Network Working Group T. Berners-Lee INTERNET-DRAFT MIT/LCS <draft-fielding-url-syntax-02> R. Fielding Expires six months after publication date. U.C. Irvine L. Masinter Xerox Corporation 07 December 1996 Uniform Resource Locators (URL) Status of this Memo This document is an Internet-Draft. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as ``work in progress.'' To learn the current status of any Internet-Draft, please check the ``1id-abstracts.txt'' listing contained in the Internet-Drafts Shadow Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe), munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or ftp.isi.edu (US West Coast). Issues: 1. We need to define a mechanism for using IPv6 addresses in the URL hostname which will not break existing systems too badly. 2. Section 6 (New URL Schemes) needs input from the Applications Area A.D.'s. Abstract A Uniform Resource Locator (URL) is a compact string representation of a location for use in identifying an abstract or physical resource. This document defines the general syntax and semantics of URLs, including both absolute and relative locators, and guidelines for their use and for the definition of new URL schemes. It revises and replaces the generic definitions in RFC 1738 and RFC 1808. 1. Introduction Uniform Resource Locators (URLs) provide a simple and extensible means for identifying a resource by its location. This specification of URL 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", RFC 1630 [1]. The specification of URLs is designed to meet the recommendations laid out in "Functional Recommendations for Internet Resource Locators", RFC 1736 [8]. This document updates and merges RFC 1738 "Uniform Resource Locators" [2] and RFC 1808 "Relative Uniform Resource Locators" [7] in order to define a single, general syntax for all URLs. It excludes those portions of RFC 1738 that defined the specific syntax of individual URL schemes; those portions will be updated as separate documents. All significant changes from the prior RFCs are noted in Appendix F. URLs are characterized by the following definitions: Uniform Uniformity of syntax and semantics allows the mechanism for referencing resources to be independent of the mechanism used to locate those resources and the operations applied to those resources once they have been located. New types of resources, access mechanisms, and operations can be introduced without changing the protocols and data formats that use URLs. 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. Locator A locator is an object that identifies a resource by its location. In the case of URLs, the object is a sequence of characters with a restricted syntax. An absolute locator identifies a location independent of any context, whereas a relative locator identifies a location relative to the context in which it is found. URLs are used to `locate' resources by providing an abstract identification of the resource location. Having located 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'. This specification is only concerned with the issue of identifying a resource by its location. 1.1. URL, URN, and URI URLs are a subset of Uniform Resource Identifiers (URI), which also includes the notion of Uniform Resource Names (URN). A URN differs from a URL in that it identifies a resource in a location-independent fashion (see RFC 1737, [10]). URNs are defined by a separate set of specifications. Although this specification restricts its discussion to URLs, the syntax defined is that of URI in general. Any requirements placed on the URL syntax also apply to the URI syntax. This uniform syntax for all resource identifiers allows a URN to be used in any data field that might otherwise hold a URL. 1.2. Example URLs The following examples illustrate URLs which are in common use. ftp://ds.internic.net/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.ics.uci.edu/pub/ietf/uri/ -- http scheme for Hypertext Transfer Protocol services mailto:masinter@parc.xerox.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 services via the TELNET Protocol Many other URL schemes have been defined. Section 6 describes how new schemes are defined and registered. The scheme defines the namespace of the URL. 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 URLs 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). 1.3. URL Transcribability The URL syntax has been designed to promote transcribability over all other concerns. A URL is a sequence of characters, i.e., letters, digits, and special characters. A URL 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 URL depends only on the characters used and not how those characters are represented on the wire. 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 URL for the research site on a napkin. Upon returning home, Sam takes out the napkin and types the URL into a computer, which then retrieves the information to which Kim referred. There are several design concerns revealed by the scenario: o A URL is a sequence of characters, which is not always represented as a sequence of octets. o A URL may be transcribed from a non-network source, and thus should consist of characters which are most likely to be able to be typed into a computer, within the constraints imposed by keyboards (and related input devices) across nationalities and languages. o A URL often needs to be remembered by people, and it is easier for people to remember a URL 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 URL component would require characters which cannot be typed on most keyboards. In such cases, the ability to access a resource is considered more important than having its URL consist of the most meaningful of components. 