>, and makes converting to
2006URIs easy. In any case, the API or protocol definition must clearly
2007define the character encoding to be used.</t>
2008
2009<t>The transfer from URI-only to IRI-capable components requires no
2010mapping, although the conversion described in <xref
2011target="URItoIRI"/> above may be performed. It is preferable not to
2012perform this inverse conversion unless it is certain this can be done
2013correctly.</t>
2014</section>
2015
2016<section title="URI/IRI Entry">
2017
2018<t>Some components allow users to enter URIs into the system
2019by typing or dictation, for example. This software must be updated to allow
2020for IRI entry.</t>
2021
2022<t>A person viewing a visual representation of an IRI (as a sequence
2023of glyphs, in some order, in some visual display) or hearing an IRI
2024will use an entry method for characters in the user's language to
2025input the IRI. Depending on the script and the input method used, this
2026may be a more or less complicated process.</t>
2027
2028<t>The process of IRI entry must ensure, as much as possible, that the
2029restrictions defined in <xref target="abnf"/> are met. This may be
2030done by choosing appropriate input methods or variants/settings
2031thereof, by appropriately converting the characters being input, by
2032eliminating characters that cannot be converted, and/or by issuing a
2033warning or error message to the user.</t>
2034
2035<t>As an example of variant settings, input method editors for East
2036Asian Languages usually allow the input of Latin letters and related
2037characters in full-width or half-width versions. For IRI input, the
2038input method editor should be set so that it produces half-width Latin
2039letters and punctuation and full-width Katakana.</t>
2040
2041<t>An input field primarily or solely used for the input of URIs/IRIs
2042might allow the user to view an IRI as it is mapped to a URI.  Places
2043where the input of IRIs is frequent may provide the possibility for
2044viewing an IRI as mapped to a URI. This will help users when some of
2045the software they use does not yet accept IRIs.</t>
2046
2047<t>An IRI input component interfacing to components that handle URIs,
2048but not IRIs, must map the IRI to a URI before passing it to these
2049components.</t>
2050
2051<t>For the input of IRIs with right-to-left characters, please see
2052<xref target="bidiInput"></xref>.</t>
2053</section>
2054
2055<section title="URI/IRI Transfer between Applications">
2056
2057<t>Many applications (for example, mail user agents) try to detect
2058URIs appearing in plain text. For this, they use some heuristics based
2059on URI syntax. They then allow the user to click on such URIs and
2060retrieve the corresponding resource in an appropriate (usually
2061scheme-dependent) application.</t>
2062
2063<t>Such applications would need to be upgraded, in order to use the
2064IRI syntax as a base for heuristics. In particular, a non-ASCII
2065character should not be taken as the indication of the end of an IRI.
2066Such applications also would need to make sure that they correctly
2067convert the detected IRI from the character encoding of the document
2068or application where the IRI appears, to the character encoding used
2069by the system-wide IRI invocation mechanism, or to a URI (according to
2070<xref target="mapping"/>) if the system-wide invocation mechanism only
2071accepts URIs.</t>
2072
2073<t>The clipboard is another frequently used way to transfer URIs and
2074IRIs from one application to another. On most platforms, the clipboard
2075is able to store and transfer text in many languages and scripts.
2076Correctly used, the clipboard transfers characters, not octets, which
2077will do the right thing with IRIs.</t>
2078</section>
2079
2080<section title="URI/IRI Generation">
2081
2082<t>Systems that offer resources through the Internet, where those
2083resources have logical names, sometimes automatically generate URIs
2084for the resources they offer. For example, some HTTP servers can
2085generate a directory listing for a file directory and then respond to
2086the generated URIs with the files.</t>
2087
2088<t>Many legacy character encodings are in use in various file systems.
2089Many currently deployed systems do not transform the local character
2090representation of the underlying system before generating URIs.</t>
2091
2092<t>For maximum interoperability, systems that generate resource
2093identifiers should make the appropriate transformations. For example,
2094if a file system contains a file named
2095"r&amp;#xE9;sum&amp;#xE9;.html", a server should expose this as
2096"r%C3%A9sum%C3%A9.html" in a URI, which allows use of
2097"r&amp;#xE9;sum&amp;#xE9;.html" in an IRI, even if locally the file
2098name is kept in a character encoding other than UTF-8.
