rfc2396

1. Introduction

   Uniform Resource Identifiers (URI) provide a simple and extensible
   means for identifying a resource.  This specification of URI syntax
   and semantics is derived from concepts introduced by the World Wide
   Web global information initiative, whose use of such objects dates
   from 1990 and is described in "Universal Resource Identifiers in WWW"
   [RFC1630].  The specification of URI is designed to meet the
   recommendations laid out in "Functional Recommendations for Internet
   Resource Locators" [RFC1736] and "Functional Requirements for Uniform
   Resource Names" [RFC1737].

   This document updates and merges "Uniform Resource Locators"
   [RFC1738] and "Relative Uniform Resource Locators" [RFC1808] in order
   to define a single, generic syntax for all URI.  It excludes those
   portions of RFC 1738 that defined the specific syntax of individual
   URL schemes; those portions will be updated as separate documents, as
   will the process for registration of new URI schemes.  This document
   does not discuss the issues and recommendation for dealing with
   characters outside of the US-ASCII character set [ASCII]; those
   recommendations are discussed in a separate document.

   All significant changes from the prior RFCs are noted in Appendix G.

1.1 Overview of URI

   URI are characterized by the following definitions:

      Uniform
         Uniformity provides several benefits: it allows different types
         of resource identifiers to be used in the same context, even
         when the mechanisms used to access those resources may differ;
         it allows uniform semantic interpretation of common syntactic
         conventions across different types of resource identifiers; it
         allows introduction of new types of resource identifiers
         without interfering with the way that existing identifiers are
         used; and, it allows the identifiers to be reused in many
         different contexts, thus permitting new applications or
         protocols to leverage a pre-existing, large, and widely-used
         set of resource identifiers.

      Resource
         A resource can be anything that has identity.  Familiar
         examples include an electronic document, an image, a service
         (e.g., "today's weather report for Los Angeles"), and a
         collection of other resources.  Not all resources are network
         "retrievable"; e.g., human beings, corporations, and bound
         books in a library can also be considered resources.

         The resource is the conceptual mapping to an entity or set of
         entities, not necessarily the entity which corresponds to that
         mapping at any particular instance in time.  Thus, a resource
         can remain constant even when its content---the entities to
         which it currently corresponds---changes over time, provided
         that the conceptual mapping is not changed in the process.

      Identifier
         An identifier is an object that can act as a reference to
         something that has identity.  In the case of URI, the object is
         a sequence of characters with a restricted syntax.

   Having identified a resource, a system may perform a variety of
   operations on the resource, as might be characterized by such words
   as `access', `update', `replace', or `find attributes'.

1.2. URI, URL, and URN

   A URI can be further classified as a locator, a name, or both.  The
   term "Uniform Resource Locator" (URL) refers to the subset of URI
   that identify resources via a representation of their primary access
   mechanism (e.g., their network "location"), rather than identifying
   the resource by name or by some other attribute(s) of that resource.
   The term "Uniform Resource Name" (URN) refers to the subset of URI
   that are required to remain globally unique and persistent even when
   the resource ceases to exist or becomes unavailable.

   The URI scheme (Section 3.1) defines the namespace of the URI, and
   thus may further restrict the syntax and semantics of identifiers
   using that scheme.  This specification defines those elements of the
   URI syntax that are either required of all URI schemes or are common
   to many URI schemes.  It thus defines the syntax and semantics that
   are needed to implement a scheme-independent parsing mechanism for
   URI references, such that the scheme-dependent handling of a URI can
   be postponed until the scheme-dependent semantics are needed.  We use
   the term URL below when describing syntax or semantics that only
   apply to locators.

   Although many URL schemes are named after protocols, this does not
   imply that the only way to access the URL's resource is via the named
   protocol.  Gateways, proxies, caches, and name resolution services
   might be used to access some resources, independent of the protocol
   of their origin, and the resolution of some URL may require the use
   of more than one protocol (e.g., both DNS and HTTP are typically used
   to access an "http" URL's resource when it can't be found in a local
   cache).

   A URN differs from a URL in that it's primary purpose is persistent
   labeling of a resource with an identifier.  That identifier is drawn
   from one of a set of defined namespaces, each of which has its own
   set name structure and assignment procedures.  The "urn" scheme has
   been reserved to establish the requirements for a standardized URN
   namespace, as defined in "URN Syntax" [RFC2141] and its related
   specifications.

   Most of the examples in this specification demonstrate URL, since
   they allow the most varied use of the syntax and often have a
   hierarchical namespace.  A parser of the URI syntax is capable of
   parsing both URL and URN references as a generic URI; once the scheme
   is determined, the scheme-specific parsing can be performed on the
   generic URI components.  In other words, the URI syntax is a superset
   of the syntax of all URI schemes.

1.3. Example URI

   The following examples illustrate URI that are in common use.

   ftp://ftp.is.co.za/rfc/rfc1808.txt
      -- ftp scheme for File Transfer Protocol services

   gopher://spinaltap.micro.umn.edu/00/Weather/California/Los%20Angeles
      -- gopher scheme for Gopher and Gopher+ Protocol services

   http://www.math.uio.no/faq/compression-faq/part1.html
      -- http scheme for Hypertext Transfer Protocol services

   mailto:[e-mail]
      -- mailto scheme for electronic mail addresses

   news:comp.infosystems.www.servers.unix
      -- news scheme for USENET news groups and articles

   telnet://melvyl.ucop.edu/
      -- telnet scheme for interactive services via the TELNET Protocol

1.4. Hierarchical URI and Relative Forms

   An absolute identifier refers to a resource independent of the
   context in which the identifier is used.  In contrast, a relative
   identifier refers to a resource by describing the difference within a
   hierarchical namespace between the current context and an absolute
   identifier of the resource.

   Some URI schemes support a hierarchical naming system, where the
   hierarchy of the name is denoted by a "/" delimiter separating the
   components in the scheme. This document defines a scheme-independent
   `relative' form of URI reference that can be used in conjunction with
   a `base' URI (of a hierarchical scheme) to produce another URI. The
   syntax of hierarchical URI is described in Section 3; the relative
   URI calculation is described in Section 5.

1.5. URI Transcribability

   The URI syntax was designed with global transcribability as one of
   its main concerns. A URI is a sequence of characters from a very
   limited set, i.e. the letters of the basic Latin alphabet, digits,
   and a few special characters.  A URI may be represented in a variety
   of ways: e.g., ink on paper, pixels on a screen, or a sequence of
   octets in a coded character set.  The interpretation of a URI depends
   only on the characters used and not how those characters are
   represented in a network protocol.

   The goal of transcribability can be described by a simple scenario.
   Imagine two colleagues, Sam and Kim, sitting in a pub at an
   international conference and exchanging research ideas.  Sam asks Kim
   for a location to get more information, so Kim writes the URI for the
   research site on a napkin.  Upon returning home, Sam takes out the
   napkin and types the URI into a computer, which then retrieves the
   information to which Kim referred.

   There are several design concerns revealed by the scenario:

      o  A URI is a sequence of characters, which is not always
         represented as a sequence of octets.

      o  A URI may be transcribed from a non-network source, and thus
         should consist of characters that are most likely to be able to
         be typed into a computer, within the constraints imposed by
         keyboards (and related input devices) across languages and
         locales.

      o  A URI often needs to be remembered by people, and it is easier
         for people to remember a URI when it consists of meaningful
         components.

   These design concerns are not always in alignment.  For example, it
   is often the case that the most meaningful name for a URI component
   would require characters that cannot be typed into some systems.  The
   ability to transcribe the resource identifier from one medium to
   another was considered more important than having its URI consist of
   the most meaningful of components.  In local and regional contexts

   and with improving technology, users might benefit from being able to
   use a wider range of characters; such use is not defined in this
   document.