1.4. Syntax Notation and Common Elements This document uses two conventions to describe and define the syntax for Uniform Resource Locators. 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 URL syntax. The grammar is that of RFC 822 [6], 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 which use a BNF-like grammar to define the bytes (octets) allowed by a protocol, the URL 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 URL 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 | hialpha 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" hialpha = "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 URL syntax is collected in Appendix A. 2. URL Characters and Character Escaping All URLs consist of a restricted set of characters, chosen to maximize their transcribability and usability across varying computer systems, natural languages, and nationalities. This restricted set corresponds to a subset of the graphic printable characters of the US-ASCII coded character set [11]. The set of characters allowed for use within URLs can be described in three categories: reserved, unreserved, and escaped. urlchar = reserved | unreserved | escaped 2.1. Reserved Characters Many URLs include components consisting of, or delimited by, certain special characters. These characters are called "reserved", since their usage within the URL component is limited to their reserved purpose. If the data characters for a URL component would conflict with the reserved purpose, then the conflicting characters must be escaped before forming the URL. reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" This specification uses the "reserved" set to refer to those characters which are allowed within a URL, but which may not be allowed within a particular component of the generic URL syntax; they are used as delimiters of the components described in Section 4.3. Characters in the "reserved" set are not always reserved. The set of characters actually reserved within any given URL component is defined by that component. In general, a character is reserved if escaping that character would change the semantics of the URL. 2.2. Unreserved Characters Data characters which are allowed in a URL but do not have a reserved purpose are called unreserved. These include upper and lower case letters, decimal digits, and a subset of the punctuation marks and symbols found in US-ASCII. unreserved = alpha | digit | mark mark = "$" | "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")" | "," Unreserved characters can be escaped without changing the semantics of the URL, but this should not be done unless the URL is being used in a context which does not allow the unescaped character to appear. 2.3. Escaped Characters A character must be escaped if it is non-printable, if it is often used to delimit a URL from its context, if it is not found in the US-ASCII coded character set, if it is known to cause problems when passed through some e-mail gateways, or if it is being used as normal data within a component in which it is reserved. Other characters should not be escaped unless the context of their use requires it. 2.3.1. Escaped Encoding An escaped character is encoded as a character triplet, consisting of the percent character "%" followed by the two hexadecimal digits representing the character's octet code in an 8-bit coded character set. For example, "%20" is the escaped encoding for the space character. escaped = "%" hex hex hex = digit | "A" | "B" | "C" | "D" | "E" | "F" | "a" | "b" | "c" | "d" | "e" | "f" The 8-bit coded character set of the octet must be a superset of the US-ASCII coded character set, such that the US-ASCII characters have the same escaped encoding regardless of the larger octet character set. The coded character set chosen must correspond to the character set of the mechanism that will interpret the URL component in which the escaped character is used. A sequence of escape triplets are used if the character is coded as a sequence of octets. Any character, from any character set, can be included in a URL via the escaped encoding, provided that the mechanism which will interpret the URL has an octet encoding for that character. However, only that mechanism (the originator of the URL) can determine which character is represented by the octet. A client without knowledge of the origination mechanism cannot unescape the character for display. 2.3.2. When to Escape and Unescape A URL is always in an escaped form, since escaping or unescaping a completed URL might change its semantics. The only time that characters within a URL can be safely escaped is when the URL is being created from its component parts. Each component may have its own set of characters which 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 URL must be separated into its components before the escaped characters within those components can be safely unescaped. 