2099</t>
2100
2101<t>This recommendation particularly applies to HTTP servers. For FTP
2102servers, similar considerations apply; see <xref target="RFC2640"/>.</t>
2103</section>
2104
2105<section title="URI/IRI Selection" anchor="selection">
2106<t>In some cases, resource owners and publishers have control over the
2107IRIs used to identify their resources. This control is mostly
2108executed by controlling the resource names, such as file names,
2109directly.</t>
2110
2111<t>In these cases, it is recommended to avoid choosing IRIs that are
2112easily confused. For example, for US-ASCII, the lower-case ell ("l") is
2113easily confused with the digit one ("1"), and the upper-case oh ("O") is
2114easily confused with the digit zero ("0"). Publishers should avoid
2115confusing users with "br0ken" or "1ame" identifiers.</t>
2116
2117<t>Outside the US-ASCII repertoire, there are many more opportunities for
2118confusion; a complete set of guidelines is too lengthy to include
2119here. As long as names are limited to characters from a single script,
2120native writers of a given script or language will know best when
2121ambiguities can appear, and how they can be avoided. What may look
2122ambiguous to a stranger may be completely obvious to the average
2123native user. On the other hand, in some cases, the UCS contains
2124variants for compatibility reasons; for example, for typographic purposes.
2125These should be avoided wherever possible. Although there may be exceptions,
2126newly created resource names should generally be in NFKC
2127<xref target="UTR15"></xref> (which means that they are also in NFC).</t>
2128
2129<t>As an example, the UCS contains the "fi" ligature at U+FB01
2130for compatibility reasons.
2131Wherever possible, IRIs should use the two letters "f" and "i" rather
2132than the "fi" ligature. An example where the latter may be used is
2133in the query part of an IRI for an explicit search for a word written
2134containing the "fi" ligature.</t>
2135
2136<t>In certain cases, there is a chance that characters from different
2137scripts look the same. The best known example is the similarity of the
2138Latin "A", the Greek "Alpha", and the Cyrillic "A". To avoid such
2139cases, IRIs should only be created where all the characters in a
2140single component are used together in a given language. This usually
2141means that all of these characters will be from the same script, but
2142there are languages that mix characters from different scripts (such
2143as Japanese).  This is similar to the heuristics used to distinguish
2144between letters and numbers in the examples above. Also, for Latin,
2145Greek, and Cyrillic, using lowercase letters results in fewer
2146ambiguities than using uppercase letters would.</t>
2147</section>
2148
2149<section title="Display of URIs/IRIs" anchor="display">
2150<t>
2151In situations where the rendering software is not expected to display
2152non-ASCII parts of the IRI correctly using the available layout and font
2153resources, these parts should be percent-encoded before being displayed.</t>
2154
2155<t>For display of Bidi IRIs, please see <xref target="visual"/>.</t>
2156</section>
2157
2158<section title="Interpretation of URIs and IRIs">
2159<t>Software that interprets IRIs as the names of local resources should
2160accept IRIs in multiple forms and convert and match them with the
2161appropriate local resource names.</t>
2162
2163<t>First, multiple representations include both IRIs in the native
2164character encoding of the protocol and also their URI counterparts.</t>
2165
2166<t>Second, it may include URIs constructed based on character
2167encodings other than UTF-8. These URIs may be produced by user agents that do
2168not conform to this specification and that use legacy character encodings to
2169convert non-ASCII characters to URIs. Whether this is necessary, and what
2170character encodings to cover, depends on a number of factors, such as
2171the legacy character encodings used locally and the distribution of
2172various versions of user agents. For example, software for Japanese
2173may accept URIs in Shift_JIS and/or EUC-JP in addition to UTF-8.</t>
2174
2175<t>Third, it may include additional mappings to be more user-friendly
2176and robust against transmission errors. These would be similar to how
2177some servers currently treat URIs as case insensitive or perform
2178additional matching to account for spelling errors. For characters
2179beyond the US-ASCII repertoire, this may, for example, include
2180ignoring the accents on received IRIs or resource names. Please note
2181that such mappings, including case mappings, are language
2182dependent.</t>
2183
2184<t>It can be difficult to identify a resource unambiguously if too
2185many mappings are taken into consideration. However, percent-encoded
2186and not percent-encoded parts of IRIs can always be clearly distinguished.
