Uniform Resource Identifier (URI): Generic Syntax

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 a 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'.¶

A URI can be further classified as a locator, a name, or both. The term "Uniform Resource Locator" (URL) refers to the subset of URIs that, in addition to identifying the resource, provide a means of locating the resource by describing its primary access mechanism (e.g., its network "location"). The term "Uniform Resource Name" (URN) refers to the subset of URIs that are required to remain globally unique and persistent even when the resource ceases to exist or becomes unavailable.¶

An individual scheme does not need to be cast into one of a discrete set of URI types such as "URL", "URN", "URC", etc. Any given URI scheme may define subspaces that have the characteristics of a name, a locator, or both, often depending on the persistence and care in the assignment of identifiers by the naming authority, rather than on any quality of the URI scheme. For that reason, this specification deprecates use of the terms URL or URN to distinguish between schemes, instead using the term URI throughout.¶

Each 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.¶

Although many URI schemes are named after protocols, this does not imply that use of such a URI will result in access to the resource via the named protocol. URIs are often used in contexts that are purely for identification, just like any other identifier. Even when a URI is used to obtain a representation of a resource, that access might be through gateways, proxies, caches, and name resolution services that are independent of the protocol of the resource origin, and the resolution of some URIs may require the use of more than one protocol (e.g., both DNS and HTTP are typically used to access an "http" URI's resource when it can't be found in a local cache).¶

A parser of the generic URI syntax is capable of parsing any URI reference into its major components; once the scheme is determined, further scheme-specific parsing can be performed on the components. In other words, the URI generic syntax is a superset of the syntax of all URI schemes.¶

The following examples illustrate URIs that are in common use.

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

   gopher://gopher.tc.umn.edu:70/11/Mailing%20Lists/
      -- gopher scheme for Gopher and Gopher+ Protocol services

   http://www.ietf.org/rfc/rfc2396.txt
      -- http scheme for Hypertext Transfer Protocol services

   mailto:John.Doe@example.com
      -- 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 TELNET services

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 the `absolute' URI form of the reference. The syntax of a hierarchical URI is described in Section 3; the relative URI calculation is described in Section 5.¶

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: ¶

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

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¶

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 formal grammar defined using the Augmented Backus-Naur Form (ABNF) notation of [RFC2234]. Although the ABNF defines syntax in terms of the ASCII character encoding [ASCII], the URI syntax should be interpreted in terms of the character that the ASCII-encoded octet represents, rather than the octet encoding itself. 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 that contains it.¶

The complete URI syntax is collected in Appendix A.¶

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

Within a delimited component of a URI, a sequence of characters is 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. For example, a new scheme "foo" might be defined such that any escaped octet is keyed to the UTF-8 encoding in order to determine the corresponding Unicode character.¶

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

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.¶

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  = ALPHA / DIGIT / 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. URI normalization processes may unescape sequences in the ranges of ALPHA (%41-%5A and %61-%7A), DIGIT (%30-%39), underscore (%5F), or tilde (%7E) without fear of creating a conflict, but unescaping the other mark characters is usually counterproductive.¶

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.¶

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 a 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.¶

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.¶

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

   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. The semicolon (";") and equals ("=") characters have the reserved purpose of delimiting parameters and parameter values within a path segment. However, parameters are not significant to the parsing of relative references.¶

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

   query         = *( pchar / "/" / "?" )

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

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      = *( pchar / "/" / "?" )

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 a 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.¶

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 [HTML], 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.¶

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.¶

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.¶

This section describes an example algorithm for resolving URI references that might be relative to a given base URI. The algorithm is intended to provide a definitive result that can be used to test the output of other implementations. Implementation of the algorithm itself is not required, but the result given by an implementation must match the result that would be given by this algorithm.¶

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 (R), the following pseudocode describes an algorithm for transforming R into its target (T), which is either an absolute URI or the current document, and R's optional fragment:

   (R.scheme, R.authority, R.path, R.query, fragment) = parse(R);
      -- The URI reference is parsed into the four components and
      -- fragment identifier, as described in Section 4.3.

   if ((not validating) and (R.scheme == Base.scheme)) then
      -- A non-validating parser may ignore a scheme in the
      -- reference if it is identical to the base URI's scheme.
      undefine(R.scheme);
   endif;

   if defined(R.scheme) then
      T.scheme    = R.scheme;
      T.authority = R.authority;
      T.path      = R.path;
      T.query     = R.query;
   else
      if defined(R.authority) then
         T.authority = R.authority;
         T.path      = R.path;
         T.query     = R.query;
      else
         if (R.path == "") then
            if defined(R.query) then
               T.path  = Base.path;
               T.query = R.query;
            else
               -- An empty reference refers to the current document
               return (current-document, fragment);
            endif;
         else
            if (R.path starts-with "/") then
               T.path = R.path;
            else
               T.path = merge(Base.path, R.path);
            endif;
            T.query = R.query;
         endif;
         T.authority = Base.authority;
      endif;
      T.scheme = Base.scheme;
   endif;

   return (T, fragment);