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 URL. 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. An exception to the unescaping rules is allowed when it is known that some older systems are escaping a character that does not need to be escaped, and when it is possible to reliably discriminate between such an escaped data character and any reserved use for that character. For example, it is generally safe to unescape "%7e" when it occurs near the beginning of an http URL path, since many older systems automatically escape the "~" character even though it is unreserved. 2.3.3. Excluded Characters Although they are not used within the URL syntax, we include here a description of those characters which have been excluded and the reasons for their exclusion. excluded = control | space | delims | unwise | national All characters corresponding to the control characters in the US-ASCII coded character set are unsafe to use within a URL, 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 URLs are transcribed or typeset or subjected to the treatment of word-processing programs. Whitespace is also used to delimit URLs 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 URLs in text documents and protocol fields. The character "#" is excluded because it is used to delimit a URL from a fragment identifier in URL references. 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. unwise = "{" | "}" | "|" | "\" | "^" | "[" | "]" | "`" Finally, all other characters besides those mentioned in the above sections are excluded because they are often difficult or impossible to transcribe using traditional computer keyboards and software. national = <Any character not in the reserved, unreserved, control, space, delims, or unwise sets> Excluded characters must be escaped in order to be properly represented within a URL. However, there do exist some systems that allow characters from the "unwise" and "national" sets to be used in URL references; a robust implementation should be prepared to handle those characters when it is possible to do so. 3. URL References A common source of confusion in the use and interpretation of Uniform Resource Locators is the distinction between a reference to a URL and the URL itself. A URL reference may be absolute or relative, and may be attached to additional information in the form of a fragment identifier. However, "the URL" which results from such a reference includes only the absolute URL after the fragment identifier (if any) is removed and after any relative URL is resolved to its absolute form. Although it is possible to limit the discussion of URL syntax and semantics to that of the absolute result, most usage of URLs is within general URL references, and it is impossible to obtain the URL from such a reference without also parsing the fragment and resolving the relative form. URL-reference = [ absoluteURL | relativeURL ] [ "#" fragment ] The syntax for relative URLs is a shortened form of that for absolute URLs, where some prefix of the URL is missing and certain path components ("." and "..") have a special meaning when interpreting a relative path. When a URL reference is used to perform a retrieval action on the identified resource, the optional fragment identifier, separated from the URL 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 URL, but is often used in conjunction with a URL. The format and interpretation of fragment identifiers is dependent on the media type of the retrieved resource. fragment = *urlchar A URL reference which does not contain a URL is a reference to the current document. In other words, an empty URL 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. 4. Generic URL Syntax 4.1. Scheme Just as there are many different methods of access to resources, there are a variety of schemes for describing the location of such resources. The URL syntax consists of a sequence of components separated by reserved characters, with the first component defining the semantics for the remainder of the URL string. In general, absolute URLs are written as follows: <scheme>:<scheme-specific-part> An absolute URL 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. Scheme names consist of a sequence of characters. The lower case letters "a"--"z", digits, and the characters plus ("+"), period ("."), and hyphen ("-") are allowed. For resiliency, programs interpreting URLs should treat upper case letters as equivalent to lower case in scheme names (e.g., allow "HTTP" as well as "http"). scheme = 1*( alpha | digit | "+" | "-" | "." ) Relative URL references are distinguished from absolute URLs in that they do not begin with a scheme name. Instead, the scheme is inherited from the base URL, as described in Section 5.2. 4.2. Opaque and Hierarchical URLs The URL syntax does not require that the scheme-specific-part have any general structure or set of semantics which is common among all URLs. However, a subset of URLs do share a common syntax for representing hierarchical relationships within the locator namespace. This generic-URL syntax is used in interpreting relative URLs. absoluteURL = generic-URL | opaque-URL opaque-URL = scheme ":" *urlchar generic-URL = scheme ":" relativeURL URLs which are hierarchical in nature use the slash "/" character for separating hierarchical components. For some file systems, the "/" used to denote the hierarchical structure of a URL corresponds to the delimiter used to construct a file name hierarchy, and thus the URL path will look similar to a file pathname. This does NOT imply that the URL is a Unix pathname. 