2187Also, the regularity of UTF-8 (see <xref target="Duerst97"/>) makes the
2188potential for collisions lower than it may seem at first.</t>
2189</section>
2190
2191<section title="Upgrading Strategy">
2192<t>Where this recommendation places further constraints on software
2193for which many instances are already deployed, it is important to
2194introduce upgrades carefully and to be aware of the various
2195interdependencies.</t>
2196
2197<t>If IRIs cannot be interpreted correctly, they should not be created,
2198generated, or transported. This suggests that upgrading URI interpreting
2199software to accept IRIs should have highest priority.</t>
2200
2201<t>On the other hand, a single IRI is interpreted only by a single or
2202very few interpreters that are known in advance, although it may be
2203entered and transported very widely.</t>
2204
2205<t>Therefore, IRIs benefit most from a broad upgrade of software to be
2206able to enter and transport IRIs. However, before an
2207individual IRI is published, care should be taken to upgrade the corresponding
2208interpreting software in order to cover the forms expected to be
2209received by various versions of entry and transport software.</t>
2210
2211<t>The upgrade of generating software to generate IRIs instead of using a
2212local character encoding should happen only after the service is upgraded
2213to accept IRIs. Similarly, IRIs should only be generated when the service
2214accepts IRIs and the intervening infrastructure and protocol is known
2215to transport them safely.</t>
2216
2217<t>Software converting from URIs to IRIs for display should be upgraded
2218only after upgraded entry software has been widely deployed to the
2219population that will see the displayed result.</t>
2220
2221
2222<t>Where there is a free choice of character encodings, it is often
2223possible to reduce the effort and dependencies for upgrading to IRIs
2224by using UTF-8 rather than another encoding. For example, when a new
2225file-based Web server is set up, using UTF-8 as the character encoding
2226for file names will make the transition to IRIs easier. Likewise, when
2227a new Web form is set up using UTF-8 as the character encoding of the
2228form page, the returned query URIs will use UTF-8 as the character
2229encoding (unless the user, for whatever reason, changes the character
2230encoding) and will therefore be compatible with IRIs.</t>
2231
2232
2233<t>These recommendations, when taken together, will allow for the
2234extension from URIs to IRIs in order to handle characters other than
2235US-ASCII while minimizing interoperability problems. For
2236considerations regarding the upgrade of URI scheme definitions, see
2237<xref target="UTF8use"/>.</t>
2238
2239</section>
2240</section> <!-- guidelines -->
2241
2242<section title="IANA Considerations" anchor="iana">
2243
2244<t>RFC Editor and IANA note: Please Replace RFC XXXX with the
2245number of this document when it issues as an RFC. </t>
2246
2247<t>IANA maintains a registry of "URI schemes". A "URI scheme" also
2248serves an "IRI scheme". </t>
2249
2250<t>To clarify that the URI scheme registration process also applies to
2251IRIs, change the description of the "URI schemes" registry
2252header to say "[RFC4395] defines an IANA-maintained registry of URI
2253Schemes. These registries include the Permanent and Provisional URI
2254Schemes.  RFC XXXX updates this registry to designate that schemes may
2255also indicate their usability as IRI schemes.</t>
2256
2257<t> Update "per RFC 4395" to "per RFC 4395 and RFC XXXX".
2258</t>
2259
2260</section> <!-- IANA -->
2261   
2262<section title="Security Considerations" anchor="security">
2263<t>The security considerations discussed in <xref target="RFC3986"/>
2264also apply to IRIs. In addition, the following issues require
2265particular care for IRIs.</t>
2266<t>Incorrect encoding or decoding can lead to security problems.
2267For example, some UTF-8 decoders do not check against overlong
2268byte sequences. See <xref target='UTR36'/> Section 3 for details.</t>
2269
2270  <t>There are serious difficulties with relying on a human to verify that a
2271    an IRI (whether presented visually or aurally)
2272    is the same as another IRI or is the one intended.
2273    These problems exist with ASCII-only URIs (bl00mberg.com vs. bloomberg.com)
2274    but are strongly exacerbated when using the much larger character repertoire of Unicode.
2275    For details, see Section 2 of <xref target='UTR36'/>.
2276    Using administrative and technical means to reduce the availability
2277    of such exploits is possible, but they are difficult to eliminate altogether.
2278    User agents SHOULD NOT rely on visual or perceptual comparison or verification of IRIs
2279    as a means of validating or assuring safety, correctness or appropriateness of an IRI.