1.6. Syntax Notation and Common Elements

   This document uses two conventions to describe and define the syntax
   for URI.  The first, called the layout form, is a general description
   of the order of components and component separators, as in

      <first>/<second>;<third>?<fourth>

   The component names are enclosed in angle-brackets and any characters
   outside angle-brackets are literal separators.  Whitespace should be
   ignored.  These descriptions are used informally and do not define
   the syntax requirements.

   The second convention is a BNF-like grammar, used to define the
   formal URI syntax.  The grammar is that of [RFC822], except that "|"
   is used to designate alternatives.  Briefly, rules are separated from
   definitions by an equal "=", indentation is used to continue a rule
   definition over more than one line, literals are quoted with "",
   parentheses "(" and ")" are used to group elements, optional elements
   are enclosed in "[" and "]" brackets, and elements may be preceded
   with <n>* to designate n or more repetitions of the following
   element; n defaults to 0.

   Unlike many specifications that use a BNF-like grammar to define the
   bytes (octets) allowed by a protocol, the URI grammar is defined in
   terms of characters.  Each literal in the grammar corresponds to the
   character it represents, rather than to the octet encoding of that
   character in any particular coded character set.  How a URI is
   represented in terms of bits and bytes on the wire is dependent upon
   the character encoding of the protocol used to transport it, or the
   charset of the document which contains it.

   The following definitions are common to many elements:

      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"

      digit    = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" |
                 "8" | "9"

      alphanum = alpha | digit

   The complete URI syntax is collected in Appendix A.

2. URI Characters and Escape Sequences

   URI consist of a restricted set of characters, primarily chosen to
   aid transcribability and usability both in computer systems and in
   non-computer communications. Characters used conventionally as
   delimiters around URI were excluded.  The restricted set of
   characters consists of digits, letters, and a few graphic symbols
   were chosen from those common to most of the character encodings and
   input facilities available to Internet users.

      uric          = reserved | unreserved | escaped

   Within a URI, characters are either used as delimiters, or to
   represent strings of data (octets) within the delimited portions.
   Octets are either represented directly by a character (using the US-
   ASCII character for that octet [ASCII]) or by an escape encoding.
   This representation is elaborated below.

2.1 URI and non-ASCII characters

   The relationship between URI and characters has been a source of
   confusion for characters that are not part of US-ASCII. To describe
   the relationship, it is useful to distinguish between a "character"
   (as a distinguishable semantic entity) and an "octet" (an 8-bit
   byte). There are two mappings, one from URI characters to octets, and
   a second from octets to original characters:

   URI character sequence->octet sequence->original character sequence

   A URI is represented as a sequence of characters, not as a sequence
   of octets. That is because URI might be "transported" by means that
   are not through a computer network, e.g., printed on paper, read over
   the radio, etc.

   A URI scheme may define a mapping from URI characters to octets;
   whether this is done depends on the scheme. Commonly, within a
   delimited component of a URI, a sequence of characters may be used to
   represent a sequence of octets. For example, the character "a"
   represents the octet 97 (decimal), while the character sequence "%",
   "0", "a" represents the octet 10 (decimal).

   There is a second translation for some resources: the sequence of
   octets defined by a component of the URI is subsequently used to
   represent a sequence of characters. A 'charset' defines this mapping.
   There are many charsets in use in Internet protocols. For example,
   UTF-8 [UTF-8] defines a mapping from sequences of octets to sequences
   of characters in the repertoire of ISO 10646.

   In the simplest case, the original character sequence contains only
   characters that are defined in US-ASCII, and the two levels of
   mapping are simple and easily invertible: each 'original character'
   is represented as the octet for the US-ASCII code for it, which is,
   in turn, represented as either the US-ASCII character, or else the
   "%" escape sequence for that octet.

   For original character sequences that contain non-ASCII characters,
   however, the situation is more difficult. Internet protocols that
   transmit octet sequences intended to represent character sequences
   are expected to provide some way of identifying the charset used, if
   there might be more than one [RFC2277].  However, there is currently
   no provision within the generic URI syntax to accomplish this
   identification. An individual URI scheme may require a single
   charset, define a default charset, or provide a way to indicate the
   charset used.

   It is expected that a systematic treatment of character encoding
   within URI will be developed as a future modification of this
   specification.

2.2. Reserved Characters

   Many URI include components consisting of or delimited by, certain
   special characters.  These characters are called "reserved", since
   their usage within the URI component is limited to their reserved
   purpose.  If the data for a URI component would conflict with the
   reserved purpose, then the conflicting data must be escaped before
   forming the URI.

      reserved    = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
                    "$" | ","

   The "reserved" syntax class above refers to those characters that are
   allowed within a URI, but which may not be allowed within a
   particular component of the generic URI syntax; they are used as
   delimiters of the components described in Section 3.

   Characters in the "reserved" set are not reserved in all contexts.
   The set of characters actually reserved within any given URI
   component is defined by that component. In general, a character is
   reserved if the semantics of the URI changes if the character is
   replaced with its escaped US-ASCII encoding.

2.3. Unreserved Characters

   Data characters that are allowed in a URI but do not have a reserved
   purpose are called unreserved.  These include upper and lower case
   letters, decimal digits, and a limited set of punctuation marks and
   symbols.

      unreserved  = alphanum | mark

      mark        = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"

   Unreserved characters can be escaped without changing the semantics
   of the URI, but this should not be done unless the URI is being used
   in a context that does not allow the unescaped character to appear.

2.4. Escape Sequences

   Data must be escaped if it does not have a representation using an
   unreserved character; this includes data that does not correspond to
   a printable character of the US-ASCII coded character set, or that
   corresponds to any US-ASCII character that is disallowed, as
   explained below.

2.4.1. Escaped Encoding

   An escaped octet is encoded as a character triplet, consisting of the
   percent character "%" followed by the two hexadecimal digits
   representing the octet code. For example, "%20" is the escaped
   encoding for the US-ASCII space character.

      escaped     = "%" hex hex
      hex         = digit | "A" | "B" | "C" | "D" | "E" | "F" |
                            "a" | "b" | "c" | "d" | "e" | "f"

2.4.2. When to Escape and Unescape

   A URI is always in an "escaped" form, since escaping or unescaping a
   completed URI might change its semantics.  Normally, the only time
   escape encodings can safely be made is when the URI is being created
   from its component parts; each component may have its own set of
   characters that are reserved, so only the mechanism responsible for
   generating or interpreting that component can determine whether or

   not escaping a character will change its semantics. Likewise, a URI
   must be separated into its components before the escaped characters
   within those components can be safely decoded.

   In some cases, data that could be represented by an unreserved
   character may appear escaped; for example, some of the unreserved
   "mark" characters are automatically escaped by some systems.  If the
   given URI scheme defines a canonicalization algorithm, then
   unreserved characters may be unescaped according to that algorithm.
   For example, "%7e" is sometimes used instead of "~" in an http URL
   path, but the two are equivalent for an http URL.

   Because the percent "%" character always has the reserved purpose of
   being the escape indicator, it must be escaped as "%25" in order to
   be used as data within a URI.  Implementers should be careful not to
   escape or unescape the same string more than once, since unescaping
   an already unescaped string might lead to misinterpreting a percent
   data character as another escaped character, or vice versa in the
   case of escaping an already escaped string.

2.4.3. Excluded US-ASCII Characters

   Although they are disallowed within the URI syntax, we include here a
   description of those US-ASCII characters that have been excluded and
   the reasons for their exclusion.

