INTERNET-DRAFT                                          Larry Masinter
                                            Adobe Systems Incorporated
                                                         Martin Duerst
draft-masinter-url-i18n-08.txt                     W3C/Keio University
Expires May 2002                                     November 20, 2001

           Internationalized Resource Identifiers (IRI)

Status of this Memo

This document is an Internet-Draft and is in full conformance with all
provisions of Section 10 of RFC2026.

Internet-Drafts are working documents of the Internet Engineering Task
Force (IETF), its areas, and its working groups.  Note that other
groups may also distribute working documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time.  It is inappropriate to use Internet- Drafts as reference
material or to cite them other than as "work in progress."

The list of current Internet-Drafts can be accessed at

The list of Internet-Draft Shadow Directories can be accessed at

This document is not a product of any working group, but may be
discussed on the mailing list <>.  For more information on
the topic of this internet-draft, please also see [W3C IRI].


This document defines a new protocol element, an Internationalized
Resource Identifier (IRI). An IRI is a sequence of characters from
the Universal Character Set [10646]. A mapping from IRIs to URIs
[RFC 2396] is defined, which means that IRIs can be used instead
of URIs where appropriate to identify resources.

Defining a new protocol element was preferred to extending or
changing the definition of URIs to allow a clear distinction and to
avoid incompatibilities with existing software. Guidelines for the
use and deployment of IRIs in various protocols, formats, and software
components that now deal with URIs are provided.

0. Change log

0.7 Changes from the -07 version

- Added applicability statement
- Allowed space and a few other characters in IRIs to be consistent
   with XML, XLink, XML Schema,... (and added some heavy warning).
- Changed the start of non-ASCII characters from U+0080 to U+00A0 to
   exclude C1 control characters. Changes to various sections.
- Fixed several problems in the IRI syntax
- Added wording in various places to give more prominence to
   URI references/IRI references.
- Added some text about pure data-based URIs
- Changed IRI-URI mapping to distinguish cases where input is already
   in UCS
- Removed section about bidirectionality, pointing to
- Rewrote section 2.4, adding a reference to [Duer97]
- Generalized section 3.3
- Added pointers to [UNIXML], [Duer01], [XLink], [XMLSchema]
- Made adoption of IRIs by W3C clearer
- Mentioned 'conversion as late as possible' principle
- Moved from key input to general input
- Various wording charges and fixes

1. Introduction

1.1 Overview and Motivation

A URI is defined in [RFC2326] as a sequence of characters chosen from
a limited subset of the repertoire of US-ASCII characters.  This
document defines a new protocol element, called IRI (Internationalized
Resource Identifier), by extending the syntax of URIs to a much wider
repertoire of characters. It also defines "internationalized" versions
corresponding to other constructs from [RFC2326], such as URI

The characters in URIs are frequently used for representing words of
natural languages.  Using words from natural languages in URIs is very
common. Such usage has many advantages: such URIs are easier to
memorize, easier to interpret, easier to transcribe, easier to create,
and easier to guess. For most languages other than English, however,
the natural script uses characters other than A-Z. For many people,
handling Latin characters is as difficult as handling the characters
of other scripts is for people who use only the Latin alphabet. Many
languages with non-Latin scripts do have transcriptions to Latin
letters and such transcriptions are now often used in URIs, but they
introduce additional ambiguities.

The infrastructure for the appropriate handling of characters from
local scripts is now widely deployed in local versions of operating
system and application software. Software that can handle a wide
variety of scripts and languages at the same time is getting more
and more widespread. Also, there are more and more protocols and
formats that can carry a wide range of characters.

Using characters outside of A-Z in IRIs brings with it some
difficulties; a discussion of potential problems and workarounds can
be found in the later sections of this document.

URIs often contain Internet host names embedded with them. There is
an ongoing discussion of internationalization and host names; the
specific issues of the relationship of IRIs and possible future
"internationalized" host names are not discussed here. See [IDN-URI]
for a separate proposal.

1.2 Applicability

IRIs are designed to work together with recent recommendations on URI
syntax [RFC 2718]. In order to be able to use an IRI (or IRI reference
in place of an URI (or URI reference) in a given protocol context, the
following conditions have to be met:

a. The protocol or format carrying the IRI has to be able to represent
    the non-ASCII characters in the IRI, either natively or by some
    protocol- or format-specific escaping mechanism (e.g. numeric
    character references in [XML1]).

b. The protocol or format element used has to have been designated
    to carry IRIs (e.g. by designating it to be of type anyURI in

c. In the URI scheme, or at least the actual URI in question (as
    recommended for new schemes in [RFC 2718]), the encoding of
    non-ASCII characters has to be based on UTF-8. This allows IRIs to
    be used with the URN syntax [RFC2141] as well as recent URL scheme
    definitions based on UTF-8, such as IMAP URLs [RFC 2192] and POP
    URLs [RFC 2384]). This condition may also apply to only a piece
    of an URI (reference), such as the fragment identifier.

