Network Working Group T. Berners-Lee INTERNET-DRAFT MIT/LCS R. Fielding Expires six months after publication date. U.C. Irvine L. Masinter Xerox Corporation 02 November 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). Abstract A Uniform Resource Locator (URL) is a compact string representation of the location for a resource that is available via the Internet. 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. This document updates and merges RFC 1738 "Uniform Resource Locators" and RFC 1808 "Relative Uniform Resource Locators" 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, since those portions will be updated as separate documents. 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 [12]. 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 are an electronic document, an image, a service (e.g., "today's weather in Los Angeles"), or a collection of other resources. Not all resources are network "retrievable", such as human beings, corporations, and physical objects like the books in a library. The resource is the conceptual entity defined by the identity, not necessarily the thing which holds that identity at one particular instance in time. Thus, a resource can remain constant even when its content---the thing(s) to which it currently corresponds---changes over time, provided that the identity 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', `find attributes'. This specification is only concerned with the issue of identifying a resource by its location. 1.1. Example URLs 1.2. 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 /;? 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 [5], 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 * to designate n or more repetitions of the following element; n defaults to 0. 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" digits = 1*digit alphanum = alpha | digit The complete URL syntax is collected in Appendix A. 2. 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. A URL reference which does not contain a URL is a reference to the current document. In other words, an empty reference embedded within a document is interpreted as a reference to the top of that document, and a reference containing only a fragment identifier is a reference to the identified fragment of that document. 3. Generic URL Syntax 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 identity of the characters used. 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 it 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. 3.1. 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: : An absolute URL contains the name of the scheme being used () followed by a colon (":") and then a string (the ) 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"). 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 below. 3.2. URL Character Encoding In most URL schemes, the sequences of characters in different parts of a URL are used to represent sequences of octets used in Internet protocols. For example, in the ftp scheme, the host name, directory name and file names are such sequences of octets, represented by parts of the URL. Within those parts, an octet may be represented by the character which has that octet as its code within the US-ASCII [20] coded character set. In addition, octets may be encoded by a character triplet consisting of the character "%" followed by the two hexadecimal digits (from "0123456789ABCDEF") which form the hexadecimal value of the octet. (The characters "abcdef" may also be used in hexadecimal encodings.) Octets must be encoded if they have no corresponding graphic character within the US-ASCII coded character set, if the use of the corresponding character is unsafe, or if the corresponding character is reserved for some other interpretation within the particular URL scheme. 3.2.1. No corresponding graphic US-ASCII URLs are written only with the graphic printable characters of the US-ASCII coded character set. Any character which cannot be represented by a printable US-ASCII coded character, including non-7bit (octets 80-FF hexadecimal) and control (octets 00-1F and 7F hexadecimal) characters, must be encoded using the hexadecimal encoding in order to be used within a URL. 3.2.2. Unsafe Characters Characters can be unsafe for a number of reasons. The space character is unsafe 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. The angle-bracket "<" and ">" and double-quote (`"') characters are unsafe because they are often used as the delimiters around URLs in text documents and protocol fields. The character "#" is unsafe because it is used to delimit a URL from a fragment identifier in URL references. The character "%" is unsafe because it is used for encodings of other characters. # [NOTE: This following bit was in RFC 1738, but seems wrong. # Perhaps we need a section for Unwise Characters. ...Roy] # Other characters are unsafe because gateways and other transport # agents are known to sometimes modify such characters. These # characters are "{", "}", "|", "\", "^", "~", "[", "]", and "`". Unsafe characters must be encoded within a URL. For example, the character "#" must be encoded within URLs even in systems that do not normally deal with fragment identifiers, so that if the URL is copied into another system that does use them, it will not be necessary to change the URL encoding. 