   The control characters in the US-ASCII coded character set are not
   used within a URI, both because they are non-printable and because
   they are likely to be misinterpreted by some control mechanisms.

   control     = <US-ASCII coded characters 00-1F and 7F hexadecimal>

   The space character is excluded because significant spaces may
   disappear and insignificant spaces may be introduced when URI are
   transcribed or typeset or subjected to the treatment of word-
   processing programs.  Whitespace is also used to delimit URI in many
   contexts.

   space       = <US-ASCII coded character 20 hexadecimal>

   The angle-bracket "<" and ">" and double-quote (") characters are
   excluded because they are often used as the delimiters around URI in
   text documents and protocol fields.  The character "#" is excluded
   because it is used to delimit a URI from a fragment identifier in URI
   references (Section 4). The percent character "%" is excluded because
   it is used for the encoding of escaped characters.

   delims      = "<" | ">" | "#" | "%" | <">

   Other characters are excluded because gateways and other transport
   agents are known to sometimes modify such characters, or they are
   used as delimiters.

   unwise      = "{" | "}" | "|" | "\" | "^" | "[" | "]" | "`"

   Data corresponding to excluded characters must be escaped in order to
   be properly represented within a URI.

3. URI Syntactic Components

   The URI syntax is dependent upon the scheme.  In general, absolute
   URI are written as follows:

      <scheme>:<scheme-specific-part>

   An absolute URI contains the name of the scheme being used (<scheme>)
   followed by a colon (":") and then a string (the <scheme-specific-
   part>) whose interpretation depends on the scheme.

   The URI syntax does not require that the scheme-specific-part have
   any general structure or set of semantics which is common among all
   URI.  However, a subset of URI do share a common syntax for
   representing hierarchical relationships within the namespace.  This
   "generic URI" syntax consists of a sequence of four main components:

      <scheme>://<authority><path>?<query>

   each of which, except <scheme>, may be absent from a particular URI.
   For example, some URI schemes do not allow an <authority> component,
   and others do not use a <query> component.

      absoluteURI   = scheme ":" ( hier_part | opaque_part )

   URI that are hierarchical in nature use the slash "/" character for
   separating hierarchical components.  For some file systems, a "/"
   character (used to denote the hierarchical structure of a URI) is the
   delimiter used to construct a file name hierarchy, and thus the URI
   path will look similar to a file pathname.  This does NOT imply that
   the resource is a file or that the URI maps to an actual filesystem
   pathname.

      hier_part     = ( net_path | abs_path ) [ "?" query ]

      net_path      = "//" authority [ abs_path ]

      abs_path      = "/"  path_segments

   URI that do not make use of the slash "/" character for separating
   hierarchical components are considered opaque by the generic URI
   parser.

      opaque_part   = uric_no_slash *uric

      uric_no_slash = unreserved | escaped | ";" | "?" | ":" | "@" |
                      "&" | "=" | "+" | "$" | ","

   We use the term <path> to refer to both the <abs_path> and
   <opaque_part> constructs, since they are mutually exclusive for any
   given URI and can be parsed as a single component.

3.1. Scheme Component

   Just as there are many different methods of access to resources,
   there are a variety of schemes for identifying such resources.  The
   URI syntax consists of a sequence of components separated by reserved
   characters, with the first component defining the semantics for the
   remainder of the URI string.

   Scheme names consist of a sequence of characters beginning with a
   lower case letter and followed by any combination of lower case
   letters, digits, plus ("+"), period ("."), or hyphen ("-").  For
   resiliency, programs interpreting URI should treat upper case letters
   as equivalent to lower case in scheme names (e.g., allow "HTTP" as
   well as "http").

      scheme        = alpha *( alpha | digit | "+" | "-" | "." )

   Relative URI references are distinguished from absolute URI in that
   they do not begin with a scheme name.  Instead, the scheme is
   inherited from the base URI, as described in Section 5.2.

3.2. Authority Component

   Many URI schemes include a top hierarchical element for a naming
   authority, such that the namespace defined by the remainder of the
   URI is governed by that authority.  This authority component is
   typically defined by an Internet-based server or a scheme-specific
   registry of naming authorities.

      authority     = server | reg_name

   The authority component is preceded by a double slash "//" and is
   terminated by the next slash "/", question-mark "?", or by the end of
   the URI.  Within the authority component, the characters ";", ":",
   "@", "?", and "/" are reserved.

   An authority component is not required for a URI scheme to make use
   of relative references.  A base URI without an authority component
   implies that any relative reference will also be without an authority
   component.

3.2.1. Registry-based Naming Authority

   The structure of a registry-based naming authority is specific to the
   URI scheme, but constrained to the allowed characters for an
   authority component.

      reg_name      = 1*( unreserved | escaped | "$" | "," |
                          ";" | ":" | "@" | "&" | "=" | "+" )

3.2.2. Server-based Naming Authority

   URL schemes that involve the direct use of an IP-based protocol to a
   specified server on the Internet use a common syntax for the server
   component of the URI's scheme-specific data:

      <userinfo>@<host>:<port>

   where <userinfo> may consist of a user name and, optionally, scheme-
   specific information about how to gain authorization to access the
   server.  The parts "<userinfo>@" and ":<port>" may be omitted.

      server        = [ [ userinfo "@" ] hostport ]

   The user information, if present, is followed by a commercial at-sign
   "@".

      userinfo      = *( unreserved | escaped |
                         ";" | ":" | "&" | "=" | "+" | "$" | "," )

   Some URL schemes use the format "user:password" in the userinfo
   field. This practice is NOT RECOMMENDED, because the passing of
   authentication information in clear text (such as URI) has proven to
   be a security risk in almost every case where it has been used.

   The host is a domain name of a network host, or its IPv4 address as a
   set of four decimal digit groups separated by ".".  Literal IPv6
   addresses are not supported.

      hostport      = host [ ":" port ]
      host          = hostname | IPv4address
      hostname      = *( domainlabel "." ) toplabel [ "." ]
      domainlabel   = alphanum | alphanum *( alphanum | "-" ) alphanum
      toplabel      = alpha | alpha *( alphanum | "-" ) alphanum
      IPv4address   = 1*digit "." 1*digit "." 1*digit "." 1*digit
      port          = *digit

   Hostnames take the form described in Section 3 of [RFC1034] and
   Section 2.1 of [RFC1123]: a sequence of domain labels separated by
   ".", each domain label starting and ending with an alphanumeric
   character and possibly also containing "-" characters.  The rightmost
   domain label of a fully qualified domain name will never start with a
   digit, thus syntactically distinguishing domain names from IPv4
   addresses, and may be followed by a single "." if it is necessary to
   distinguish between the complete domain name and any local domain.
   To actually be "Uniform" as a resource locator, a URL hostname should
   be a fully qualified domain name.  In practice, however, the host
   component may be a local domain literal.

      Note: A suitable representation for including a literal IPv6
      address as the host part of a URL is desired, but has not yet been
      determined or implemented in practice.

   The port is the network port number for the server.  Most schemes
   designate protocols that have a default port number.  Another port
   number may optionally be supplied, in decimal, separated from the
   host by a colon.  If the port is omitted, the default port number is
   assumed.

3.3. Path Component

   The path component contains data, specific to the authority (or the
   scheme if there is no authority component), identifying the resource
   within the scope of that scheme and authority.

      path          = [ abs_path | opaque_part ]

      path_segments = segment *( "/" segment )
      segment       = *pchar *( ";" param )
      param         = *pchar

      pchar         = unreserved | escaped |
                      ":" | "@" | "&" | "=" | "+" | "$" | ","

   The path may consist of a sequence of path segments separated by a
   single slash "/" character.  Within a path segment, the characters
   "/", ";", "=", and "?" are reserved.  Each path segment may include a
   sequence of parameters, indicated by the semicolon ";" character.
   The parameters are not significant to the parsing of relative
   references.
3.4. Query Component

   The query component is a string of information to be interpreted by
   the resource.

      query         = *uric

   Within a query component, the characters ";", "/", "?", ":", "@",
   "&", "=", "+", ",", and "$" are reserved.