    In cases and for pieces where another encoding than UTF-8 is used,
    and for raw binary data encoded in URIs (see [RFC 2397]), the
    functionalities of IRIs cannot be used.

Section 2 discusses the IRI syntax and conversion between IRIs and
URIs. Limitations on characters appropriate for use in IRIs and
processing of IRIs are dealt with in Section 3. Section 4 discusses
software requirements for IRIs from an operational viewpoint.
For additional discussion and examples, please see also [Duer01].

1.3 Definitions

The following definitions are used in this document; they follow the
terms in [RFC 2130] and [RFC2277]:

  character               An abstract object with a separate identity.
                          "LATIN CAPITAL LETTER A" is a character.
  octet                   8 bits
  character repertoire    A set of characters (in the Mathematical
  sequence of characters  A sequence (one after another) of characters
  sequence of octets      A sequence (one after another) of octets
  (character) encoding    A method of representing a sequence of
                          characters as a sequence of octets (maybe
                          with variants). A method of (unambiguously!)
                          converting a sequence of octets into a
                          sequence of characters.
  code point              A placeholder for a character in a character
                          encoding, for example to encode additional
                          characters in future versions of the
                          character encoding.
  charset                 The name of a parameter or attribute used
                          to identify a character encoding.

2. IRI Syntax

This section defines the syntax of Internationalized Resource
Identifiers (IRIs).

As with URIs, an IRI is defined as a sequence of characters. This
definition accommodates the fact that IRIs may be written on paper or
read over the radio as well as being transmitted over the network.
Also, the same IRI may be represented as different sequences of octets
in different protocols or documents if these protocols or documents
use different character encodings. Using the same character encoding
as the containing protocol or document assures that the characters in
the IRI can be handled (searched, converted, displayed,...) in the
same way as the rest of the protocol or document.

2.1 Summary of IRI syntax

IRIs are defined similarly to URIs in [RFC2386] (as modified by
[RFC2732]), but the class of unreserved characters is extended by
adding all the characters of the UCS (Universal Character Set,
[ISO10646]) beyond U+0080, subject to the limitations given in
Section 3.

Otherwise, the syntax and use of components and reserved characters is
the same as that in [RFC2396]. All the operations defined in
[RFC2396], such as the resolution of relative URIs, can be applied to
IRIs by IRI-processing software exactly in the same way as this is
done by URI-processing software.

Characters outside the ASCII range MUST NOT be used for syntactical
purposes such as to delimit components in newly defined schemes. As an
example, it is not allowed to use U+00A2, CENT SIGN, as a delimiter,
because it is in the 'iunreserved' category, in the same way as it is
not possible to use '-' as a delimiter, because it is in the
'unreserved' category.

2.2 ABNF for IRI References and IRIs

While it might be possible to define IRI references and IRIs merely by
their transformation to URIs, IRI references and IRIs can also be
accepted and processed directly. Therefore, an ABNF definition for
IRI references and IRIs is given here.

The following are different form [RFC2396]:

       IRI-reference  = [ absoluteIRI | relativeIRI ] [ "#" ifragment ]
       absoluteIRI    = scheme ":" ( ihier_part | iopaque_part )
       relativeIRI    = ( inet_path | iabs_path | irel_path )
                        [ "?" iquery ]
       ihier_part     = ( inet_path | iabs_path ) [ "?" iquery ]
       iopaque_part   = iric_no_slash *iric
       iric_no_slash  = iunreserved | escaped | ";" | "?" | ":" | "@" |
                       "&" | "=" | "+" | "$" | ","
       inet_path      = "//" iauthority [ iabs_path ]
       iabs_path      = "/"  ipath_segments
       irel_path      = irel_segment [ iabs_path ]
       irel_segment   = 1*( iunreserved | escaped |
                           ";" | "@" | "&" | "=" | "+" | "$" | "," )
       iauthority     = server | ireg_name
       ireg_name      = 1*( iunreserved | escaped | "$" | "," |
                           ";" | ":" | "@" | "&" | "=" | "+" )
       ipath_segments = isegment *( "/" isegment )
       isegment       = *ipchar *( ";" iparam )
       iparam         = *ipchar
       ipchar         = iunreserved | escaped |
                         ":" | "@" | "&" | "=" | "+" | "$" | ","
       iquery         = *iric
       ifragment      = *iric
       iric           = reserved | iunreserved | escaped
       iunreserved    = ichar | unreserved
       ichar          = << any character of the UCS [ISO10646] of U+00A0
                           and beyond, subject to limitations in Section
                            3.1. >> | space | delims | unwise

Please note that the space character and various delimiters are
allowed in IRIs and IRI references. This is further discussed in
section 3.1, point b.