2280    Other means of presenting users with the validity, safety, or appropriateness
2281    of visited sites are being developed in the browser community
2282    as an alternative means of avoiding these difficulties.</t>
2283
2284<t>Besides the large character repertoire of Unicode, reasons for
2285  confusion include different forms of normalization and different normalization
2286  expectations, use of percent-encoding with various legacy encodings,
2287  and bidirectionality issues. See also <xref target='UTR36'/>.</t>
2288
2289<t>Confusion can occur in various IRI components, such as the
2290domain name part or the path part, or between IRI components. For considerations specific
2291to the domain name part, see <xref target="RFC5890"/>. For considerations specific to
2292particular protocols or schemes, see the security sections of the relevant specifications
2293and registration templates.
2294Administrators of sites that allow independent
2295users to create resources in the same sub area have to be careful.
2296Details are discussed in <xref target="selection"/>.</t>
2297
2298<t>Confusion can occur with bidirectional IRIs, if the restrictions
2299in <xref target="bidi-structure"/> are not followed. The same visual
2300representation may be interpreted as different logical representations,
2301and vice versa. It is also very important that a correct Unicode bidirectional
2302implementation be used.</t>
2303  <t>The characters additionally allowed in Legacy Extended IRIs
2304    introduce additional security issues. For details, see <xref target='notAllowed'/>.</t>
2305</section><!-- security -->
2306
2307<section title="Acknowledgements">
2308<t>This document was derived from <xref target="RFC3987"/>; the acknowledgments from
2309that specification still apply.</t>
2310<t>We would like to thank Ian Hickson, Michael Sperberg-McQueen,
2311  and Dan Connolly for their work on HyperText References, and Norman Walsh, Richard Tobin,
2312  Henry S. Thomson, John Cowan, Paul Grosso, and the XML Core Working Group of the W3C for their work on LEIRIs.</t>
2313<t>In addition, this document was influenced by contributions from (in no particular order) Chris
2314  Lilley, Bjoern Hoehrmann,
2315Felix Sasaki, Jeremy Carroll, Frank Ellermann, Michael Everson, Cary Karp, Matitiahu Allouche,
2316Richard Ishida, Addison Phillips, Jonathan Rosenne, Najib Tounsi, Debbie Garside, Mark Davis, Sarmad
2317Hussain, Ted Hardie, Konrad Lanz, Thomas Roessler, Lisa Dusseault, Julian Reschke, Giovanni
2318Campagna, Anne van Kesteren, Mark Nottingham, Erik van der Poel, Marcin Hanclik, Marcos Caceres, Roy
2319Fielding, Greg Wilkins, Pieter Hintjens, Daniel R. Tobias, Marko Martin, Maciej Stanchowiak, Wil
2320Tan, Yui Naruse, Michael A. Puls II, Dave Thaler, Tom Perch, John Klensin, Shawn Steele, Peter
2321Saint-Andre, Geoffrey Sneddon, Chris Weber, Alex Melnikov, Slim Amamou, SM, Tim Berners-Lee, Yaron
2322Goland, Sam Ruby, Adam Barth, Abdulrahman I. ALGhadir, Aharon Lanin, Thomas Milo, Murray Sargent,
2323Marc Blanchet, and Mykyta Yevstifeyev.</t>
2324</section>
2325
2326<section title="Main Changes Since RFC 3987">
2327  <t>This section describes the main changes since <xref target="RFC3987"></xref>.</t>
2328  <section title="Major restructuring of IRI processing model" anchor="forkChanges">
2329    <t>Major restructuring of IRI processing model to make scheme-specific translation
2330      necessary to handle IDNA requirements and for consistency with web implementations. </t>
2331    <t>Starting with IRI, you want one of:
2332      <list style="hanging">
2333        <t hangText="a"> IRI components (IRI parsed into UTF8 pieces)</t>
2334        <t hangText="b"> URI components (URI parsed into ASCII pieces, encoded correctly) </t>
2335        <t hangText="c"> whole URI  (for passing on to some other system that wants whole URIs) </t>
2336      </list></t>
2337   
2338    <section title="OLD WAY">
2339      <t><list style="numbers">
2340       
2341        <t>Pct-encoding on the whole thing to a URI.
2342          (c1) If you want a (maybe broken) whole URI, you might
2343          stop here.</t>
2344       
2345        <t>Parsing the URI into URI components.
2346          (b1) If you want (maybe broken) URI components, stop here.</t>
2347       
2348        <t> Decode the components (undoing the pct-encoding).