4. URI References

   The term "URI-reference" is used here to denote the common usage of a
   resource identifier.  A URI reference may be absolute or relative,
   and may have additional information attached in the form of a
   fragment identifier.  However, "the URI" that results from such a
   reference includes only the absolute URI after the fragment
   identifier (if any) is removed and after any relative URI is resolved
   to its absolute form.  Although it is possible to limit the
   discussion of URI syntax and semantics to that of the absolute
   result, most usage of URI is within general URI references, and it is
   impossible to obtain the URI from such a reference without also
   parsing the fragment and resolving the relative form.

      URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]

   The syntax for relative URI is a shortened form of that for absolute
   URI, where some prefix of the URI is missing and certain path
   components ("." and "..") have a special meaning when, and only when,
   interpreting a relative path.  The relative URI syntax is defined in
   Section 5.

4.1. Fragment Identifier

   When a URI reference is used to perform a retrieval action on the
   identified resource, the optional fragment identifier, separated from
   the URI by a crosshatch ("#") character, consists of additional
   reference information to be interpreted by the user agent after the
   retrieval action has been successfully completed.  As such, it is not
   part of a URI, but is often used in conjunction with a URI.

      fragment      = *uric

   The semantics of a fragment identifier is a property of the data
   resulting from a retrieval action, regardless of the type of URI used
   in the reference.  Therefore, the format and interpretation of
   fragment identifiers is dependent on the media type [RFC2046] of the
   retrieval result.  The character restrictions described in Section 2

   for URI also apply to the fragment in a URI-reference.  Individual
   media types may define additional restrictions or structure within
   the fragment for specifying different types of "partial views" that
   can be identified within that media type.

   A fragment identifier is only meaningful when a URI reference is
   intended for retrieval and the result of that retrieval is a document
   for which the identified fragment is consistently defined.

4.2. Same-document References

   A URI reference that does not contain a URI is a reference to the
   current document.  In other words, an empty URI reference within a
   document is interpreted as a reference to the start of that document,
   and a reference containing only a fragment identifier is a reference
   to the identified fragment of that document.  Traversal of such a
   reference should not result in an additional retrieval action.
   However, if the URI reference occurs in a context that is always
   intended to result in a new request, as in the case of HTML's FORM
   element, then an empty URI reference represents the base URI of the
   current document and should be replaced by that URI when transformed
   into a request.

4.3. Parsing a URI Reference

   A URI reference is typically parsed according to the four main
   components and fragment identifier in order to determine what
   components are present and whether the reference is relative or
   absolute.  The individual components are then parsed for their
   subparts and, if not opaque, to verify their validity.

   Although the BNF defines what is allowed in each component, it is
   ambiguous in terms of differentiating between an authority component
   and a path component that begins with two slash characters.  The
   greedy algorithm is used for disambiguation: the left-most matching
   rule soaks up as much of the URI reference string as it is capable of
   matching.  In other words, the authority component wins.

   Readers familiar with regular expressions should see Appendix B for a
   concrete parsing example and test oracle.

5. Relative URI References

   It is often the case that a group or "tree" of documents has been
   constructed to serve a common purpose; the vast majority of URI in
   these documents point to resources within the tree rather than

   outside of it.  Similarly, documents located at a particular site are
   much more likely to refer to other resources at that site than to
   resources at remote sites.

   Relative addressing of URI allows document trees to be partially
   independent of their location and access scheme.  For instance, it is
   possible for a single set of hypertext documents to be simultaneously
   accessible and traversable via each of the "file", "http", and "ftp"
   schemes if the documents refer to each other using relative URI.
   Furthermore, such document trees can be moved, as a whole, without
   changing any of the relative references.  Experience within the WWW
   has demonstrated that the ability to perform relative referencing is
   necessary for the long-term usability of embedded URI.

   The syntax for relative URI takes advantage of the <hier_part> syntax
   of <absoluteURI> (Section 3) in order to express a reference that is
   relative to the namespace of another hierarchical URI.

      relativeURI   = ( net_path | abs_path | rel_path ) [ "?" query ]

   A relative reference beginning with two slash characters is termed a
   network-path reference, as defined by <net_path> in Section 3.  Such
   references are rarely used.

   A relative reference beginning with a single slash character is
   termed an absolute-path reference, as defined by <abs_path> in
   Section 3.

   A relative reference that does not begin with a scheme name or a
   slash character is termed a relative-path reference.

      rel_path      = rel_segment [ abs_path ]

      rel_segment   = 1*( unreserved | escaped |
                          ";" | "@" | "&" | "=" | "+" | "$" | "," )

   Within a relative-path reference, the complete path segments "." and
   ".." have special meanings: "the current hierarchy level" and "the
   level above this hierarchy level", respectively.  Although this is
   very similar to their use within Unix-based filesystems to indicate
   directory levels, these path components are only considered special
   when resolving a relative-path reference to its absolute form
   (Section 5.2).

   Authors should be aware that a path segment which contains a colon
   character cannot be used as the first segment of a relative URI path
   (e.g., "this:that"), because it would be mistaken for a scheme name.

   It is therefore necessary to precede such segments with other
   segments (e.g., "./this:that") in order for them to be referenced as
   a relative path.

   It is not necessary for all URI within a given scheme to be
   restricted to the <hier_part> syntax, since the hierarchical
   properties of that syntax are only necessary when relative URI are
   used within a particular document.  Documents can only make use of
   relative URI when their base URI fits within the <hier_part> syntax.
   It is assumed that any document which contains a relative reference
   will also have a base URI that obeys the syntax.  In other words,
   relative URI cannot be used within a document that has an unsuitable
   base URI.

   Some URI schemes do not allow a hierarchical syntax matching the
   <hier_part> syntax, and thus cannot use relative references.

5.1. Establishing a Base URI

   The term "relative URI" implies that there exists some absolute "base
   URI" against which the relative reference is applied.  Indeed, the
   base URI is necessary to define the semantics of any relative URI
   reference; without it, a relative reference is meaningless.  In order
   for relative URI to be usable within a document, the base URI of that
   document must be known to the parser.

   The base URI of a document can be established in one of four ways,
   listed below in order of precedence.  The order of precedence can be
   thought of in terms of layers, where the innermost defined base URI
   has the highest precedence.  This can be visualized graphically as:

      .----------------------------------------------------------.
      |  .----------------------------------------------------.  |
      |  |  .----------------------------------------------.  |  |
      |  |  |  .----------------------------------------.  |  |  |
      |  |  |  |  .----------------------------------.  |  |  |  |
      |  |  |  |  |       <relative_reference>       |  |  |  |  |
      |  |  |  |  `----------------------------------'  |  |  |  |
      |  |  |  | (5.1.1) Base URI embedded in the       |  |  |  |
      |  |  |  |         document's content             |  |  |  |
      |  |  |  `----------------------------------------'  |  |  |
      |  |  | (5.1.2) Base URI of the encapsulating entity |  |  |
      |  |  |         (message, document, or none).        |  |  |
      |  |  `----------------------------------------------'  |  |
      |  | (5.1.3) URI used to retrieve the entity            |  |
      |  `----------------------------------------------------'  |
      | (5.1.4) Default Base URI is application-dependent        |
      `----------------------------------------------------------'

5.1.1. Base URI within Document Content

   Within certain document media types, the base URI of the document can
   be embedded within the content itself such that it can be readily
   obtained by a parser.  This can be useful for descriptive documents,
   such as tables of content, which may be transmitted to others through
   protocols other than their usual retrieval context (e.g., E-Mail or
   USENET news).