   The following are the same as [RFC2396] as modified by [RFC2732]:

       reserved    = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
                     "$" | "," | "[" | "]"
       unreserved    = alphanum | mark
       mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" |
                       "(" | ")"
       escaped       = "%" HEXDIG HEXDIG
       server        = [ [ userinfo "@" ] hostport ]
       userinfo      = *( unreserved | escaped |
                          ";" | ":" | "&" | "=" | "+" | "$" | "," )
       hostport      = host [ ":" port ]
       host          = hostname | IPv4address | IPv6reference
       ipv6reference = "[" IPv6address "]"
       hostname      = *( domainlabel "." ) toplabel [ "." ]
       domainlabel   = alphanum | alphanum *( alphanum | "-" ) alphanum
       toplabel      = alpha | alpha *( alphanum | "-" ) alphanum
       IPv6address   = hexpart [ ":" IPv4address ]
       IPv4address   = 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT
       IPv6prefix    = hexpart "/" 1*2DIGIT
       hexpart       = hexseq | hexseq "::" [ hexseq ] | "::"
                       [ hexseq ]
       hexseq        = hex4 *( ":" hex4)
       hex4          = 1*4HEXDIG
       port          = *DIGIT

       scheme        = alpha *( alpha | digit | "+" | "-" | "." )
       alphanum      = alpha | digit
       alpha         = lowalpha | upalpha
       lowalpha = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" |
                  "j" | "k" | "l" | "m" | "n" | "o" | "p" | "q" | "r" |
                  "s" | "t" | "u" | "v" | "w" | "x" | "y" | "z"
       upalpha  = "A" | "B" | "C" | "D" | "E" | "F" | "G" | "H" | "I" |
                  "J" | "K" | "L" | "M" | "N" | "O" | "P" | "Q" | "R" |
                  "S" | "T" | "U" | "V" | "W" | "X" | "Y" | "Z"
       space    = <US-ASCII coded character 20 hexadecimal>
       delims   = "<" | ">" | "#" | "%" | <">
       unwise   = "{" | "}" | "|" | "\" | "^" | "[" | "]" | "`"

2.3 Mapping of IRIs to URIs

This section defines how to map an IRI to a URI. Everything in
this section applies also to IRI references and URI references, as
well as components thereoff (e.g. fragment identifiers).

This mapping has two purposes:

  a) Syntactical: Many URI schemes and components define additional
     syntactical restrictions not captured in Section 2.2. Such
     restrictions can be applied to IRIs by noting that IRIs are only
     valid if they map to syntactically valid URIs. This means that
     such syntactical restrictions do not have to be defined again
     on the IRI level.

  b) Interpretational: URIs identify resources in various ways. IRIs
     also indentify resources. The resource that an IRI identifies is
     the same as the one identified by the URI obtained after
     converting the IRI according to the procedure defined here.
     This means that there is no need to define the association
     between identifier and resource again on the IRI level.

This mapping is accomplished in two parts. Part I is skipped if the
input is already in an UCS-based encoding (e.g. UTF-8 or UTF-16). In
that case, it is assumed that the IRI is already in NFC.

   Part I)

   1) Represent the IRI characters as a sequence of characters from the

   2) Normalize the character sequence according to Normalization Form
      C, as defined in [UNI15].  (See further discussion in Section

   Part II)

   For each character that is disallowed in URI references, apply
   steps 3.1 through 3.3. The disallowed characters consist of all
   non-ASCII characters, plus the excluded characters listed in
   Section 2.4 of [RFC 2396], except for the number sign (#) and
   percent sign (%) and the square bracket characters re-allowed
   in [RFC 2732].

   1) Convert the character to a sequence of one or more octets
      using UTF-8 [RFC 2279].

   2) Convert each octet to %HH, where HH is the hexadecimal
      notation of the octet value. Note: This is identical to the
      escaping mechanism in Section 2.4.1 of [RFC 2396].

   3) Replace the original character by the resulting character

In step 3) of part II), characters allowed in URI references, or
octets already escaped, MUST NOT be escaped further, because they are
already in their correct escaping stage in IRIs. This means that this
mapping is similar to, but different from, the escaping applied when
including arbitrary content into some part of an URI.