3.2.3. Reserved Characters Many URL schemes reserve certain characters for a special meaning: their appearance in the scheme-specific part of the URL has a designated semantics. If the character corresponding to an octet is reserved in a scheme, the octet must be encoded if it is not intended to represent that reserved purpose. The characters ";", "/", "?", ":", "@", "=", "+", and "&" are the characters which may be reserved for special meaning within a scheme. No other characters may be reserved within a scheme. # [Note: I am afraid that this notion of reserving a single set of # characters for all reserved uses just doesn't exist in practice. # It might be better to just define the concept, and let each scheme # define the set of reserved characters for each component. ...Roy] Usually a URL has the same interpretation when an octet is represented by a character and when it is encoded. However, this is not true for reserved characters: encoding a character reserved for a particular scheme may change the semantics of a URL. 3.3. 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-RL syntax is used in interpreting relative URLs. 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 embedded URLs. Experience within the World Wide Web has demonstrated that the ability to perform relative referencing is necessary for the long-term usability of embedded URLs. It is not necessary for all URLs within a given scheme to be restricted to the generic-RL 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-RL 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 documents that have unsuitable base URLs. 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 filename. 3.4. 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 five main components to define the location of a resource :///;? each of which, except , may be absent from a particular URL. The order of the components is important. If both and are present, the information must occur after the . 3.4.1. 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: :@: Some or all of the parts ":@", ":", and ":" 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. The different components obey the following rules: user An optional user name. Some schemes (e.g., ftp) allow the specification of a user name. password An optional password. If present, it follows the user name, separated from it by a colon. The user name (and password), if present, are followed by a commercial at-sign "@". Within the user and password field, any ":", "@", or "/" must be encoded. 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., has an empty user name and no password, has no user name, while has a user name of "foo" and an empty password. host The fully qualified domain name of a network host, or its IP address as a set of four decimal digit groups separated by ".". Fully qualified domain names take the form as described in Section 3.5 of RFC 1034 [13] 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. port The port number to connect to. 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 colon is as well. 3.4.2. Path Component path The rest of the locator consists of data specific to the scheme, and is known as the "path". It supplies the details of how the specified resource can be accessed. Note that the "/" between the host (or port) and the url-path is NOT part of the url-path. The path syntax depends on the scheme being used, as does the manner in which it is interpreted. 3.4.3. Parameters Component 3.4.4. Query Component 4. BNF for URLs This is a BNF-like description of the Uniform Resource Locator syntax, using the conventions of RFC 822 [5], except that "|" is used to designate alternatives. Briefly, literals are quoted with "", parentheses "(" and ")" are used to group elements, optional elements are enclosed in [brackets], and elements may be preceded with * to designate n or more repetitions of the following element; n defaults to 0. ; The generic form of a URL is: genericurl = scheme ":" schemepart ; the scheme is in lower case; interpreters should use case-ignore scheme = 1*[ lowalpha | digit | "+" | "-" | "." ] schemepart = *xchar | ip-schemepart ; URL schemeparts for ip based protocols: ip-schemepart = "//" login [ "/" urlpath ] login = [ user [ ":" password ] "@" ] hostport hostport = host [ ":" port ] host = hostname | hostnumber hostname = *[ domainlabel "." ] toplabel domainlabel = alphadigit | alphadigit *[ alphadigit | "-" ] alphadigit toplabel = alpha | alpha *[ alphadigit | "-" ] alphadigit alphadigit = alpha | digit hostnumber = digits "." digits "." digits "." digits port = digits user = *[ uchar | ";" | "?" | "&" | "=" ] password = *[ uchar | ";" | "?" | "&" | "=" ] urlpath = *xchar ; depends on protocol see section 3.1 This BNF also describes the generic-RL syntax for valid base URLs. Note that this differs from the URL syntax defined in RFC 1738 [2] in that all schemes are required to use a single set of reserved characters and use them consistently within the major URL components. URL = ( absoluteURL | relativeURL ) [ "#" fragment ] absoluteURL = generic-RL | ( scheme ":" *( uchar | reserved ) ) generic-RL = scheme ":" relativeURL relativeURL = net_path | abs_path | rel_path net_path = "//" server [ abs_path ] abs_path = "/" rel_path rel_path = [ path ] [ ";" params ] [ "?" query ] path = fsegment *( "/" segment ) fsegment = 1*pchar segment = *pchar params = param *( ";" param ) param = *( pchar | "/" ) scheme = 1*( alpha | digit | "+" | "-" | "." ) server = *( pchar | ";" | "?" ) query = *( uchar | reserved ) fragment = *( uchar | reserved ) pchar = uchar | ":" | "@" | "&" | "=" | "+" uchar = unreserved | escape xchar = unreserved | reserved | escape unreserved = alpha | digit | safe | extra escape = "%" hex hex hex = digit | "A" | "B" | "C" | "D" | "E" | "F" | "a" | "b" | "c" | "d" | "e" | "f" safe = "$" | "-" | "_" | "." extra = "!" | "*" | "'" | "(" | ")" | "," national = "{" | "}" | "|" | "\" | "^" | "~" | "[" | "]" | "`" reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" punctuation = "<" | ">" | "#" | "%" | <"> 5. Specific Schemes The mapping for some existing standard and experimental protocols is outlined in the BNF syntax definition. Notes on particular protocols follow. The schemes covered are: ftp File Transfer protocol http Hypertext Transfer Protocol gopher The Gopher protocol mailto Electronic mail address news USENET news nntp USENET news using NNTP access telnet Reference to interactive sessions wais Wide Area Information Servers file Host-specific file names prospero Prospero Directory Service Other schemes may be specified by future specifications. Section 4 of this document describes how new schemes may be registered, and lists some scheme names that are under development. 5.3. Specific Schemes and their Syntactic Categories Each URL scheme has its own rules regarding the presence or absence of the syntactic components described in Sections 2.1 and 2.2. In addition, some schemes are never appropriate for use with relative URLs. However, since relative URLs will only be used within contexts in which they are useful, these scheme-specific differences can be ignored by the resolution process. Within this section, we include as examples only those schemes that have a defined URL syntax in RFC 1738 [2]. The following schemes are never used with relative URLs: mailto Electronic Mail news USENET news telnet TELNET Protocol for Interactive Sessions Some URL schemes allow the use of reserved characters for purposes outside the generic-RL syntax given above. However, such use is rare. Relative URLs can be used with these schemes whenever the applicable base URL follows the generic-RL syntax. gopher Gopher and Gopher+ Protocols prospero Prospero Directory Service wais Wide Area Information Servers Protocol Users of gopher URLs should note that gopher-type information is almost always included at the beginning of what would be the generic-RL path. If present, this type information prevents relative-path references to documents with differing gopher-types. Finally, the following schemes can always be parsed using the generic-RL syntax. This does not necessarily imply that relative URLs will be useful with these schemes -- that decision is left to the system implementation and the author of the base document. file Host-specific Files ftp File Transfer Protocol http Hypertext Transfer Protocol nntp USENET news using NNTP access 6. REGISTRATION OF NEW SCHEMES A new scheme may be introduced by defining a mapping onto a conforming URL syntax, using a new prefix. URLs for experimental schemes may be used by mutual agreement between parties. Scheme names starting with the characters "x-" are reserved for experimental purposes. The Internet Assigned Numbers Authority (IANA) will maintain a registry of URL schemes. Any submission of a new URL scheme must include a definition of an algorithm for accessing of resources within that scheme and the syntax for representing such a scheme. 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. New schemes should try to follow the same syntactic conventions of existing schemes, where appropriate. It is likewise recommended that, where a protocol allows for retrieval by URL, that the client software have provision for being configured to use specific gateway locators for indirect access through new naming schemes. The following scheme have been proposed at various times, but this document does not define their syntax or use at this time. It is suggested that IANA reserve their scheme names for future definition: afs Andrew File System global file names. mid Message identifiers for electronic mail. cid Content identifiers for MIME body parts. nfs Network File System (NFS) file names. tn3270 Interactive 3270 emulation sessions. mailserver Access to data available from mail servers. z39.50 Access to ANSI Z39.50 services. We recommend that new schemes be designed to be parsable via the generic-RL syntax if they are intended to be used with relative URLs. A description of the allowed relative forms should be included when a new scheme is registered, as per Section 4 of RFC 1738 [2]. 7. Parsing a URL Reference An accepted method for parsing URL references is useful to clarify the generic-RL syntax of Section 2.2 and to describe the algorithm for resolving relative URLs presented in Section 4. This section describes the parsing rules for breaking down a URL (relative or absolute) into the component parts described in Section 2.1. The rules assume that the URL has already been separated from any surrounding text and copied to a "parse string". The rules are listed in the order in which they would be applied by the parser. 