   It is beyond the scope of this document to specify how, for each
   media type, the base URI can be embedded.  It is assumed that user
   agents manipulating such media types will be able to obtain the
   appropriate syntax from that media type's specification.  An example
   of how the base URI can be embedded in the Hypertext Markup Language
   (HTML) [RFC1866] is provided in Appendix D.

   A mechanism for embedding the base URI within MIME container types
   (e.g., the message and multipart types) is defined by MHTML
   [RFC2110].  Protocols that do not use the MIME message header syntax,
   but which do allow some form of tagged metainformation to be included
   within messages, may define their own syntax for defining the base
   URI as part of a message.

5.1.2. Base URI from the Encapsulating Entity

   If no base URI is embedded, the base URI of a document is defined by
   the document's retrieval context.  For a document that is enclosed
   within another entity (such as a message or another document), the
   retrieval context is that entity; thus, the default base URI of the
   document is the base URI of the entity in which the document is
   encapsulated.

5.1.3. Base URI from the Retrieval URI

   If no base URI is embedded and the document is not encapsulated
   within some other entity (e.g., the top level of a composite entity),
   then, if a URI was used to retrieve the base document, that URI shall
   be considered the base URI.  Note that if the retrieval was the
   result of a redirected request, the last URI used (i.e., that which
   resulted in the actual retrieval of the document) is the base URI.

5.1.4. Default Base URI

   If none of the conditions described in Sections 5.1.1--5.1.3 apply,
   then the base URI is defined by the context of the application.
   Since this definition is necessarily application-dependent, failing
   to define the base URI using one of the other methods may result in
   the same content being interpreted differently by different types of
   application.

   It is the responsibility of the distributor(s) of a document
   containing relative URI to ensure that the base URI for that document
   can be established.  It must be emphasized that relative URI cannot
   be used reliably in situations where the document's base URI is not
   well-defined.

5.2. Resolving Relative References to Absolute Form

   This section describes an example algorithm for resolving URI
   references that might be relative to a given base URI.

   The base URI is established according to the rules of Section 5.1 and
   parsed into the four main components as described in Section 3.  Note
   that only the scheme component is required to be present in the base
   URI; the other components may be empty or undefined.  A component is
   undefined if its preceding separator does not appear in the URI
   reference; the path component is never undefined, though it may be
   empty.  The base URI's query component is not used by the resolution
   algorithm and may be discarded.

   For each URI reference, the following steps are performed in order:

   1) The URI reference is parsed into the potential four components and
      fragment identifier, as described in Section 4.3.

   2) If the path component is empty and the scheme, authority, and
      query components are undefined, then it is a reference to the
      current document and we are done.  Otherwise, the reference URI's
      query and fragment components are defined as found (or not found)
      within the URI reference and not inherited from the base URI.

   3) If the scheme component is defined, indicating that the reference
      starts with a scheme name, then the reference is interpreted as an
      absolute URI and we are done.  Otherwise, the reference URI's
      scheme is inherited from the base URI's scheme component.

      Due to a loophole in prior specifications [RFC1630], some parsers
      allow the scheme name to be present in a relative URI if it is the
      same as the base URI scheme.  Unfortunately, this can conflict
      with the correct parsing of non-hierarchical URI.  For backwards
      compatibility, an implementation may work around such references
      by removing the scheme if it matches that of the base URI and the
      scheme is known to always use the <hier_part> syntax.  The parser
      can then continue with the steps below for the remainder of the
      reference components.  Validating parsers should mark such a
      misformed relative reference as an error.

   4) If the authority component is defined, then the reference is a
      network-path and we skip to step 7.  Otherwise, the reference
      URI's authority is inherited from the base URI's authority
      component, which will also be undefined if the URI scheme does not
      use an authority component.

   5) If the path component begins with a slash character ("/"), then
      the reference is an absolute-path and we skip to step 7.

   6) If this step is reached, then we are resolving a relative-path
      reference.  The relative path needs to be merged with the base
      URI's path.  Although there are many ways to do this, we will
      describe a simple method using a separate string buffer.

      a) All but the last segment of the base URI's path component is
         copied to the buffer.  In other words, any characters after the
         last (right-most) slash character, if any, are excluded.

      b) The reference's path component is appended to the buffer
         string.

      c) All occurrences of "./", where "." is a complete path segment,
         are removed from the buffer string.

      d) If the buffer string ends with "." as a complete path segment,
         that "." is removed.

      e) All occurrences of "<segment>/../", where <segment> is a
         complete path segment not equal to "..", are removed from the
         buffer string.  Removal of these path segments is performed
         iteratively, removing the leftmost matching pattern on each
         iteration, until no matching pattern remains.

      f) If the buffer string ends with "<segment>/..", where <segment>
         is a complete path segment not equal to "..", that
         "<segment>/.." is removed.

      g) If the resulting buffer string still begins with one or more
         complete path segments of "..", then the reference is
         considered to be in error.  Implementations may handle this
         error by retaining these components in the resolved path (i.e.,
         treating them as part of the final URI), by removing them from
         the resolved path (i.e., discarding relative levels above the
         root), or by avoiding traversal of the reference.
      h) The remaining buffer string is the reference URI's new path
         component.

   7) The resulting URI components, including any inherited from the
      base URI, are recombined to give the absolute form of the URI
      reference.  Using pseudocode, this would be

         result = ""

         if scheme is defined then
             append scheme to result
             append ":" to result

         if authority is defined then
             append "//" to result
             append authority to result

         append path to result

         if query is defined then
             append "?" to result
             append query to result

         if fragment is defined then
             append "#" to result
             append fragment to result

         return result

      Note that we must be careful to preserve the distinction between a
      component that is undefined, meaning that its separator was not
      present in the reference, and a component that is empty, meaning
      that the separator was present and was immediately followed by the
      next component separator or the end of the reference.

   The above algorithm is intended to provide an example by which the
   output of implementations can be tested -- implementation of the
   algorithm itself is not required.  For example, some systems may find
   it more efficient to implement step 6 as a pair of segment stacks
   being merged, rather than as a series of string pattern replacements.

      Note: Some WWW client applications will fail to separate the
      reference's query component from its path component before merging
      the base and reference paths in step 6 above.  This may result in
      a loss of information if the query component contains the strings
      "/../" or "/./".

   Resolution examples are provided in Appendix C.
6. URI Normalization and Equivalence

   In many cases, different URI strings may actually identify the
   identical resource. For example, the host names used in URL are
   actually case insensitive, and the URL <http://www.XEROX.com> is
   equivalent to <http://www.xerox.com>. In general, the rules for
   equivalence and definition of a normal form, if any, are scheme
   dependent. When a scheme uses elements of the common syntax, it will
   also use the common syntax equivalence rules, namely that the scheme
   and hostname are case insensitive and a URL with an explicit ":port",
   where the port is the default for the scheme, is equivalent to one
   where the port is elided.

7. Security Considerations

   A URI does not in itself pose a security threat.  Users should beware
   that there is no general guarantee that a URL, which at one time
   located a given resource, will continue to do so.  Nor is there any
   guarantee that a URL will not locate a different resource at some
   later point in time, due to the lack of any constraint on how a given
   authority apportions its namespace.  Such a guarantee can only be
   obtained from the person(s) controlling that namespace and the
   resource in question.  A specific URI scheme may include additional
   semantics, such as name persistence, if those semantics are required
   of all naming authorities for that scheme.