The above mapping produces an URI fully conforming to [RFC 2396] out
of each IRI. In addition, it results in the identity transformation
for URIs, i.e. applying the mapping a second time will not change
anything anymore. Every URI is therefore by definition an IRI.
However, this mapping SHOULD only be applied when necessary, as late
as possible.

2.4 Converting URIs to IRIs

In some situations, it may be desirable to convert a URI into an
equivalent IRI. This section gives a procedure to do such a conversion.
However, it is important to note that it is not always possible to
eliminate all escaping in an IRI. There are many reasons for this.

a. Some escape sequences (all below %80) are necessary to distinguish
    escaped and unescaped uses of reserved characters.

b. Some escape sequences cannot be interpreted as sequences of UTF-8
    Note: Due to the regularities in the octet patterns of UTF-8,
    there is a very high probability, but no guarantee, that
    escape sequences that can be interpreted as sequences of UTF-8
    octets actually originated from UTF-8. For a detailed discussion
    of the odds, see [Duer97].

c. The conversion may result in a character that is not appropriate.
    See section 3.1 for further details.

Conversion from an URI to an IRI is done using the following steps:

   1) Convert all hexadecimal escapes (% followed by two hexadecimal
      digits) of %80 and higher to the corresponding octets. (The
      result is a sequence of intermixed 'characters' and 'octets'; it
      is very important to distinguish strictly between characters and
      octets in this procedure.)

   2) Convert all sequences of octets that are strictly legal UTF-8
      sequences to the corresponding sequences of characters based on
      UTF-8. Any subsequence that is found to correspond to a legal
      UTF-8 sequence can be converted. Note: The properties of UTF-8
      make sure that this will in all cases lead to the same result.

   3) Using UTF-8, convert back to octets all resulting characters
      that are identified as not appropriate according to Section 3.1.

   4) Convert all the remaining octets (i.e. those not affected by step
      2) and those produced by step 3) back to hexadecimal escapes.

This procedure will convert as many escaped non-ASCII characters as
possible to characters in an IRI. Because there are some choices
when applying step 3) (see Section 3.1), results may differ.

3. Considerations for use of IRIs

3.1 IRI Character Limitations

Not all characters of the UCS are appropriate for use as resource
identifiers. This section discusses the limitations on characters and
character sequences usable for IRIs. The considerations in this
section are relevant when creating IRIs and when converting from URIs
to IRIs.

Because of the large and increasing number of characters in
the UCS and the large number of situations where IRIs can be used,
it is impossible to give general rules for which characters are
allowed and which not. The following considerations are relevant:

a. The repertoire of characters allowed in each IRI component is
    limited by the definition of that component.  For example, the
    definition of host names in URIs does not currently allow hex
    escapes, or "_", or many other punctuation characters. This
    specification does not relax those limits, and so IRIs currently
    may not contain any non-ASCII characters in host names. This
    specification likewise does not extended the scheme component
    beyond US-ASCII.

    Please note that in accordance with URI practice, generic IRI
    software cannot and should not check for such limitations.

b. In the URI syntax, characters that are likely to be used to delimit
    URIs in text and print ("space", "delims", and "unwise") were
    excluded. They are included in the IRI syntax, for the following
    reasons: 1) The syntax includes many other characters that are
    not appropriate in many cases; 2) Some implementation practice
    already allows them in URI references (e.g. spaces in fragment
    identifiers); 3) It is very convenient in some cases, e.g. for
    XPointers in XML attributes; and 4) Considering context is already
    necessary in the case of URIs, for example for "&" in XML.

    However, these characters should be used only with utmost care.
    Whenever there is a chance that an IRI will get used out of a
    context where these characters can be used without harm, they
    should be escaped from the start.

c. The UCS contains many areas of "characters" which have no simple
    way of inputing them. These should be avoided. Characters that
    fall into this category include Dingbats, Mathematical
    and other symbols, ligatures and presentation forms.

d. The UCS contains many areas of characters for which there are
    strong visual look-alikes. Because of the likelihood of
    transcription errors, these also should be avoided. This includes
    the full-width equivalents of ASCII characters, half-width
    Katakana characters for Japanese, and many others. This also
    includes many look-alikes of "space", "delims", and "unwise",
    characters excluded in [RFC 2396].

e. Characters with no visual representation may not be interoperably
    entered. "Control characters" MUST NOT be used. This includes
    the traditional ranges of control characters (U+0000-U+001F and
    U+007F-U+009F) as well as other cases such as plane-14 language tag

f. Some code points are reserved for private use or for special
    encoding purposes. They are not interoperable. Code points reserved
    for private use MUST NOT be used. Code points reserved for
    surrogates MUST NOT be used.

g. Where there exist duplicate ways of encoding a certain character as
    visible to the user, Normalization Form C as defined in [UNI15]
    MUST be used.