7.4.1. Parsing the Fragment Identifier If the parse string contains a crosshatch "#" character, then the substring after the first (left-most) crosshatch "#" and up to the end of the parse string is the identifier. If the crosshatch is the last character, or no crosshatch is present, then the fragment identifier is empty. The matched substring, including the crosshatch character, is removed from the parse string before continuing. Note that the fragment identifier is not considered part of the URL. However, since it is often attached to the URL, parsers must be able to recognize and set aside fragment identifiers as part of the process. 7.4.2. Parsing the Scheme If the parse string contains a colon ":" after the first character and before any characters not allowed as part of a scheme name (i.e., any not an alphanumeric, plus "+", period ".", or hyphen "-"), the of the URL is the substring of characters up to but not including the first colon. These characters and the colon are then removed from the parse string before continuing. 7.4.3. Parsing the Network Location/Login If the parse string begins with a double-slash "//", then the substring of characters after the double-slash and up to, but not including, the next slash "/" character is the network location/login () of the URL. If no trailing slash "/" is present, the entire remaining parse string is assigned to . The double- slash and are removed from the parse string before continuing. 7.4.4. Parsing the Query Information If the parse string contains a question mark "?" character, then the substring after the first (left-most) question mark "?" and up to the end of the parse string is the information. If the question mark is the last character, or no question mark is present, then the query information is empty. The matched substring, including the question mark character, is removed from the parse string before continuing. 7.4.5. Parsing the Parameters If the parse string contains a semicolon ";" character, then the substring after the first (left-most) semicolon ";" and up to the end of the parse string is the parameters (). If the semicolon is the last character, or no semicolon is present, then is empty. The matched substring, including the semicolon character, is removed from the parse string before continuing. 7.4.6. Parsing the Path After the above steps, all that is left of the parse string is the URL and the slash "/" that may precede it. Even though the initial slash is not part of the URL path, the parser must remember whether or not it was present so that later processes can differentiate between relative and absolute paths. Often this is done by simply storing the preceding slash along with the path. 8. 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 embedded relative URLs; 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: .----------------------------------------------------------. | .----------------------------------------------------. | | | .----------------------------------------------. | | | | | .----------------------------------------. | | | | | | | (8.1) Base URL embedded in the | | | | | | | | document's content | | | | | | | `----------------------------------------' | | | | | | (8.2) Base URL of the encapsulating entity | | | | | | (message, document, or none). | | | | | `----------------------------------------------' | | | | (8.3) URL used to retrieve the entity | | | `----------------------------------------------------' | | (8.4) Base URL = "" (undefined) | `----------------------------------------------------------' 8.1. 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 an Appendix (Section 10). 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 RFC 822 [5], we recommend that the format of this header be: base-header = "Content-Base" ":" absoluteURL where "Content-Base" is case-insensitive and any whitespace inside the field-value (including that used for line folding) is ignored. 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. 8.2. 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 base-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. 8.3. 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. 8.4. Default Base URL If none of the conditions described in Sections 3.1 -- 3.3 apply, then the base URL is considered to be the empty string and all embedded URLs 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. 9. Resolving Relative URLs This section describes an example algorithm for resolving URLs within a context in which the URLs may be relative, such that the result is always a URL in absolute form. Although this algorithm cannot guarantee that the resulting URL will equal that intended by the original author, it does guarantee that any valid URL (relative or absolute) can be consistently transformed to an absolute form given a valid base URL. The following steps are performed in order: Step 1: The base URL is established according to the rules of Section 3. If the base URL is the empty string (unknown), the embedded URL is interpreted as an absolute URL and we are done. Step 2: Both the base and embedded URLs are parsed into their component parts as described in Section 2.4. a) If the embedded URL is entirely empty, it inherits the entire base URL (i.e., is set