4.3. URL Syntactic Components The URL syntax is dependent upon the scheme. Some schemes use reserved characters like "?" and ";" to indicate special components, while others just consider them to be part of the path. However, most URL schemes use a common sequence of four main components to define the location of a resource <scheme>://<server><path>?<query> each of which, except <scheme>, may be absent from a particular URL. For example, some URL schemes do not allow a server component, and others do not use a query component. 4.3.1. Server Component URL schemes that involve the direct use of an IP-based protocol to a specified host on the Internet use a common syntax for the server component of the URL's scheme-specific data: <user>:<password>@<host>:<port> Some or all of the parts "<user>:<password>@", ":<password>", and ":<port>" may be excluded. The server component is preceded by a double slash "//" and is terminated by the next slash "/" or by the end of the URL. Within the server component, the characters ":", "@", "?", and "/" are reserved. server = [ [ user [ ":" password ] "@" ] hostport ] The user name and password, if present, are followed by a commercial at-sign "@". user = *( unreserved | escaped | ";" | "&" | "=" | "+" ) password = *( unreserved | escaped | ";" | "&" | "=" | "+" ) Note that an empty user name or password is different than no user name or password; there is no way to specify a password without specifying a user name. E.g., <ftp://@host.com/> has an empty user name and no password, <ftp://host.com/> has no user name, while <ftp://foo:@host.com/> has a user name of "foo" and an empty password. The host is a domain name of a network host, or its IPv4 address as a set of four decimal digit groups separated by ".". A suitable representation for IPv6 addresses has not yet been determined. hostport = host [ ":" port ] host = hostname | hostnumber hostname = *( domainlabel "." ) toplabel domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum toplabel = alpha | alpha *( alphanum | "-" ) alphanum hostnumber = 1*digit "." 1*digit "." 1*digit "." 1*digit port = *digit Domain names take the form as described in Section 3.5 of RFC 1034 [9] and Section 2.1 of RFC 1123 [5]: a sequence of domain labels separated by ".", each domain label starting and ending with an alphanumerical character and possibly also containing "-" characters. The rightmost domain label will never start with a digit, though, which syntactically distinguishes all domain names from the IP addresses. 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. A server component is not required for a URL scheme to make use of relative references. A base URL without a server component implies that any relative reference will also be without a server component. 4.3.2. Path Component The path component contains data, specific to the scheme or server, regarding the details of how the resource can be accessed. path = [ "/" ] path_segments 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. 4.3.3. Query Component The query component is a string of information to be interpreted by the resource. query = *urlchar Within a query component, the characters "/", "&", "=", and "+" are reserved. 4.4. Parsing a URL Reference A URL reference is typically parsed according to the four main components in order to determine what components are present and whether or not the reference is relative or absolute. The individual components are then parsed for their subparts and to verify their validity. A reference is parsed as if it is a generic-URL, even though it might be considered opaque by later processes. Although the BNF defines what is allowed in each component, it is ambiguous in terms of differentiating between a server 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 URL reference string as it is capable of matching. In other words, the server component wins. Readers familiar with regular expressions should see Appendix B for a concrete parsing example and test oracle. 5. Relative URL 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 URLs in these documents point to locations within the tree rather than outside of it. Similarly, documents located at a particular server are much more likely to refer to other resources on that server than to resources at remote sites. Relative addressing of URLs 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 URLs. 