Additional information is available from [UNIXML]. Although this
is written in a different context, it discusses many of the
categories of characters and code points not appropriate for IRIs.

For reasons of transcribability, many characters have been
excluded from IRIs above. These can nevertheless be encoded in an
IRI if necessary. They have to be escaped using the procedure
in Section 2.3. For example, a space can always be encoded in an
URI and in an IRI as %20. A non-breaking space (U+00A0) has to be
encoded as %C2%A0.

3.2 Bidirectional IRIs for Right-to-Left languages

Some UCS characters, such as those used in the Arabic and Hebrew
script, have an inherent right-to-left writing direction. IRIs
containing such characters, called bidirectional IRIs or Bidi IRIs)
require additional attention because of the non-trivial relation
between logical representation (used for digital representation as
well as when reading/spelling) and visual representation (used for
display/printing). This document does not address Bidi-specific
issues. A proposal for addressing these issues can be found in [Bidi].

3.3. Processing IRIs

Processing of relative forms of IRIs against a base is handled
straightforwardly; the algorithms of RFC 2396 may be applied directly,
treating the characters additionally allowed in IRIs in the same way
as unreserved characters in URIs. Other processing operations on IRIs
and IRI references similarly work analogous to their URI complements.

Such processing and mapping to URIs is commutative, i.e. the same
result is obtained independent of whether the processing or the
mapping is done first. If both IRIs and URIs are involved in
processing, the IRI parts SHOULD be preserved as long as possible.
For example, it is possible to create an absolute IRI from a
relative IRI and an URI base. When IRIs are compared, temporary
mapping to URIs MAY be advisable to eliminate potential differences
in the degree of escaping.

4. Software requirements

This section explains the issues and difficulties in supporting IRIs
in the same software components and operations that currently process
URIs: software interfaces that handle URIs, software that allows users
to enter URIs, software that generates URIs, software that displays
URIs, formats and protocol  that transport URIs, and software that
interprets URIs, may all require more or less modification before
functioning properly with IRIs. The considerations in this section
also apply to URI references and IRI references.

4.1 URI/IRI software interfaces

Software interfaces that handle URIs, such as URI-handling APIs and
protocols transferring URIs, need interfaces and protocol elements
that are designed to carry IRIs.

Note that although an IRI is defined as a sequence of characters,
software interfaces for URIs typically function on sequences of
octets. Thus, it is necessary to define clearly which character
encoding is used.

In case the current handling in an API or protocol is based on
US-ASCII, UTF-8 is recommended as the encoding for IRIs, because
this is compatible with US-ASCII, is in accordance with the
recommendations of [RFC 2277], and makes it easy to convert to URIs
where necessary. In any case, the encoding used must not be left

Intermediate software interfaces between IRI-capable components and
URI-only components MUST map the IRIs as per section 2.3 above, when
transferring from IRI-capable to URI-only components. However, such
a mapping SHOULD be applied as late as possible. It should not be
applied between components that are known to be able to handle IRIs.

The transfer from URI-only to IRI-capable components requires no
mapping, although the conversion described in section 2.4 above may be
performed. It is preferable not to perform this inverse conversion
when there is a chance that this cannot be done correctly.

4.2 URI/IRI entry

There are components that allow users to enter URIs into the system,
e.g., by typing or dictation.  This software must be updated to allow
for IRI entry.

A person viewing a visual representation of an IRI (as a sequence of
glyphs, in some order, in some visual display) or hearing an IRI,
will use a entry method for characters in that language to input
the IRI. Depending on the script and the input method used, this may
be a more or less complicated process.

The process of IRI entry must assure, as far as possible, that the
limitations defined in Section 3.1 are met. This may be done by
choosing appropriate input methods or variants/settings thereof, by
appropriately converting the characters being input, by eliminating
characters that cannot be converted, and/or by issuing a warning or
error message to the user.

An input field primarily or only used for the input of URIs/IRIs
should allow the user to view an IRI as converted to an URI.  Places
where the input of IRIs is frequent should provide the possibility for
viewing an IRI as converted to an URI. This will help users when some
of the software they use does not yet accept IRIs.