equal to the base URL) and we are done. b) If the embedded URL starts with a scheme name, it is interpreted as an absolute URL and we are done. c) Otherwise, the embedded URL inherits the scheme of the base URL. Step 3: If the embedded URL's is non-empty, we skip to Step 7. Otherwise, the embedded URL inherits the (if any) of the base URL. Step 4: If the embedded URL path is preceded by a slash "/", the path is not relative and we skip to Step 7. Step 5: If the embedded URL path is empty (and not preceded by a slash), then the embedded URL inherits the base URL path, and a) if the embedded URL's is non-empty, we skip to step 7; otherwise, it inherits the of the base URL (if any) and b) if the embedded URL's is non-empty, we skip to step 7; otherwise, it inherits the of the base URL (if any) and we skip to step 7. Step 6: The last segment of the base URL's path (anything following the rightmost slash "/", or the entire path if no slash is present) is removed and the embedded URL's path is appended in its place. The following operations are then applied, in order, to the new path: a) All occurrences of "./", where "." is a complete path segment, are removed. b) If the path ends with "." as a complete path segment, that "." is removed. c) All occurrences of "/../", where is a complete path segment not equal to "..", are removed. Removal of these path segments is performed iteratively, removing the leftmost matching pattern on each iteration, until no matching pattern remains. d) If the path ends with "/..", where is a complete path segment not equal to "..", that "/.." is removed. Step 7: The resulting URL components, including any inherited from the base URL, are recombined to give the absolute form of the embedded URL. Parameters, regardless of their purpose, do not form a part of the URL path and thus do not affect the resolving of relative paths. In particular, the presence or absence of the ";type=d" parameter on an ftp URL does not affect the interpretation of paths relative to that URL. Fragment identifiers are only inherited from the base URL when the entire embedded URL is empty. 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 matches. 10. Security Considerations The URL scheme does not in itself pose a security threat. Users should beware that there is no general guarantee that a URL which at one time points to a given object continues to do so, and does not even at some later time point to a different object due to the movement of objects on servers. A URL-related security threat is that it is sometimes possible to construct a URL such that an attempt to perform a harmless idempotent operation such as the retrieval of the object 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 embedded encoded delimiters for a given protocol (for example, CR and LF characters for telnet protocols) that these are not unencoded before transmission. This would violate the protocol but could be used to simulate an extra operation or parameter, again causing an unexpected and possible harmful remote operation to be performed. The use of URLs containing passwords that should be secret is clearly unwise. There are no security considerations in the use or parsing of relative URLs. However, once a relative URL has been resolved to its absolute form, the same security considerations apply as those described in RFC 1738 [2]. 11. Acknowledgements This paper builds on the basic WWW design (RFC 1630) and much discussion of these issues by many people on the network. The discussion was particularly stimulated by articles by Clifford Lynch, Brewster Kahle [10] and Wengyik Yeong [18]. Contributions from John Curran, Clifford Neuman, Ed Vielmetti and later the IETF URL BOF and URI working group were incorporated. Most recently, careful readings and comments by Dan Connolly, Ned Freed, Roy Fielding, Guido van Rossum, Michael Dolan, Bert Bos, John Kunze, Olle Jarnefors, Peter Svanberg and many others have helped refine this RFC. This work is derived from concepts introduced by Tim Berners-Lee and the World-Wide Web global information initiative. Relative URLs are described as "Partial URLs" in RFC 1630 [1]. This document is intended to fulfill the recommendations for Internet Resource Locators as stated in [6]. It has benefited greatly from the comments of all those participating in the URI-WG. Particular thanks go to Larry Masinter, Michael A. Dolan, Guido van Rossum, Dave Kristol, David Robinson, and Brad Barber for identifying problems/deficiencies in earlier drafts. 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] Kunze, J., "Functional Recommendations for Internet Resource Locators", RFC 1736, IS&T, UC Berkeley, February 1995. [8] Huitema, C., "Naming: Strategies and Techniques", Computer Networks and ISDN Systems 23 (1991) 107-110. [9] Kahle, B., "Document Identifiers, or International Standard Book Numbers for the Electronic Age", 1991. [10] Mockapetris, P., "Domain Names - Concepts and Facilities", STD 13, RFC 1034, USC/Information Sciences Institute, November 1987. [11] Sollins, K. and L. Masinter, "Functional Requirements for Uniform Resource Names", RFC 1737, MIT/LCS, Xerox Corporation, December 1994. [12] Yeong, W. "Towards Networked Information Retrieval", Technical report 91-06-25-01, Performance Systems International, Inc., June 