   It is sometimes possible to construct a URL such that an attempt to
   perform a seemingly harmless, idempotent operation, such as the
   retrieval of an entity associated with the resource, will in fact
   cause a possibly damaging remote operation to occur.  The unsafe URL
   is typically constructed by specifying a port number other than that
   reserved for the network protocol in question.  The client
   unwittingly contacts a site that is in fact running a different
   protocol.  The content of the URL contains instructions that, when
   interpreted according to this other protocol, cause an unexpected
   operation.  An example has been the use of a gopher URL to cause an
   unintended or impersonating message to be sent via a SMTP server.

   Caution should be used when using any URL that specifies a port
   number other than the default for the protocol, especially when it is
   a number within the reserved space.

   Care should be taken when a URL contains escaped delimiters for a
   given protocol (for example, CR and LF characters for telnet
   protocols) that these are not unescaped before transmission.  This
   might violate the protocol, but avoids the potential for such
   characters to be used to simulate an extra operation or parameter in
   that protocol, which might lead to an unexpected and possibly harmful
   remote operation to be performed.

   It is clearly unwise to use a URL that contains a password which is
   intended to be secret. In particular, the use of a password within
   the 'userinfo' component of a URL is strongly disrecommended except
   in those rare cases where the 'password' parameter is intended to be
   public.

8. Acknowledgements

   This document was derived from RFC 1738 [RFC1738] and RFC 1808
   [RFC1808]; the acknowledgements in those specifications still apply.
   In addition, contributions by Gisle Aas, Martin Beet, Martin Duerst,
   Jim Gettys, Martijn Koster, Dave Kristol, Daniel LaLiberte, Foteos
   Macrides, James Marshall, Ryan Moats, Keith Moore, and Lauren Wood
   are gratefully acknowledged.

9. References

   [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and
             Languages", BCP 18, RFC 2277, January 1998.

   [RFC1630] Berners-Lee, T., "Universal Resource Identifiers in WWW: A
             Unifying Syntax for the Expression of Names and Addresses
             of Objects on the Network as used in the World-Wide Web",
             RFC 1630, June 1994.

   [RFC1738] Berners-Lee, T., Masinter, L., and M. McCahill, Editors,
             "Uniform Resource Locators (URL)", RFC 1738, December 1994.

   [RFC1866] Berners-Lee T., and D. Connolly, "HyperText Markup Language
             Specification -- 2.0", RFC 1866, November 1995.

   [RFC1123] Braden, R., Editor, "Requirements for Internet Hosts --
             Application and Support", STD 3, RFC 1123, October 1989.

   [RFC822]  Crocker, D., "Standard for the Format of ARPA Internet Text
             Messages", STD 11, RFC 822, August 1982.

   [RFC1808] Fielding, R., "Relative Uniform Resource Locators", RFC
             1808, June 1995.

   [RFC2046] Freed, N., and N. Borenstein, "Multipurpose Internet Mail
             Extensions (MIME) Part Two: Media Types", RFC 2046,
             November 1996.
   [RFC1736] Kunze, J., "Functional Recommendations for Internet
             Resource Locators", RFC 1736, February 1995.

   [RFC2141] Moats, R., "URN Syntax", RFC 2141, May 1997.

   [RFC1034] Mockapetris, P., "Domain Names - Concepts and Facilities",
             STD 13, RFC 1034, November 1987.

   [RFC2110] Palme, J., and A. Hopmann, "MIME E-mail Encapsulation of
             Aggregate Documents, such as HTML (MHTML)", RFC 2110, March
             1997.

   [RFC1737] Sollins, K., and L. Masinter, "Functional Requirements for
             Uniform Resource Names", RFC 1737, December 1994.

   [ASCII]   US-ASCII. "Coded Character Set -- 7-bit American Standard
             Code for Information Interchange", ANSI X3.4-1986.

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

10. Authors' Addresses

   Tim Berners-Lee
   World Wide Web Consortium
   MIT Laboratory for Computer Science, NE43-356
   545 Technology Square
   Cambridge, MA 02139

   Fax: [phone]
   EMail: [e-mail]


   Roy T. Fielding
   Department of Information and Computer Science
   University of California, Irvine
   Irvine, CA  92697-3425

   Fax: [phone]
   EMail: [e-mail]


   Larry Masinter
   Xerox PARC
   3333 Coyote Hill Road
   Palo Alto, CA 94034

   Fax: [phone]
   EMail: [e-mail]

A. Collected BNF for URI

      URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
      absoluteURI   = scheme ":" ( hier_part | opaque_part )
      relativeURI   = ( net_path | abs_path | rel_path ) [ "?" query ]

      hier_part     = ( net_path | abs_path ) [ "?" query ]
      opaque_part   = uric_no_slash *uric

      uric_no_slash = unreserved | escaped | ";" | "?" | ":" | "@" |
                      "&" | "=" | "+" | "$" | ","

      net_path      = "//" authority [ abs_path ]
      abs_path      = "/"  path_segments
      rel_path      = rel_segment [ abs_path ]

      rel_segment   = 1*( unreserved | escaped |
                          ";" | "@" | "&" | "=" | "+" | "$" | "," )

      scheme        = alpha *( alpha | digit | "+" | "-" | "." )

      authority     = server | reg_name

      reg_name      = 1*( unreserved | escaped | "$" | "," |
                          ";" | ":" | "@" | "&" | "=" | "+" )

      server        = [ [ userinfo "@" ] hostport ]
      userinfo      = *( unreserved | escaped |
                         ";" | ":" | "&" | "=" | "+" | "$" | "," )

      hostport      = host [ ":" port ]
      host          = hostname | IPv4address
      hostname      = *( domainlabel "." ) toplabel [ "." ]
      domainlabel   = alphanum | alphanum *( alphanum | "-" ) alphanum
      toplabel      = alpha | alpha *( alphanum | "-" ) alphanum
      IPv4address   = 1*digit "." 1*digit "." 1*digit "." 1*digit
      port          = *digit

      path          = [ abs_path | opaque_part ]
      path_segments = segment *( "/" segment )
      segment       = *pchar *( ";" param )
      param         = *pchar
      pchar         = unreserved | escaped |
                      ":" | "@" | "&" | "=" | "+" | "$" | ","

      query         = *uric

      fragment      = *uric

      uric          = reserved | unreserved | escaped
      reserved      = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
                      "$" | ","
      unreserved    = alphanum | mark
      mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" |
                      "(" | ")"

      escaped       = "%" hex hex
      hex           = digit | "A" | "B" | "C" | "D" | "E" | "F" |
                              "a" | "b" | "c" | "d" | "e" | "f"

      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"
      digit    = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" |
                 "8" | "9"





























Berners-Lee, et. al.        Standards Track                    [Page 28]

RFC 2396                   URI Generic Syntax                August 1998


B. Parsing a URI Reference with a Regular Expression

   As described in Section 4.3, the generic URI syntax is not sufficient
   to disambiguate the components of some forms of URI.  Since the
   "greedy algorithm" described in that section is identical to the
   disambiguation method used by POSIX regular expressions, it is
   natural and commonplace to use a regular expression for parsing the
   potential four components and fragment identifier of a URI reference.

   The following line is the regular expression for breaking-down a URI
   reference into its components.

      ^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?
       12            3  4          5       6  7        8 9

   The numbers in the second line above are only to assist readability;
   they indicate the reference points for each subexpression (i.e., each
   paired parenthesis).  We refer to the value matched for subexpression
   <n> as $<n>.  For example, matching the above expression to

      http://www.ics.uci.edu/pub/ietf/uri/#Related

   results in the following subexpression matches:

      $1 = http:
      $2 = http
      $3 = //www.ics.uci.edu
      $4 = www.ics.uci.edu
      $5 = /pub/ietf/uri/
      $6 = <undefined>
      $7 = <undefined>
      $8 = #Related
      $9 = Related

   where <undefined> indicates that the component is not present, as is
   the case for the query component in the above example.  Therefore, we
   can determine the value of the four components and fragment as

      scheme    = $2
      authority = $4
      path      = $5
      query     = $7
      fragment  = $9

   and, going in the opposite direction, we can recreate a URI reference
   from its components using the algorithm in step 7 of Section 5.2.