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 URLs. relativeURL = net_path | abs_path | rel_path A relative reference beginning with two slash characters is termed a network-path reference. Such references are rarely used. net_path = "//" server [ abs_path ] A relative reference beginning with a single slash character is termed an absolute-path reference. abs_path = "/" rel_path A relative reference which does not begin with a scheme name or a slash character is termed a relative-path reference. rel_path = [ path_segments ] [ "?" query ] 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 URL 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 URLs within a given scheme to be restricted to the generic-URL syntax, since the hierarchical properties of that syntax are only necessary when relative URLs are used within a particular document. Documents can only make use of relative URLs when their base URL fits within the generic-URL syntax. It is assumed that any document which contains a relative reference will also have a base URL that obeys the syntax. In other words, relative URLs cannot be used within a document that has an unsuitable base URL. 5.1. Establishing a Base URL The term "relative URL" implies that there exists some absolute "base URL" against which the relative reference is applied. Indeed, the base URL is necessary to define the semantics of any relative URL reference; without it, a relative reference is meaningless. In order for relative URLs to be usable within a document, the base URL of that document must be known to the parser. The base URL 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 URL has the highest precedence. This can be visualized graphically as: .----------------------------------------------------------. | .----------------------------------------------------. | | | .----------------------------------------------. | | | | | .----------------------------------------. | | | | | | | .----------------------------------. | | | | | | | | | <relative_reference> | | | | | | | | | `----------------------------------' | | | | | | | | (5.1.1) Base URL embedded in the | | | | | | | | document's content | | | | | | | `----------------------------------------' | | | | | | (5.1.2) Base URL of the encapsulating entity | | | | | | (message, document, or none). | | | | | `----------------------------------------------' | | | | (5.1.3) URL used to retrieve the entity | | | `----------------------------------------------------' | | (5.1.4) Base URL = "" (undefined) | `----------------------------------------------------------' 5.1.1. Base URL within Document Content Within certain document media types, the base URL 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 URL 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 URL can be embedded in the Hypertext Markup Language (HTML) [3] is provided in Appendix D. Messages are considered to be composite documents. The base URL of a message can be specified within the message headers (or equivalent tagged metainformation) of the message. For protocols that make use of message headers like those described in MIME [4], the base URL can be specified by the Content-Base or Content-Location header fields. Content-Base = "Content-Base" ":" absoluteURL Content-Location = "Content-Location" ":" ( absoluteURL | relativeURL ) The field names are case-insensitive and any whitespace inside the field value (including that used for line folding) is ignored. Content-Base takes precedence over any Content-Location. If the latter is relative, it must be resolved to its absolute form (like any relative URL) before it can be used as the base URL for other references. For example, the header field Content-Base: http://www.ics.uci.edu/Test/a/b/c would indicate that the base URL for that message is the string "http://www.ics.uci.edu/Test/a/b/c". The base URL for a message serves as both the base for any relative URLs within the message headers and the default base URL for documents enclosed within the message, as described in the next section. Protocols which do not use the RFC 822 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 URL as part of a message. 5.1.2. Base URL from the Encapsulating Entity If no base URL is embedded, the base URL 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 URL of the document is the base URL of the entity in which the document is encapsulated. Composite media types, such as the "multipart/*" and "message/*" media types defined by MIME (RFC 1521, [4]), define a hierarchy of retrieval context for their enclosed documents. In other words, the retrieval context of a component part is the base URL of the composite entity of which it is a part. Thus, a composite entity can redefine the retrieval context of its component parts via the inclusion of a Content-Base or Content-Location header, and this redefinition applies recursively for a hierarchy of composite parts. Note that this might not change the base URL of the components, since each component may include an embedded base URL or base-header that takes precedence over the retrieval context. 5.1.3. Base URL from the Retrieval URL If no base URL is embedded and the document is not encapsulated within some other entity (e.g., the top level of a composite entity), then, if a URL was used to retrieve the base document, that URL shall be considered the base URL. Note that if the retrieval was the result of a redirected request, the last URL used (i.e., that which resulted in the actual retrieval of the document) is the base URL. 5.1.4. Default Base URL If none of the conditions described in Sections 5.1.1--5.1.3 apply, then the base URL is considered to be the empty string and all URL references within that document are assumed to be absolute URLs. It is the responsibility of the distributor(s) of a document containing relative URLs to ensure that the base URL for that document can be established. It must be emphasized that relative URLs cannot be used reliably in situations where the document's base URL is not well-defined. 