1991. [13] Yeong, W., "Representing Public Archives in the Directory", Work in Progress, November 1991. [14] "Coded Character Set -- 7-bit American Standard Code for Information Interchange", ANSI X3.4-1986. Editors' 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, CA 92717-3425 U.S.A. Fax: +1(714)824-4056 EMail: fielding@ics.uci.edu Larry Masinter Xerox PARC 3333 Coyote Hill Road Palo Alto, CA 94034 Phone: (415) 812-4365 Fax: (415) 812-4333 EMail: masinter@parc.xerox.com Appendices 12. Recommendations for URLs in Context URIs, including URLs, are intended to be transmitted through protocols which provide a context for their interpretation. In some cases, it will be necessary to distinguish URLs from other possible data structures in a syntactic structure. In this case, is recommended that URLs be preceeded with a prefix consisting of the characters "URL:". For example, this prefix may be used to distinguish URLs from other kinds of URIs. In addition, there are many occasions when URLs are included in other kinds of text; examples include electronic mail, USENET news messages, or printed on paper. In such cases, it is convenient to have a separate syntactic wrapper that delimits the URL and separates it from the rest of the text, and in particular from punctuation marks that might be mistaken for part of the URL. For this purpose, is recommended that angle brackets ("<" and ">"), along with the prefix "URL:", be used to delimit the boundaries of the URL. This wrapper does not form part of the URL and should not be used in contexts in which delimiters are already specified. In the case where a fragment/anchor identifier is associated with a URL (following a "#"), the identifier would be placed within the brackets as well. 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 unencoded 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 but you can probably pick it up from . Note the warning in . 13. Examples and Recommended Practice 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: 13.1. Normal Examples g:h = g = ./g = g/ = /g = //g = ?y = g?y = g?y/./x = #s = g#s = g#s/./x = g?y#s = ;x = g;x = g;x?y#s = . = ./ = .. = ../ = ../g = ../.. = ../../ = ../../g = 13.2. 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 resolves to the complete base URL: <> = 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 of a URL. ../../../g = ../../../../g = Similarly, parsers must avoid treating "." and ".." as special when they are not complete components of a relative path. /./g = /../g = g. = .g = g.. = ..g = Less likely are cases where the relative URL uses unnecessary or nonsensical forms of the "." and ".." complete path segments. ./../g = ./g/. = g/./h = g/../h = 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: = 13.3. Recommended Practice Authors should be aware that path names which contain a colon ":" character cannot be used as the first component of a relative URL path (e.g., "this:that") because they will likely be mistaken for a scheme name. It is therefore necessary to precede such cases with other components (e.g., "./this:that"), or to escape the colon character (e.g., "this%3Athat"), in order for them to be correctly parsed. The former solution is preferred because it does not affect the absolute form of the URL. There is an ambiguity in the semantics for the ftp URL scheme regarding the use of a trailing slash ("/") character and/or a parameter ";type=d" to indicate a resource that is an ftp directory. If the result of retrieving that directory includes embedded relative URLs, it is necessary that the base URL path for that result include a trailing slash. For this reason, we recommend that the ";type=d" parameter value not be used within contexts that allow relative URLs. 14. 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: An example HTML document ... a hypertext anchor ... A parser reading the example document should interpret the given relative URL "../x" as representing the absolute URL regardless of the context in which the example document was obtained. 15. Summary of Non-editorial Changes 15.1. Additions Section 2 on 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. Design rationale for the chosen URL character set has been added to Section 3 in order to address common misconceptions about the lack of non-US-ASCII characters in the allowed character set. 15.2. Modifications from both RFC 1738 and RFC 1808 The "+" character was added to those in the reserved set, since it is treated as reserved within some URLs. 15.3. Modifications from RFC 1738 The definition of specific URL schemes and their scheme-specific syntax and semantics has been moved to separate documents. 15.4. 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 2. The description of the Base header field has been replaced with the Content-Base and Content-Location header fields defined by HTTP/1.1 [4]. RFC 1808 described various schemes as either having or not having the properties of the generic-RL syntax. However, the only requirement is that the particular document containing the relative references have a base URL which abides by the generic-RL syntax, regardless of the URL scheme, so the associated description has been updated to reflect that. The BNF term has been replaced with , since the latter more accurately describes its use and purpose.