2349          (a) if you want IRI components, stop here.</t>
2350       
2351        <t> reencode:  Either using a different encoding some components
2352          (for domain names, and query components in web pages, which
2353          depends on the component, scheme and context), and otherwise
2354          using pct-encoding.
2355          (b2) if you want (good) URI components, stop here.</t>
2356       
2357        <t> reassemble the reencoded components.
2358          (c2) if you want a (*good*) whole URI stop here.</t>
2359      </list>
2360       
2361      </t>
2362     
2363    </section>
2364   
2365    <section title="NEW WAY">
2366      <t>
2367        <list style="numbers">
2368         
2369          <t> Parse the IRI into IRI components using the generic syntax.
2370            (a) if you want IRI components, stop here.</t>
2371         
2372          <t> Encode each components, using pct-encoding, IDN encoding, or
2373            special query part encoding depending on the component
2374            scheme or context. (b) If you want URI components, stop here.</t>
2375          <t> reassemble the a whole URI from URI components.
2376            (c) if you want a whole URI stop here.</t>
2377        </list></t>
2378    </section>
2379    <section title="Extension of Syntax">
2380      <t>Added the tag range (U+E0000-E0FFF) to the iprivate production.
2381        Some IRIs generated with the new syntax may fail to pass very strict checks
2382        relying on the old syntax. But characters in this range should be extremely infrequent
2383        anyway.</t>
2384    </section>
2385    <section title="More to be added"><t>TODO: There are more main changes that need to be
2386      documented in this section.</t></section>
2387</section>
2388
2389<section title="Change Log">
2390
2391<t>Note to RFC Editor: Please completely remove this section before publication.</t>
2392
2393<section title='Changes after draft-ietf-iri-3987bis-01'>
2394    <t>Changes from draft-ietf-iri-3987bis-01 onwards are available as changesets
2395      in the IETF tools subversion repository at
2396      http://trac.tools.ietf.org/wg/iri/trac/log/draft-ietf-iri-3987bis/draft-ietf-iri-3987bis.xml.</t>
2397</section>
2398 
2399<section title='Changes from draft-duerst-iri-bis-07 to draft-ietf-iri-3987bis-00'>
2400     <t>Changed draft name, date, last paragraph of abstract, and titles in change log, and added this section
2401     in moving from draft-duerst-iri-bis-07 (personal submission) to draft-ietf-iri-3987bis-00 (WG document).</t>
2402</section>
2403
2404<section title="Changes from -06 to -07 of draft-duerst-iri-bis">
2405  <t>Major restructuring of the processing model, see <xref target="forkChanges"></xref>.</t>
2406</section>
2407</section>
2408
2409<section title='Changes from -00 to -01'><t><list style="symbols">
2410  <t>Removed 'mailto:' before mail addresses of authors.</t>
2411  <t>Added "&lt;to be done&gt;" as right side of 'href-strip' rule. Fixed '|' to '/' for
2412    alternatives.</t>
2413</list></t>
2414</section>
2415
2416<section title="Changes from -05 to -06 of draft-duerst-iri-bis-00"><t><list style="symbols">
2417<t>Add HyperText Reference, change abstract, acks and references for it</t>
2418<t>Add Masinter back as another editor.</t>
2419<t>Masinter integrates HRef material from HTML5 spec.</t>
2420<t>Rewrite introduction sections to modernize.</t>
2421</list></t>
2422</section>
2423
2424<section title="Changes from -04 to -05 of draft-duerst-iri-bis">
2425  <t><list style="symbols">
2426    <t>Updated references.</t>
2427    <t>Changed IPR text to pre5378Trust200902.</t></list></t>
2428</section>
2429
2430<section title="Changes from -03 to -04 of draft-duerst-iri-bis">
2431  <t><list style="symbols">
2432    <t>Added explicit abbreviation for LEIRIs.</t>
2433    <t>Mentioned LEIRI references.</t>
2434    <t>Completed text in LEIRI section about tag characters and about specials.</t></list></t>
2435</section>
2436
2437<section title="Changes from -02 to -03 of draft-duerst-iri-bis">
2438  <t><list style="symbols">
2439    <t>Updated some references.</t>
2440    <t>Updated Michel Suginard's coordinates.</t></list></t>
2441</section>
2442
2443<section title="Changes from -01 to -02 of draft-duerst-iri-bis">
2444  <t><list style="symbols">
2445    <t>Added tag range to iprivate (issue private-include-tags-115).</t>
2446    <t>Added Specials (U+FFF0-FFFD) to Legacy Extended IRIs.</t></list></t>
2447</section>
2448<section title="Changes from -00 to -01 of draft-duerst-iri-bis">
2449  <t><list style="symbols">
2450    <t>Changed from "IRIs with Spaces/Controls" to "Legacy Extended IRI"
2451      based on input from the W3C XML Core WG.