The pseudocode above refers to a merge routine for merging a relative-path reference with the path of the base URI to obtain the target path. Although there are many ways to do this, we will describe a simple method using a separate string buffer: ¶

1. 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. If the base URI's path component is the empty string, then a single slash character ("/") is copied to the buffer.
2. The reference's path component is appended to the buffer string.
3. All occurrences of "./", where "." is a complete path segment, are removed from the buffer string.
4. If the buffer string ends with "." as a complete path segment, that "." is removed.
5. 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.
6. If the buffer string ends with "<segment>/..", where <segment> is a complete path segment not equal to "..", that "<segment>/.." is removed.
7. 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.
8. The remaining buffer string is the target URI's path component.

Some systems may find it more efficient to implement the merge algorithm as a pair of path 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. This may result in a loss of information if the query component contains the strings "/../" or "/./".

The resulting target URI components and fragment can be recombined to provide the absolute form of the URI reference. Using pseudocode, this would be:

   result = ""

   if defined(T.scheme) then
      append T.scheme to result;
      append ":" to result;
   endif;

   if defined(T.authority) then
      append "//" to result;
      append T.authority to result;
   endif;

   append T.path to result;

   if defined(T.query) then
      append "?" to result;
      append T.query to result;
   endif;

   if defined(fragment) then
      append "#" to result;
      append fragment to result;
   endif;

   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.

Resolution examples are provided in Appendix C.¶

Since URIs exist to identify resources, presumably they should be considered equivalent when they identify the same resource. However, such a definition of equivalence is not of much practical use, since there is no way for software to compare two resources without knowledge of their origin. For this reason, determination of equivalence or difference of URIs is based on string comparison, perhaps augmented by reference to additional rules provided by URI scheme definitions. We use the terms "different" and "equivalent" to describe the possible outcomes of such comparisons, but there are many application-dependent versions of equivalence.¶

Even though it is possible to determine that two URIs are equivalent, it is never possible to be sure that two URIs identify different resources. Therefore, comparison methods are designed to minimize false negatives while strictly avoiding false positives.¶

In testing for equivalence, it is generally unwise to directly compare relative URI references; they should be converted to their absolute forms before comparison. Furthermore, when URI references are being compared for the purpose of selecting (or avoiding) a network action, such as retrieval of a representation, it is often necessary to separate fragment identifiers from the URIs prior to comparison.¶

A variety of methods are used in practice to test URI equivalence. These methods fall into a range, distinguished by the amount of processing required and the degree to which the probability of false negatives is reduced. As noted above, false negatives cannot in principle be eliminated. In practice, their probability can be reduced, but this reduction requires more processing and is not cost-effective for all applications.¶

If this range of comparison practices is considered as a ladder, the following discussion will climb the ladder, starting with those that are cheap but have a relatively higher chance of producing false negatives, and proceeding to those that have higher computational cost and lower risk of false negatives.¶

It is in the best interests of everyone to avoid false-negatives in comparing URIs, and to only require the minimum amount of software processing for such comparisons. Those who generate and make reference to URIs can reduce the cost of processing and the risk of false negatives by consistently providing them in a form that is reasonably canonical with respect to their scheme. Specifically: ¶

• Always provide the URI scheme in lower-case characters.
• Always provide the hostname, if any, in lower-case characters.
• Only perform %-escaping where it is essential.
• Always use upper-case A-through-F characters when %-escaping.
• Use the UTF-8 character-to-octet mapping, whenever possible.
• Prevent /./ and /../ from appearing in absolute URI paths.