5.2. Resolving Relative References to Absolute Form This section describes an example algorithm for resolving URL references which might be relative to a given base URL. The base URL is established according to the rules of Section 5.1 and parsed into the four main components as described in Section 4.4. Note that only the scheme component is required to be present in the base URL; the other components may be empty or undefined. A component is undefined if its preceding separator does not appear in the URL reference; the path component is never undefined, though it may be empty. The base URL's query component is not used by the resolution algorithm and may be discarded. For each URL reference, the following steps are performed in order: 1) The URL reference is parsed into the potential four components and fragment identifier, as described in Section 4.4. 2) If the path component is empty and the scheme, server, and query components are undefined, then it is a reference to the current document and we are done. 3) If the scheme component is defined, indicating that the reference starts with a scheme name, then the reference is interpreted as an absolute URL and we are done. Otherwise, the reference URL's scheme is inherited from the base URL's scheme component. 4) If the server component is defined, then the reference is a network-path and we skip to step 7. Otherwise, the reference URL's server is inherited from the base URL's server component, which will also be undefined if the URL scheme does not use a server 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 URL'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 URL'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) If the reference's query component is defined, then a "?" character is appended to the buffer string, followed by the query component. d) All occurrences of "./", where "." is a complete path segment, are removed from the buffer string. e) If the buffer string ends with "." as a complete path segment, that "." is removed. f) 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. g) If the buffer string ends with "<segment>/..", where <segment> is a complete path segment not equal to "..", that "<segment>/.." is removed. h) If the buffer string contains a question-mark "?" character, then the reference URL's query component is the substring after the first (left-most) question-mark. Otherwise, the reference URL's query component is set undefined. i) The reference URL's new path component is the buffer string up to, but not including, the first question-mark character or the end of the buffer string. 7) The resulting URL components, including any inherited from the base URL, are recombined to give the absolute form of the URL reference. Using pseudocode, this would be result = "" if scheme is defined then append scheme to result append ":" to result if server is defined then append "//" to result append server 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. Resolution examples are provided in Appendix C. 6. Adding New Schemes The Internet Assigned Numbers Authority (IANA) maintains a registry of URL schemes. The current process for defining URL schemes is via the Internet standards process: new URL schemes should be described in standards-track RFCs. Over time, other methods of registering URL schemes may be added. URL schemes must have demonstrable utility and operability. One way to provide such a demonstration is via a gateway which provides objects in the new scheme for clients using an existing protocol. If the new scheme does not locate resources that are data objects, the properties of names in the new space must be clearly defined. URL schemes should follow the same syntactic conventions of existing schemes when appropriate. URL schemes should use the generic-URL syntax if they are intended to be used with relative URLs. A description of the allowed relative forms should be included in the scheme's definition. URL schemes cannot redefine the algorithm for resolving relative references. The resolution algorithm must remain independent of the scheme name in order to preserve the mobility of relative references between naming schemes and the ability to parse and resolve a relative reference without knowing the properties of any particular scheme. 7. Security Considerations A URL 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 server apportions its namespace. Such a guarantee can only be obtained from the person(s) controlling that namespace and the resource in question. 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 server which is in fact running a different protocol. The content of the URL contains instructions which, when interpreted according to this other protocol, cause an unexpected operation. An example has been the use of gopher URLs to cause a rude message to be sent via a SMTP server. Caution should be used when using any URL which 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 URLs contain 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. 