2452      Moved the relevant subsections to the back and promoted them to a section.</t>
2453    <t>Added some text re. Legacy Extended IRIs to the security section.</t>
2454    <t>Added a IANA Consideration Section.</t>
2455    <t>Added this Change Log Section.</t>
2456    <t>Added a section about "IRIs with Spaces/Controls" (converting from a Note in RFC 3987).</t></list></t>
2457</section>
2458<section title="Changes from RFC 3987 to -00 of draft-duerst-iri-bis">
2459  <t><list>
2460    <t>Fixed errata (see http://www.rfc-editor.org/cgi-bin/errataSearch.pl?rfc=3987).</t></list></t>
2461</section>
2462</section>
2463</middle>
2464
2465<back>
2466<references title="Normative References">
2467
2468<reference anchor="ASCII">
2469<front>
2470<title>Coded Character Set -- 7-bit American Standard Code for Information
2471Interchange</title>
2472<author>
2473<organization>American National Standards Institute</organization>
2474</author>
2475<date year="1986"/>
2476</front>
2477<seriesInfo name="ANSI" value="X3.4"/>
2478</reference>
2479
2480<reference anchor="ISO10646">
2481<front>
2482<title>ISO/IEC 10646:2003: Information Technology -
2483Universal Multiple-Octet Coded Character Set (UCS)</title>
2484<author>
2485<organization>International Organization for Standardization</organization>
2486</author>
2487<date month="December" year="2003"/>
2488</front>
2489<seriesInfo name="ISO" value="Standard 10646"/>
2490</reference>
2491
2492&rfc2119;
2493&rfc3490;
2494&rfc3491;
2495&rfc3629;
2496&rfc3986;
2497
2498<reference anchor="STD68">
2499<front>
2500<title abbrev="ABNF">Augmented BNF for Syntax Specifications: ABNF</title>
2501<author initials="D." surname="Crocker" fullname="Dave Crocker"><organization/></author>
2502<author initials="P." surname="Overell" fullname="Paul Overell"><organization/></author>
2503<date month="January" year="2008"/></front>
2504<seriesInfo name="STD" value="68"/><seriesInfo name="RFC" value="5234"/>
2505</reference>
2506 
2507&rfc5890;
2508&rfc5891;
2509
2510<reference anchor="UNIV6">
2511<front>
2512<title>The Unicode Standard, Version 6.0.0 (Mountain View, CA, The Unicode Consortium, 2011, ISBN 978-1-936213-01-6)</title>
2513<author><organization>The Unicode Consortium</organization></author>
2514<date year="2010" month="October"/>
2515</front>
2516</reference>
2517
2518<reference anchor="UNI9" target="http://www.unicode.org/reports/tr9/tr9-13.html">
2519<front>
2520<title>The Bidirectional Algorithm</title>
2521<author initials="M." surname="Davis" fullname="Mark Davis"><organization/></author>
2522<date year="2004" month="March"/>
2523</front>
2524<seriesInfo name="Unicode Standard Annex" value="#9"/>
2525</reference>
2526
2527<reference anchor="UTR15" target="http://www.unicode.org/unicode/reports/tr15/tr15-23.html">
2528<front>
2529<title>Unicode Normalization Forms</title>
2530<author initials="M." surname="Davis" fullname="Mark Davis"><organization/></author>
2531<author initials="M.J." surname="Duerst" fullname="Martin Duerst"><organization/></author>
2532<date year="2008" month="March"/>
2533</front>
2534<seriesInfo name="Unicode Standard Annex" value="#15"/>
2535</reference>
2536
2537</references>
2538
2539<references title="Informative References">
2540
2541<reference anchor="BidiEx" target="http://www.w3.org/International/iri-edit/BidiExamples">
2542<front>
2543<title>Examples of bidirectional IRIs</title>
2544<author><organization/></author>
2545<date year="" month=""/>
2546</front>
2547</reference>
2548
2549<reference anchor="CharMod" target="http://www.w3.org/TR/charmod-resid">
2550<front>
2551<title>Character Model for the World Wide Web: Resource Identifiers</title>
2552<author initials="M." surname="Duerst" fullname="Martin Duerst"><organization/></author>
2553<author initials="F." surname="Yergeau" fullname="Francois Yergeau"><organization/></author>
2554<author initials="R." surname="Ishida" fullname="Richard Ishida"><organization/></author>