An IRI input component that interfaces to components that handle URIs,
but not IRIs, must escape the IRI before passing it to such a

The input of IRIs with right-to-left characters requires additional
care to keep the visual and the internal representation in synch, and
to eliminate control characters and marks used to control the display
before passing the IRI over to a resolver. IRI input fields that allow
the input of right-to-left characters MUST provide this functionality.
IRI input fields that do not provide this functionality MUST NOT allow
the input of right-to-left characters.

4.3 URI/IRI generation

Systems that are offering resources through the Internet, where those
resources have logical names, sometimes automatically generate URIs
for the resources they offer. For example, some HTTP servers can
generate a 'directory listing' for a file directory under their purview,
and then respond to the generated URIs with the files.

Many legacy character encodings are in use in various file systems.
Many currently deployed systems do not transform the local character
representation of the underlying system before generating URIs.

For maximum interoperability, systems that generate resource
identifiers should do the appropriate transformations. They should use
IRIs converted to URIs in cases where it cannot be expected that the
recipient is able to handle IRIs. Due to the way most user agents
currently work, native IRIs, encoded in UTF-8, may be used if the
recipient announces that it can interpret UTF-8. This requires that
the whole page is sent as UTF-8. If this is not possible, escaping
can always be used.

This recommendation in particular applies to HTTP servers. For FTP
servers, similar considerations apply, see in particular [RFC 2640].

4.4 URI/IRI selection

In some cases, resource owners and publishers have control over the
IRIs used to identify their resources. Such control is mostly
executed by controlling the resource names, such as file names,

In such cases, it is recommended to avoid choosing IRIs that are
easily confused. For example, for US-ASCII, the lower-case ell "l" is
easily confused with the digit one "1", and the upper-case oh "O" is
easily confused with the digit zero "0". Publishers should avoid to
unintentionally confuse users with "br0ken" or "1ame" identifiers.

Outside of the US-ASCII range, there are many more opportunities for
confusion; a complete set of guidelines is too lengthy to include
here. As long as names are limited to characters from a single script,
native writers of a given script or language will know best when
ambiguities can appear, and how they can be avoided. What may look
ambiguous to a stranger may be completely obvious to the average
native user.

Note that the limitations defined in Section 3.1 and the
recommendations given here are of a different nature.  The limitations
defined in Section 3.1 are necessary to avoid duplicate encodings that
are artifacts of digital representation and that the user has no way
to distinguish visually. On the other hand, in a given context, an
identifier such as "BOX0021" can be completely appropriate, and it is
impossible to find an algorithm that distinguishes the appropriate
from the confusing identifiers.

In certain cases, there is a chance that letters from different
scripts look the same. The best know example is the Latin 'A', the
Greek 'Alpha', and the Cyrillic 'A'. To disambiguate such cases, it
should be assumed that all letters in a component are from the same
script. This is similar to the heuristics used to distinguish between
letters and numbers in the examples above. Also, for the above three
scripts, using lower case letters results in much fewer ambiguities
than using upper-case letters.

4.5 Display of URI/IRIs

Many systems contain software that presents URIs to users as part of
the system's user interface (sometimes presenting 'friendly' URIs,
i.e., a shortened or more legible subset of the URI.)  This section
applies to this presentation, as well as to the strategy for printing
URIs in magazines, newspapers, or reading them over the radio.

Software that displays identifiers to users should follow a general
principle: "Don't display something to a user that the user would not
be able to enter." The consequences of this principle require
judgement about the availability of software that implements the entry
methods described in Section 3.2.

a) In situations where a viewer is not likely to have software that
   implements non-ASCII character entry (as described in Section 3.1),
   or where it can be expected that only a limited range of non-ASCII
   characters can be entered, any part of an IRI containing characters
   outside the range allowed in [RFC 2396] or any additions should be
   escaped before being displayed.

b) In situations where a viewer _is_ likely to have such software,
    IRIs may be displayed directly.

For display of BIDI IRIs, please see [Bidi].

4.6 Interpretation of URI/IRIs

Software that interprets IRIs as the names of local resources should
accept IRIs in multiple forms, and convert and match them with the
appropriate local resource names.

First, multiple representations includes both IRIs in the native
character encoding of the protocol and also their URI counterparts.

Second, it may include URIs constructed based on other character
encodings than UTF-8. Such URIs may be produced by user agents that do
not conform to this specification and use legacy encodings to convert
non-ASCII characters to URIs. Whether this is necessary, and what
character encodings to cover, depends on a number of factors, such as
the legacy character encodings used locally and the distribution of
various versions of user agents. For example, software for Japanese
may accept URIs in Shift_JIS and/or EUC-JP in addition to UTF-8.