Berners-Lee, et. al.        Standards Track                    [Page 29]

RFC 2396                   URI Generic Syntax                August 1998


C. Examples of Resolving Relative URI References

   Within an object with a well-defined base URI of

      http://a/b/c/d;p?q

   the relative URI would be resolved as follows:

C.1.  Normal Examples

      g:h           =  g:h
      g             =  http://a/b/c/g
      ./g           =  http://a/b/c/g
      g/            =  http://a/b/c/g/
      /g            =  http://a/g
      //g           =  http://g
      ?y            =  http://a/b/c/?y
      g?y           =  http://a/b/c/g?y
      #s            =  (current document)#s
      g#s           =  http://a/b/c/g#s
      g?y#s         =  http://a/b/c/g?y#s
      ;x            =  http://a/b/c/;x
      g;x           =  http://a/b/c/g;x
      g;x?y#s       =  http://a/b/c/g;x?y#s
      .             =  http://a/b/c/
      ./            =  http://a/b/c/
      ..            =  http://a/b/
      ../           =  http://a/b/
      ../g          =  http://a/b/g
      ../..         =  http://a/
      ../../        =  http://a/
      ../../g       =  http://a/g

C.2.  Abnormal Examples

   Although the following abnormal examples are unlikely to occur in
   normal practice, all URI parsers should be capable of resolving them
   consistently.  Each example uses the same base as above.

   An empty reference refers to the start of the current document.

      <>            =  (current document)

   Parsers must be careful in handling the case where there are more
   relative path ".." segments than there are hierarchical levels in the
   base URI's path.  Note that the ".." syntax cannot be used to change
   the authority component of a URI.




Berners-Lee, et. al.        Standards Track                    [Page 30]

RFC 2396                   URI Generic Syntax                August 1998


      ../../../g    =  http://a/../g
      ../../../../g =  http://a/../../g

   In practice, some implementations strip leading relative symbolic
   elements (".", "..") after applying a relative URI calculation, based
   on the theory that compensating for obvious author errors is better
   than allowing the request to fail.  Thus, the above two references
   will be interpreted as "http://a/g" by some implementations.

   Similarly, parsers must avoid treating "." and ".." as special when
   they are not complete components of a relative path.

      /./g          =  http://a/./g
      /../g         =  http://a/../g
      g.            =  http://a/b/c/g.
      .g            =  http://a/b/c/.g
      g..           =  http://a/b/c/g..
      ..g           =  http://a/b/c/..g

   Less likely are cases where the relative URI uses unnecessary or
   nonsensical forms of the "." and ".." complete path segments.

      ./../g        =  http://a/b/g
      ./g/.         =  http://a/b/c/g/
      g/./h         =  http://a/b/c/g/h
      g/../h        =  http://a/b/c/h
      g;x=1/./y     =  http://a/b/c/g;x=1/y
      g;x=1/../y    =  http://a/b/c/y

   All client applications remove the query component from the base URI
   before resolving relative URI.  However, some applications fail to
   separate the reference's query and/or fragment components from a
   relative path before merging it with the base path.  This error is
   rarely noticed, since typical usage of a fragment never includes the
   hierarchy ("/") character, and the query component is not normally
   used within relative references.

      g?y/./x       =  http://a/b/c/g?y/./x
      g?y/../x      =  http://a/b/c/g?y/../x
      g#s/./x       =  http://a/b/c/g#s/./x
      g#s/../x      =  http://a/b/c/g#s/../x










Berners-Lee, et. al.        Standards Track                    [Page 31]

RFC 2396                   URI Generic Syntax                August 1998


   Some parsers allow the scheme name to be present in a relative URI if
   it is the same as the base URI scheme.  This is considered to be a
   loophole in prior specifications of partial URI [RFC1630]. Its use
   should be avoided.

      http:g        =  http:g           ; for validating parsers
                    |  http://a/b/c/g   ; for backwards compatibility












































Berners-Lee, et. al.        Standards Track                    [Page 32]

RFC 2396                   URI Generic Syntax                August 1998


D. Embedding the Base URI in HTML documents

   It is useful to consider an example of how the base URI of a document
   can be embedded within the document's content.  In this appendix, we
   describe how documents written in the Hypertext Markup Language
   (HTML) [RFC1866] can include an embedded base URI.  This appendix
   does not form a part of the URI specification and should not be
   considered as anything more than a descriptive example.

   HTML defines a special element "BASE" which, when present in the
   "HEAD" portion of a document, signals that the parser should use the
   BASE element's "HREF" attribute as the base URI for resolving any
   relative URI.  The "HREF" attribute must be an absolute URI.  Note
   that, in HTML, element and attribute names are case-insensitive.  For
   example:

      <!doctype html public "-//IETF//DTD HTML//EN">
      <HTML><HEAD>
      <TITLE>An example HTML document</TITLE>
      <BASE href="http://www.ics.uci.edu/Test/a/b/c">
      </HEAD><BODY>
      ... <A href="../x">a hypertext anchor</A> ...
      </BODY></HTML>

   A parser reading the example document should interpret the given
   relative URI "../x" as representing the absolute URI

      <http://www.ics.uci.edu/Test/a/x>

   regardless of the context in which the example document was obtained.





















Berners-Lee, et. al.        Standards Track                    [Page 33]

RFC 2396                   URI Generic Syntax                August 1998


E. Recommendations for Delimiting URI in Context

   URI are often transmitted through formats that do not provide a clear
   context for their interpretation.  For example, there are many
   occasions when URI are included in plain text; examples include text
   sent in electronic mail, USENET news messages, and, most importantly,
   printed on paper.  In such cases, it is important to be able to
   delimit the URI from the rest of the text, and in particular from
   punctuation marks that might be mistaken for part of the URI.

   In practice, URI are delimited in a variety of ways, but usually
   within double-quotes "http://test.com/", angle brackets
   <http://test.com/>, or just using whitespace

                             http://test.com/

   These wrappers do not form part of the URI.

   In the case where a fragment identifier is associated with a URI
   reference, the fragment would be placed within the brackets as well
   (separated from the URI with a "#" character).

   In some cases, extra whitespace (spaces, linebreaks, tabs, etc.) may
   need to be added to break long URI across lines. The whitespace
   should be ignored when extracting the URI.

   No whitespace should be introduced after a hyphen ("-") character.
   Because some typesetters and printers may (erroneously) introduce a
   hyphen at the end of line when breaking a line, the interpreter of a
   URI containing a line break immediately after a hyphen should ignore
   all unescaped whitespace around the line break, and should be aware
   that the hyphen may or may not actually be part of the URI.

   Using <> angle brackets around each URI is especially recommended as
   a delimiting style for URI that contain whitespace.

   The prefix "URL:" (with or without a trailing space) was recommended
   as a way to used to help distinguish a URL from other bracketed
   designators, although this is not common in practice.

   For robustness, software that accepts user-typed URI should attempt
   to recognize and strip both delimiters and embedded whitespace.