2555<author initials="M." surname="Wolf" fullname="Misha Wolf"><organization/></author>
2556<author initials="T." surname="Texin" fullname="Tex Texin"><organization/></author>
2557<date year="2004" month="November" day="25"/>
2558</front>
2559<seriesInfo name="World Wide Web Consortium" value="Candidate Recommendation"/>
2560</reference>
2561
2562<reference anchor="Duerst97" target="http://www.ifi.unizh.ch/mml/mduerst/papers/PDF/IUC11-UTF-8.pdf">
2563<front>
2564<title>The Properties and Promises of UTF-8</title>
2565<author initials="M.J." surname="Duerst" fullname="Martin Duerst"><organization/></author>
2566<date year="1997" month="September"/>
2567</front>
2568<seriesInfo name="Proc. 11th International Unicode Conference, San Jose" value=""/>
2569</reference>
2570
2571<reference anchor="Gettys" target="http://www.w3.org/DesignIssues/ModelConsequences">
2572<front>
2573<title>URI Model Consequences</title>
2574<author initials="J." surname="Gettys" fullname="Jim Gettys"><organization/></author>
2575<date month="" year=""/>
2576</front>
2577</reference>
2578
2579<reference anchor="HTML4" target="http://www.w3.org/TR/html401/appendix/notes.html#h-B.2">
2580<front>
2581<title>HTML 4.01 Specification</title>
2582<author initials="D." surname="Raggett" fullname="Dave Raggett"><organization/></author>
2583<author initials="A." surname="Le Hors" fullname="Arnaud Le Hors"><organization/></author>
2584<author initials="I." surname="Jacobs" fullname="Ian Jacobs"><organization/></author>
2585<date year="1999" month="December" day="24"/>
2586</front>
2587<seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2588</reference>
2589
2590<reference anchor="LEIRI" target="http://www.w3.org/TR/leiri/">
2591<front>
2592<title>Legacy extended IRIs for XML resource identification</title>
2593<author initials="H." surname="Thompson" fullname="Henry Thompson"><organization/></author>
2594<author initials="R." surname="Tobin"    fullname="Richard Tobin"><organization/></author>
2595<author initials="N." surname="Walsh" fullname="Norman Walsh"><organization/></author>
2596  <date year="2008" month="November" day="3"/>
2597
2598</front>
2599<seriesInfo name="World Wide Web Consortium" value="Note"/>
2600</reference>
2601
2602
2603&rfc2045;
2604&rfc2130;
2605&rfc2141;
2606&rfc2192;
2607&rfc2277;
2608&rfc2368;
2609&rfc2384;
2610&rfc2396;
2611&rfc2397;
2612&rfc2616;
2613&rfc1738;
2614&rfc2640;
2615&rfc3987;
2616&rfc6055;
2617<reference anchor='RFC4395bis'>
2618  <front>
2619    <title>Guidelines and Registration Procedures for New URI/IRI Schemes</title>
2620    <author initials='T.' surname='Hansen' fullname="Tony Hansen"><organization/></author>
2621    <author initials='T.' surname='Hardie' fullname="Ted Hardie"><organization/></author>
2622    <author initials='L.' surname='Masinter' fullname="Larry Masinter"><organization/></author>
2623    <date year="2010" month='September' day="30"/>
2624    <workgroup>IRI</workgroup>
2625  </front>
2626  <seriesInfo name="Internet-Draft" value="draft-hansen-iri-4395bis-irireg-00"/>
2627</reference>
2628 
2629 
2630<reference anchor="UNIXML" target="http://www.w3.org/TR/unicode-xml/">
2631<front>
2632<title>Unicode in XML and other Markup Languages</title>
2633<author initials="M.J." surname="Duerst" fullname="Martin Duerst"><organization/></author>
2634<author initials="A." surname="Freytag" fullname="Asmus Freytag"><organization/></author>
2635<date year="2003" month="June" day="18"/>
2636</front>
2637<seriesInfo name="Unicode Technical Report" value="#20"/>
2638<seriesInfo name="World Wide Web Consortium" value="Note"/>
2639</reference>
2640 
2641<reference anchor="UTR36" target="http://unicode.org/reports/tr36/">
2642<front>
2643<title>Unicode Security Considerations</title>
2644<author initials="M." surname="Davis" fullname="Mark Davis"><organization/></author>
2645<author initials="M." surname="Suignard" fullname="Michel Suignard"><organization/></author>
2646<date year="2010" month="August" day="4"/>