8. Acknowledgements This document was derived from RFC 1738 [2] and RFC 1808 [7]; the acknowledgements in those specifications still apply. In addition, this draft has benefited from comments by Lauren Wood. 9. References [1] 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, CERN, June 1994. [2] Berners-Lee, T., Masinter, L., and M. McCahill, Editors, "Uniform Resource Locators (URL)", RFC 1738, CERN, Xerox Corporation, University of Minnesota, December 1994. [3] Berners-Lee T., and D. Connolly, "HyperText Markup Language Specification -- 2.0", RFC 1866, MIT/W3C, November 1995. [4] Borenstein, N., and N. Freed, "MIME (Multipurpose Internet Mail Extensions): Mechanisms for Specifying and Describing the Format of Internet Message Bodies", RFC 1521, Bellcore, Innosoft, September 1993. [5] Braden, R., Editor, "Requirements for Internet Hosts -- Application and Support", STD 3, RFC 1123, IETF, October 1989. [6] Crocker, D., "Standard for the Format of ARPA Internet Text Messages", STD 11, RFC 822, UDEL, August 1982. [7] Fielding, R., "Relative Uniform Resource Locators", RFC 1808, UC Irvine, June 1995. [8] Kunze, J., "Functional Recommendations for Internet Resource Locators", RFC 1736, IS&T, UC Berkeley, February 1995. [9] Mockapetris, P., "Domain Names - Concepts and Facilities", STD 13, RFC 1034, USC/Information Sciences Institute, November 1987. [10] Sollins, K., and L. Masinter, "Functional Requirements for Uniform Resource Names", RFC 1737, MIT/LCS, Xerox Corporation, December 1994. [11] US-ASCII. "Coded Character Set -- 7-bit American Standard Code for Information Interchange", ANSI X3.4-1986. 10. Authors' Addresses Tim Berners-Lee World Wide Web Consortium MIT Laboratory for Computer Science, NE43-356 545 Technology Square Cambridge, MA 02139 Fax: +1(617)258-8682 EMail: timbl@w3.org Roy T. Fielding Department of Information and Computer Science University of California, Irvine Irvine, CA 92697-3425 Fax: +1(714)824-4056 EMail: fielding@ics.uci.edu Larry Masinter Xerox PARC 3333 Coyote Hill Road Palo Alto, CA 94034 Fax: +1(415)812-4333 EMail: masinter@parc.xerox.com Appendices A. Collected BNF for URLs URL-reference = [ absoluteURL | relativeURL ] [ "#" fragment ] absoluteURL = generic-URL | opaque-URL opaque-URL = scheme ":" *urlchar generic-URL = scheme ":" relativeURL relativeURL = net_path | abs_path | rel_path net_path = "//" server [ abs_path ] abs_path = "/" rel_path rel_path = [ path_segments ] [ "?" query ] scheme = 1*( alpha | digit | "+" | "-" | "." ) server = [ [ user [ ":" password ] "@" ] hostport ] user = *( unreserved | escaped | ";" | "&" | "=" | "+" ) password = *( unreserved | escaped | ";" | "&" | "=" | "+" ) hostport = host [ ":" port ] host = hostname | hostnumber hostname = *( domainlabel "." ) toplabel domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum toplabel = alpha | alpha *( alphanum | "-" ) alphanum hostnumber = 1*digit "." 1*digit "." 1*digit "." 1*digit port = *digit path = [ "/" ] path_segments path_segments = segment *( "/" segment ) segment = *pchar *( ";" param ) param = *pchar pchar = unreserved | escaped | ":" | "@" | "&" | "=" | "+" query = *urlchar fragment = *urlchar urlchar = reserved | unreserved | escaped reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" unreserved = alpha | digit | mark mark = "$" | "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")" | "," escaped = "%" hex hex hex = digit | "A" | "B" | "C" | "D" | "E" | "F" | "a" | "b" | "c" | "d" | "e" | "f" alphanum = alpha | digit alpha = lowalpha | hialpha 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" hialpha = "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" B. Parsing a URL Reference with a Regular Expression As described in Section 4.4, the generic-URL syntax is not sufficient to disambiguate the components of some forms of URL. 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 URL reference. The following line is the regular expression for breaking-down a URL 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ドル = http: 2ドル = http 3ドル = //www.ics.uci.edu 4ドル = www.ics.uci.edu 5ドル = /pub/ietf/uri/ 6ドル = <undefined> 7ドル = <undefined> 8ドル = #Related 9ドル = 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ドル server = 4ドル path = 5ドル query = 7ドル fragment = 9ドル and, going in the opposite direction, we can recreate a URL reference from its components using the algorithm in step 7 of Section 5.2. C. Examples of Resolving Relative URL References Within an object with a well-defined base URL of Content-Base: http://a/b/c/d;p?q the relative URLs 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 URL 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 URL's path. Note that the ".." syntax cannot be used to change the server component of a URL. ../../../g = http://a/../g ../../../../g = http://a/../../g 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 URL 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 g?y/./x = http://a/b/c/g?y/x g?y/../x = http://a/b/c/x g#s/./x = http://a/b/c/g#s/./x g#s/../x = http://a/b/c/g#s/../x Finally, some older parsers allow the scheme name to be present in a relative URL if it is the same as the base URL scheme. This is considered to be a loophole in prior specifications of partial URLs [1] and should be avoided by future parsers. http:g = http:g http: = http: D. Embedding the Base URL in HTML documents It is useful to consider an example of how the base URL 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) [3] can include an embedded base URL. This appendix does not form a part of the relative URL 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 URL for resolving any relative URLs. The "HREF" attribute must be an absolute URL. 