   For example, the text:







Berners-Lee, et. al.        Standards Track                    [Page 34]

RFC 2396                   URI Generic Syntax                August 1998


      Yes, Jim, I found it under "http://www.w3.org/Addressing/",
      but you can probably pick it up from <ftp://ds.internic.
      net/rfc/>.  Note the warning in <http://www.ics.uci.edu/pub/
      ietf/uri/historical.html#WARNING>.

   contains the URI references

      http://www.w3.org/Addressing/
      ftp://ds.internic.net/rfc/
      http://www.ics.uci.edu/pub/ietf/uri/historical.html#WARNING









































Berners-Lee, et. al.        Standards Track                    [Page 35]

RFC 2396                   URI Generic Syntax                August 1998


F. Abbreviated URLs

   The URL syntax was designed for unambiguous reference to network
   resources and extensibility via the URL scheme.  However, as URL
   identification and usage have become commonplace, traditional media
   (television, radio, newspapers, billboards, etc.) have increasingly
   used abbreviated URL references.  That is, a reference consisting of
   only the authority and path portions of the identified resource, such
   as

      www.w3.org/Addressing/

   or simply the DNS hostname on its own.  Such references are primarily
   intended for human interpretation rather than machine, with the
   assumption that context-based heuristics are sufficient to complete
   the URL (e.g., most hostnames beginning with "www" are likely to have
   a URL prefix of "http://").  Although there is no standard set of
   heuristics for disambiguating abbreviated URL references, many client
   implementations allow them to be entered by the user and
   heuristically resolved.  It should be noted that such heuristics may
   change over time, particularly when new URL schemes are introduced.

   Since an abbreviated URL has the same syntax as a relative URL path,
   abbreviated URL references cannot be used in contexts where relative
   URLs are expected.  This limits the use of abbreviated URLs to places
   where there is no defined base URL, such as dialog boxes and off-line
   advertisements.
























Berners-Lee, et. al.        Standards Track                    [Page 36]

RFC 2396                   URI Generic Syntax                August 1998


G. Summary of Non-editorial Changes

G.1. Additions

   Section 4 (URI References) was added to stem the confusion regarding
   "what is a URI" and how to describe fragment identifiers given that
   they are not part of the URI, but are part of the URI syntax and
   parsing concerns.  In addition, it provides a reference definition
   for use by other IETF specifications (HTML, HTTP, etc.) that have
   previously attempted to redefine the URI syntax in order to account
   for the presence of fragment identifiers in URI references.

   Section 2.4 was rewritten to clarify a number of misinterpretations
   and to leave room for fully internationalized URI.

   Appendix F on abbreviated URLs was added to describe the shortened
   references often seen on television and magazine advertisements and
   explain why they are not used in other contexts.

G.2. Modifications from both RFC 1738 and RFC 1808

   Changed to URI syntax instead of just URL.

   Confusion regarding the terms "character encoding", the URI
   "character set", and the escaping of characters with %<hex><hex>
   equivalents has (hopefully) been reduced.  Many of the BNF rule names
   regarding the character sets have been changed to more accurately
   describe their purpose and to encompass all "characters" rather than
   just US-ASCII octets.  Unless otherwise noted here, these
   modifications do not affect the URI syntax.

   Both RFC 1738 and RFC 1808 refer to the "reserved" set of characters
   as if URI-interpreting software were limited to a single set of
   characters with a reserved purpose (i.e., as meaning something other
   than the data to which the characters correspond), and that this set
   was fixed by the URI scheme.  However, this has not been true in
   practice; any character that is interpreted differently when it is
   escaped is, in effect, reserved.  Furthermore, the interpreting
   engine on a HTTP server is often dependent on the resource, not just
   the URI scheme.  The description of reserved characters has been
   changed accordingly.

   The plus "+", dollar "$", and comma "," characters have been added to
   those in the "reserved" set, since they are treated as reserved
   within the query component.






Berners-Lee, et. al.        Standards Track                    [Page 37]

RFC 2396                   URI Generic Syntax                August 1998


   The tilde "~" character was added to those in the "unreserved" set,
   since it is extensively used on the Internet in spite of the
   difficulty to transcribe it with some keyboards.

   The syntax for URI scheme has been changed to require that all
   schemes begin with an alpha character.

   The "user:password" form in the previous BNF was changed to a
   "userinfo" token, and the possibility that it might be
   "user:password" made scheme specific. In particular, the use of
   passwords in the clear is not even suggested by the syntax.

   The question-mark "?" character was removed from the set of allowed
   characters for the userinfo in the authority component, since testing
   showed that many applications treat it as reserved for separating the
   query component from the rest of the URI.

   The semicolon ";" character was added to those stated as being
   reserved within the authority component, since several new schemes
   are using it as a separator within userinfo to indicate the type of
   user authentication.

   RFC 1738 specified that the path was separated from the authority
   portion of a URI by a slash.  RFC 1808 followed suit, but with a
   fudge of carrying around the separator as a "prefix" in order to
   describe the parsing algorithm.  RFC 1630 never had this problem,
   since it considered the slash to be part of the path.  In writing
   this specification, it was found to be impossible to accurately
   describe and retain the difference between the two URI
      <foo:/bar>   and   <foo:bar>
   without either considering the slash to be part of the path (as
   corresponds to actual practice) or creating a separate component just
   to hold that slash.  We chose the former.

G.3. Modifications from RFC 1738

   The definition of specific URL schemes and their scheme-specific
   syntax and semantics has been moved to separate documents.

   The URL host was defined as a fully-qualified domain name.  However,
   many URLs are used without fully-qualified domain names (in contexts
   for which the full qualification is not necessary), without any host
   (as in some file URLs), or with a host of "localhost".

   The URL port is now *digit instead of 1*digit, since systems are
   expected to handle the case where the ":" separator between host and
   port is supplied without a port.




Berners-Lee, et. al.        Standards Track                    [Page 38]

RFC 2396                   URI Generic Syntax                August 1998


   The recommendations for delimiting URI in context (Appendix E) have
   been adjusted to reflect current practice.

G.4. Modifications from RFC 1808

   RFC 1808 (Section 4) defined an empty URL reference (a reference
   containing nothing aside from the fragment identifier) as being a
   reference to the base URL.  Unfortunately, that definition could be
   interpreted, upon selection of such a reference, as a new retrieval
   action on that resource.  Since the normal intent of such references
   is for the user agent to change its view of the current document to
   the beginning of the specified fragment within that document, not to
   make an additional request of the resource, a description of how to
   correctly interpret an empty reference has been added in Section 4.

   The description of the mythical Base header field has been replaced
   with a reference to the Content-Location header field defined by
   MHTML [RFC2110].

   RFC 1808 described various schemes as either having or not having the
   properties of the generic URI syntax.  However, the only requirement
   is that the particular document containing the relative references
   have a base URI that abides by the generic URI syntax, regardless of
   the URI scheme, so the associated description has been updated to
   reflect that.

   The BNF term <net_loc> has been replaced with <authority>, since the
   latter more accurately describes its use and purpose.  Likewise, the
   authority is no longer restricted to the IP server syntax.

   Extensive testing of current client applications demonstrated that
   the majority of deployed systems do not use the ";" character to
   indicate trailing parameter information, and that the presence of a
   semicolon in a path segment does not affect the relative parsing of
   that segment.  Therefore, parameters have been removed as a separate
   component and may now appear in any path segment.  Their influence
   has been removed from the algorithm for resolving a relative URI
   reference.  The resolution examples in Appendix C have been modified
   to reflect this change.

   Implementations are now allowed to work around misformed relative
   references that are prefixed by the same scheme as the base URI, but
   only for schemes known to use the <hier_part> syntax.








Berners-Lee, et. al.        Standards Track                    [Page 39]

RFC 2396                   URI Generic Syntax                August 1998


H.  Full Copyright Statement

   Copyright (C) The Internet Society (1998).  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
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
   BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
   HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
























Berners-Lee, et. al.        Standards Track                    [Page 40]