2647</front>
2648<seriesInfo name="Unicode Technical Report" value="#36"/>
2649</reference>
2650
2651<reference anchor="XLink" target="http://www.w3.org/TR/xlink/#link-locators">
2652<front>
2653<title>XML Linking Language (XLink) Version 1.0</title>
2654<author initials="S." surname="DeRose" fullname="Steve DeRose"><organization/></author>
2655<author initials="E." surname="Maler" fullname="Eve Maler"><organization/></author>
2656<author initials="D." surname="Orchard" fullname="David Orchard"><organization/></author>
2657<date year="2001" month="June" day="27"/>
2658</front>
2659<seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2660</reference>
2661
2662<reference anchor="XML1" target="http://www.w3.org/TR/REC-xml">
2663  <front>
2664    <title>Extensible Markup Language (XML) 1.0 (Forth Edition)</title>
2665    <author initials="T." surname="Bray" fullname="Tim Bray"><organization/></author>
2666    <author initials="J." surname="Paoli" fullname="Jean Paoli"><organization/></author>
2667    <author initials="C.M." surname="Sperberg-McQueen" fullname="C. M. Sperberg-McQueen">
2668      <organization/></author>
2669    <author initials="E." surname="Maler" fullname="Eve Maler"><organization/></author>
2670    <author initials="F." surname="Yergeau" fullname="Francois Yergeau"><organization/></author>
2671    <date day="16" month="August" year="2006"/>
2672  </front>
2673  <seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2674</reference>
2675
2676<reference anchor="XMLNamespace" target="http://www.w3.org/TR/REC-xml-names">
2677  <front>
2678    <title>Namespaces in XML (Second Edition)</title>
2679    <author initials="T." surname="Bray" fullname="Tim Bray"><organization/></author>
2680    <author initials="D." surname="Hollander" fullname="Dave Hollander"><organization/></author>
2681    <author initials="A." surname="Layman" fullname="Andrew Layman"><organization/></author>
2682    <author initials="R." surname="Tobin" fullname="Richard Tobin"><organization></organization></author>
2683    <date day="16" month="August" year="2006"/>
2684  </front>
2685  <seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2686</reference>
2687
2688<reference anchor="XMLSchema" target="http://www.w3.org/TR/xmlschema-2/#anyURI">
2689<front>
2690<title>XML Schema Part 2: Datatypes</title>
2691<author initials="P." surname="Biron" fullname="Paul Biron"><organization/></author>
2692<author initials="A." surname="Malhotra" fullname="Ashok Malhotra"><organization/></author>
2693<date year="2001" month="May" day="2"/>
2694</front>
2695<seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2696</reference>
2697
2698<reference anchor="XPointer" target="http://www.w3.org/TR/xptr-framework/#escaping">
2699<front>
2700<title>XPointer Framework</title>
2701<author initials="P." surname="Grosso" fullname="Paul Grosso"><organization/></author>
2702<author initials="E." surname="Maler" fullname="Eve Maler"><organization/></author>
2703<author initials="J." surname="Marsh" fullname="Jonathan Marsh"><organization/></author>
2704<author initials="N." surname="Walsh" fullname="Norman Walsh"><organization/></author>
2705<date year="2003" month="March" day="25"/>
2706</front>
2707<seriesInfo name="World Wide Web Consortium" value="Recommendation"/>
2708</reference>
2709
2710<reference anchor="HTML5" target="http://www.w3.org/TR/2009/WD-html5-20090423/">
2711<front>
2712<title>A vocabulary and associated APIs for HTML and XHTML</title>
2713<author initials="I." surname="Hickson" fullname="Ian Hickson"><organization>Google, Inc.</organization></author>
2714<author initials="D." surname="Hyatt" fullname="David Hyatt"><organization>Apple, Inc.</organization></author>
2715<date year="2009"  month="April" day="23"/>
2716</front>
2717<seriesInfo name="World Wide Web Consortium" value="Working Draft"/>
2718</reference>
2719
2720</references>
2721
2722</back>
2723</rfc>
Note: See TracBrowser for help on using the repository browser.