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 URL "../x" as representing the absolute URL <http://www.ics.uci.edu/Test/a/x> regardless of the context in which the example document was obtained. E. Recommendations for Delimiting URLs in Context URLs are often transmitted through formats which do not provide a clear context for their interpretation. For example, there are many occasions when URLs 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 URL from the rest of the text, and in particular from punctuation marks that might be mistaken for part of the URL. In practice, URLs 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/ The prefix "URL:", with or without a trailing space, is sometimes used to help distinguish a URL from normal text. These wrappers do not form part of the URL. In the case where a fragment identifier is associated with a URL reference, the fragment would be placed within the brackets as well (separated from the URL with a "#" character). In some cases, extra whitespace (spaces, linebreaks, tabs, etc.) may need to be added to break long URLs across lines. The whitespace should be ignored when extracting the URL. 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 URL 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 URL. Examples: Yes, Jim, I found it under "http://www.w3.org/pub/WWW/", but you can probably pick it up from <ftp://ds.internic. net/rfc/>. Note the warning in <http://ds.internic.net/ instructions/overview.html#WARNING>. F. Summary of Non-editorial Changes F.1. Additions Section 1 (Introduction) is entirely new. Design rationale for the scope of URLs and the chosen URL character set has been added in order to address common misconceptions about what would and would not be appropriate for additional URL schemes, and why the allowed character set is limited to US-ASCII characters. A definition of URI is also given, and how the URI syntax equates to the URL syntax, so that other IETF specifications (e.g., HTTP, HTML, etc.) can refer to a single definition of URI. Section 3 (URL References) was added to stem the confusion regarding "what is a URL" and how to describe fragment identifiers given that they are not part of the URL, but are part of the URL syntax and parsing concerns. In addition, it provides a reference definition for use by other IETF specifications (HTML, HTTP, etc.) which have previously attempted to redefine the URL syntax in order to account for the presence of fragment identifiers in URL references. Section 2.3.2 (When to Escape and Unescape) was added in response to many (mis)implementation questions on the subject. F.2. Modifications from both RFC 1738 and RFC 1808 Confusion regarding the terms "character encoding", the URL "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 URL syntax. Both RFC 1738 and RFC 1808 refer to the "reserved" set of characters as if URL-interpreting servers 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 URL scheme. However, this has not been true in practice; any character which is interpreted differently when it is escaped is, in effect, reserved. Furthermore, the interpreting engine on a server is often dependent on the resource, not just the URL scheme. The description of reserved characters has been changed accordingly. The plus "+" character was added to those in the "reserved" set, since it is treated as reserved within some URL components. 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 question-mark "?" character was removed from the set of allowed characters for the user and password in the server component, since testing showed that many applications treat it as reserved for separating the query component from the rest of the URL. RFC 1738 specified that the path was separated from the server portion of a URL 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 URLs <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. F.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. The recommendations for delimiting URLs in context (Appendix E) have been adjusted to reflect current practice. F.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 3. The description of the mythical Base header field has been replaced with the Content-Base and Content-Location header fields defined by HTTP/1.1 and MHTML. RFC 1808 described various schemes as either having or not having the properties of the generic-URL syntax. However, the only requirement is that the particular document containing the relative references have a base URL which abides by the generic-URL syntax, regardless of the URL scheme, so the associated description has been updated to reflect that. The BNF term <net_loc> has been replaced with <server>, since the latter more accurately describes its use and purpose. 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 URL reference. The resolution examples in Appendix C have been modified to reflect this change. Testing has also revealed that most client applications remove the query component from the base URL before resolving relative URLs, and append the reference's query component to a relative path before merging it with the base path. The resolution algorithm has been changed accordingly.