The choices listed above are motivated by observations that a high proportion of deployed software already use these techniques in practice for the purposes of normalization.¶

There is no guarantee that, having once used a given URI to retrieve some information, that the same information will be retievable by that URI in the future. Nor is there any guarantee that the information retrievable via that URI in the future will be observably similar to that retrieved in the past. The URI syntax does not constrain how a given scheme or authority apportions its namespace or maintains it over time. Such a guarantee can only be obtained from the person(s) controlling that namespace and the resource in question. A specific URI scheme may define 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 URI such that an attempt to perform a seemingly harmless, idempotent operation, such as the retrieval of a representation associated with a resource, will in fact cause a possibly damaging remote operation to occur. The unsafe URI 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 URI contains instructions that, when interpreted according to this other protocol, cause an unexpected operation. An example has been the use of a gopher URI to cause an unintended or impersonating message to be sent via a SMTP server.¶

Caution should be used when using any URI that specifies a TCP 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 URI 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 being performed.¶

Although the URI syntax for IPv4address only allows the common, dotted-decimal form of IPv4 address literal, many implementations that process URIs make use of platform-dependent system routines, such as gethostbyname() and inet_aton(), to translate the string literal to an actual IP address. Unfortunately, such system routines often allow and process a much larger set of formats than those described in Section 3.2.2.¶

For example, many implementations allow dotted forms of three numbers, wherein the last part is interpreted as a 16-bit quantity and placed in the right-most two bytes of the network address (e.g., a Class B network). Likewise, a dotted form of two numbers means the last part is interpreted as a 24-bit quantity and placed in the right most three bytes of the network address (Class A), and a single number (without dots) is interpreted as a 32-bit quantity and stored directly in the network address. Adding further to the confusion, some implementations allow each dotted part to be interpreted as decimal, octal, or hexadecimal, as specified in the C language (i.e., a leading 0x or 0X implies hexadecimal; otherwise, a leading 0 implies octal; otherwise, the number is interpreted as decimal).¶

These additional IP address formats are not allowed in the URI syntax due to differences between platform implementations. However, they can become a security concern if an application attempts to filter access to resources based on the IP address in string literal format. If such filtering is performed, it is recommended that literals be converted to numeric form and filtered based on the numeric value, rather than a prefix or suffix of the string form.¶

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

Because the userinfo component is rarely used and appears before the hostname in the authority component, it can be used to construct a URI that is intended to mislead a human user by appearing to identify one (trusted) naming authority while actually identifying a different authority hidden behind the noise. For example

   http://www.example.com&story=breaking_news@10.0.0.1/top_story.htm

might lead a human user to assume that the authority is 'www.example.com', whereas it is actually '10.0.0.1'. Note that the misleading userinfo could be much longer than the example above.

A misleading URI, such as the one above, is an attack on the user's preconceived notions about the meaning of a URI, rather than an attack on the software itself. User agents may be able to reduce the impact of such attacks by visually distinguishing the various components of the URI when rendered, such as by using a different color or tone to render userinfo if any is present, though there is no general panacea. More information on URI-based semantic attacks can be found in [Siedzik].¶

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.¶

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.¶

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.¶

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

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.¶

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, but is allowed for backwards compatibility.¶

IPv6 literals have been added to the list of possible identifiers for the host portion of a server component, as described by [RFC2732], with the addition of "[" and "]" to the reserved, uric, and uric-no-slash sets. Square brackets are now specified as reserved for the authority component, allowed within the opaque part of an opaque URI, and not allowed in the hierarchical syntax except for their use as delimiters for an IPv6reference within host. In order to make this change without changing the technical definition of the path, query, and fragment components, those rules were redefined to directly specify the characters allowed rather than continuing to be defined in terms of uric.¶

Since [RFC2732] defers to [RFC2373] for definition of an IPv6 literal address, which unfortunately has an incorrect ABNF description of IPv6address, we created a new ABNF rule for IPv6address that matches the text representations defined by Section 2.2 of [RFC2373]. Likewise, the definition of IPv4address has been improved in order to limit each decimal octet to the range 0-255, and the definition of hostname has been improved to better specify length limitations and partially-qualified domain names.¶

Section 6 on URI normalization and comparison has been completely rewritten and extended using input from Tim Bray and discussion within the W3C Technical Architecture Group.¶

The ad-hoc BNF syntax has been replaced with the ABNF of [RFC2234]. This change required all rule names that formerly included underscore characters to be renamed with a dash instead. Likewise, absoluteURI and relativeURI have been changed to absolute-URI and relative-URI, respectively, for consistency.¶

The ABNF of hier-part and relative-URI (Section 3) has been corrected to allow a relative URI path to be empty. This also allows an absolute-URI to consist of nothing after the "scheme:", as is present in practice with the "DAV:" namespace [RFC2518] and the "about:" URI used by many browser implementations.¶

The ABNF of qualified has been simplified to remove a parsing ambiguity without changing the allowed syntax.¶

The resolving relative references algorithm of [RFC2396] has been rewritten using pseudocode for this revision to improve clarity and fix the following issues: ¶

• [RFC2396] section 5.2, step 6a, failed to account for a base URI with no path.
• Restored the behavior of [RFC1808] where, if the the reference contains an empty path and a defined query component, then the target URI inherits the base URI's path component.