Third, it may include additional mappings to be more user-friendly and
robust against transmission errors. These would be similar to how
currently some servers treat URIs as case-insensitive, or perform
additional matchings to account for spelling errors.  For characters
beyond the ASCII repertoire, this may e.g.  include ignoring the
accents on received IRIs or resource names where appropriate. Please
note that such mappings, including case mappings, are

It can be difficult to unambiguously identify a resource if too
many mappings are taken into consideration. However, escaped and
non-escaped parts of IRIs can always clearly be distinguished.
Also, the regularity of UTF-8 (see [Duer97] makes the potential
for collisions lower than it may seem at first sight.

4.7 Transportation of URI/IRIs in document formats and protocols

Document formats that transport URIs may need to be upgraded to allow
the transport of IRIs. In those cases where the document as a whole
has a native character encoding, IRIs should also be encoded in this
encoding, and converted accordingly by a parser or interpreter. IRI
characters that are not expressible in the native encoding should be
escaped according to Section 2.2, or may be escaped in another way if
the document format provides a way to do this. For example, in HTML,
XML, or SGML, numeric character references can be used.

Please note that some formats already IRIs, although they use
different terminology. HTML 4.0 [HTML4] defines the conversion from
IRIs to URIs as error-avoiding behavior. XML 1.0 [XML1], XLink
[XLink], and XML Schema [XMLSchema] and specifications based upon
them allow IRIs. Also, it is expected that all relevant new  W3C
formats and protocols will be required to handle IRIs [CharMod].

5. Upgrading strategy

As this recommendation places further constraints on software for
which many instances are already deployed, it is important to
introduce upgrades carefully, and to be aware of the various

If IRIs cannot be interpreted correctly, they should not be generated
or transported. This suggests that upgrading URI interpreting software
to accept IRIs should have highest priority.

On the other hand, a single IRI is interpreted only by a single or
very few interpreters that are known in advance, while it may be
entered and transported very widely.

Therefore, IRIs benefit most from a broad upgrade of software to be
able to enter and transport IRIs, but before publishing any
individual IRI, care should be taken to upgrade the corresponding
interpreting software in order to cover the forms expected to be
received by various versions of entry and transport software.

The upgrade of generating software to generate IRIs instead of a local
encoding should happen only after the service is upgraded to accept
IRIs. Similarly, IRIs should only be generated when the service
accepts IRIs and the intervening infrastructure and protocol is known
to transport them safely.

Display software should be upgraded only after upgraded entry software
has been widely deployed to the population that will see the displayed

These recommendations, when taken together, will allow for the
extension from URIs to IRIs in order to handle scripts other than
ASCII while minimizing interoperability problems.

6. Security Considerations

If IRI entry software normalizes the characters entered, but the
resource names on the interpreting side are not normalized
accordingly, and the interpreting software does not take this into
account, there is a possibility of "spoofing". Similar possibilities
turn up when interpreting software accepts URIs in various native
encodings or allows accents and similar things to be ignored.

"Spoofing" means that somebody may add a resource name that looks the
same or similar to the user while actually being different, or a
resource name that contains the same characters, but in a different
encoding. The added resource may pretend to be the real resource by
looking very similar, but may contain all kinds of changes that may be
difficult to spot but can cause all kinds of problems.

Conceptually, this is no different from the problems surrounding the
use of case-insensitive web servers. For example, a popular web page
with a mixed case name ( might be
"spoofed" by someone who obtains access to

However, the introduction of character normalization, of additional
mappings for user convenience, and of mappings for various encodings
may increase the number of spoofing possibilities. In some cases, in
particular for Latin-based resource names, this is usually easy to
detect because UTF-8-encoded names, when interpreted and viewed as
legacy encodings, produce mostly garbage. In other cases, when
concurrently used encodings have a similar structure, but there are no
characters that have exactly the same encoding, detection is more
difficult. A good example may be the concurrent use of Shift_JIS and
EUC-JP on a Japanese server.

Administrators of large sites which allow independent users to
create subareas may need to be careful that the aliasing rules
do not create chances for spoofing.

7. Acknowledgements

The issue addressed here has been discussed at numerous times over the
last many years; for example, there was a thread in the HTML working
group in August 1995 (under the topic of "Globalizing URIs") in the
www-international mailing list in July 1996 (under the topic of
"Internationalization and URLs"), and ad-hoc meetings at the Unicode
conferences in September 1995 and September 1997.

Thanks to Francois Yergeau, Chris Wendt, Yaron Goland, Graham Klyne,
Roy Fielding, M.T. Carrasco Benitez, James Clark, Andrea Vine, Leslie
Daigle, Makoto MURATA and many others for help with understanding the
issues and possible solutions. Thanks also to the members of the W3C
I18N Working Group and Interest Group for their work on [CharMod],
to the members of many other W3C WGs for adopting the ideas, and to
the members of the Montreal IAB Workshop on Internationalization and
Localization for a healthy review.

8. Copyright

Copyright (C) The Internet Society, 1997. All Rights Reserved.

This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph
are included on all such copies and derivative works.  However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other
than English.

The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.

This document and the information contained herein is provided on an

9. Author's addresses

          Larry Masinter
          Adobe Systems Incorporated
          Mail Stop W14
          345 Park Ave
          San Jose, CA 95110

          Tel: +1 408 536-3024

          Martin J. Duerst
          W3C/Keio University
          5322 Endo, Fujisawa
          252-8520 Japan

          Tel/Fax: +81 466 49 1170

          Note: The homepage URI of the second author is the escaped
                form of an IRI.

          Note: Please write "Duerst" with u-umlaut wherever
                possible, e.g. as "D&#252;rst" in XML and HTML.

10. References

[Bidi] M. Duerst, "Internet Identifiers and Bidirectionality", Internet
  Draft, July 2001,
  work in progress.

[CharMod] M. Duerst, F. Yergeau et al., Ed., "Character Model for the
  World Wide Web", <>, work in progress.

[Duer97] M. Duerst, "The Properties and Promizes of UTF-8", Proc.
  11th International Unicode Conference, San Jose, September 1997.

[Duer01] M. Duerst, "Internationalized Resource Identifiers:
  From Specification to Testing", Proc. 19th International Unicode
  Conference, San Jose, September 2001,

[HTML4] "HTML 4.0", World Wide Web Consortium,

[IDN-URI] M. Duerst, "Internationalized Domain Names in URIs and
  IRIs", Internet Draft, November 2001,
  work in progress.

[ISO10646] ISO/IEC, Information Technology - Universal Multiple-Octet
  Coded Character Set (UCS) - Part 1: Architecture and Basic
  Multilingual Plane, Oct. 2000, with amendments.

[RFC 2119] S. Bradner, "Key words for use in RFCs to Indicate
  Requirement Levels", March 1997.

[RFC 2141] R. Moats, "URN Syntax", May 1997.

[RFC 2192] C. Newman, "IMAP URL Scheme", September 1997.

[RFC 2277] H. Alvestrad, "IETF Policy on Character Sets and

[RFC 2279] F. Yergeau. "UTF-8, a transformation format of ISO 10646.",
  January 1998.

[RFC 2384] R. Gellens, "POP URL Scheme", August 1998.

[RFC 2396] T.Berners-Lee, R.Fielding, L.Masinter. "Uniform Resource
  Identifiers (URI): Generic Syntax." August, 1998.

[RFC 2397] L. Masinter, "The "data" URL scheme", August 1998.

[RFC 2616] R.Fielding, J.Gettys, et al, "Hypertext Transfer Protocol
  -- HTTP/1.1", June 1999.

[RFC 2640] B. Curtis, "Internationalization of the File Transfer
  Protocol", July 1999.

[RFC 2718] L. Masinter, H. Alvestrand, D. Zigmond, R. Petke,
  "Guidelines for new URL Schemes", November 1999.

[RFC 2732] R. Hinden, B. Carpenter, L. Masinter, "Format for Literal
  IPv6 Addresses in URL's", December 1999.

[UNIV3] The Unicode Consortium, "The Unicode Standard Version 3.0",
  Addison-Wesley, Reading, MA, 2000.

[UNI15] M.Davis and M.Duerst, "Unicode Normalization Forms", Unicode
  Technical Report #15, November 1999.

[UNIXML] M. Duerst and A. Freytag, "Unicode in XML and other Markup
  Languages", Unicode Technical Report #20, W3C Note 15 December
  2000. <> or

[W3C IRI] Internationalization - URIs and other identifiers

[XLink] Steve DeRose et al., "XML Linking Language (XLink) Version
  1.0", World Wide Web Consortium Recommendation 27 June 2001.

[XML1] Tim Bray et al., "Extensible Markup Language (XML) 1.0
  (Second Edition)", World Wide Web Consortium Recommendation,
  Oct. 2000. <>,
  including Erratum 26 at

[XMLSchema] P. Biron, A. Malhotra, "XML Schema Part 2: Datatypes",
  World Wide Web Consortium Recommendation 2 May 2001.