Hypertext Markup Language (HTML) Tim Berners-Lee, CERN Internet Draft Daniel Connolly, Atrium IIIR Working Group June 1993 Hypertext Markup Language (HTML) A Representation of Textual Information and MetaInformation for Retrieval and Interchange Status of this Document 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 working documents valid for a maximum of six months. Internet Drafts may be updated, replaced, or obsoleted by other documents at any time. It is not appropriate to use Internet Drafts as reference material or to cite them other than as a "working draft" or "work in progress". Distribution of this document is unlimited. The document is a draft form of a standard for interchange of information on the network which is proposed to be registered as a MIME (RFC1341) content type. Please send comments to timbl@info.cern.ch or the discussion list www-talk@info.cern.ch. This is version 1.2 of this draft. This document is available in hypertext on the World-Wide Web as http://info.cern.ch/hypertext/WWW/MarkUp/HTML.html Abstract HyperText Markup Language (HTML) can be used to represent Hypertext news, mail, online documentation, and collaborative hypermedia; Menus of options; Database query results; Simple structured documents with inlined graphics. Hypertext views of existing bodies of information The World Wide Web (W3) initiative links related information throughout the globe. HTML provides one simple format for providing linked information, and all W3 compatible programs are required to be capable of handling HTML. W3 uses an Internet Berners-Lee and Connolly 1 protocol (Hypertext Transfer Protocol, HTTP), which allows transfer representations to be negotiated between client and server, the result being returned in an extended MIME message. HTML is therefore just one, but an important one, of the representations used with W3. HTML is proposed as a MIME content type. HTML refers to the URL specification of RFCxxxx. Implementations of HTML parsers and generators can be found in the various W3 servers and browsers, in the public domain W3 code, and may also be built using various public domain SGML parsers such as [SGMLS] . HTML is an SGML document type with fairly generic semantics appropriate for representing information from a wide range of applications. It is more generic than many specific SGML applications, but is still completely device-independent. IN THIS DOCUMENT This document contains the following parts: Vocabulary used in this document, degrees of imperative. HTML and MIME with discussion of character sets. HTML and SGML and the relationship between them, and Structured text : an introduction for beginners to SGML. HTML Elements A list with description, example, and typical rendering. HTML Entities Entities used to describe characters. The HTML DTD The text of the SGML DTD for HTML Link relationship values . A provisional list. Not part of the standard. Registration Authority The authority for extending lists of valid vales. Acknowledgements and a change history of the document References to related documents Authors addresses Contact information. table of contents Berners-Lee and Connolly 2 Vocabulary This specification uses the words below with the precise meaning given. Representation The encoding of information for interchange. For example, HTML is a representation of hypertext. Rendering The form of presentation to information to the human reader. IMPERATIVES may The implementation is not obliged to follow this in any way. must If this is not followed, the implementation does not conform to this specification. shall as "must" should If this is not followed, though the implementation officially conforms to the standard, undesirable results may occur in practice. typical Typical rendering is described for many elements. This is not a mandatory part of the standard but is given as guidance for designers and to help explain the uses for which the elements were intended. NOTES Sections marked "Note:" are not mandatory parts of the specification but for guidance only. STATUS OF FEATURES Mainstream All parsers must recognize these features. Features are mainstream unless otherwise mentioned. Extra Standard HTML features which may safely be ignored by parsers. It is legal to ignore these, treat the contents as though the tags were not there. (e.g. EM, and any undefined elements) Obsolete Not standard HTML. Parsers should implement these features as far as possible in order to preserve back-compatibility with previous Berners-Lee and Connolly 3 versions of this specification. HTML AND MIME The definition of the HTML content subtype is MIME Type name text MIME subtype name: html Required parameters: none Optional parameters: charset Character sets The base character set (the SGML BASESET) for HTML is ISO Latin-1. This is the set referred to by any numeric character references . The actual character set used in the representation of an HTML document may be ISO Latin 1, or its 7-bit subset which is ASCII. There is no obligation for an HTML document to contain any characters above decimal 127. It is possible that a transport medium such as electronic mail imposes constraints on the number of bits in a representation of a document, though the HTTP access protocol used by W3 always allows 8 bit transfer. When an HTML document is encoded using 7-bit characters, then the mechanisms of character references and entity references may be used to encode characters in the upper half of the ISO Latin-1 set. In this way, documents may be prepared which are suitable for mailing through 7-bit limited systems. INTRODUCTION The HyperText Markup Language is defined in terms of the ISO Standard Generalized Markup Language []. SGML is a system for defining structured document types and markup languages to represent instances of those document types. Every SGML document has three parts: An SGML declaration, which binds SGML processing quantities and syntax token names to specific values. For example, the SGML declaration in the HTML DTD specifies that the string that opens a tag is A sample HTML instance

An Example of Structure

Here's a typical paragraph.

Some elements (e.g. P, LI) are empty. They have no content. They show up as just a start tag. For the rest of the elements, the content is a sequence of data characters and nested elements. Note that the HTML DTD in fact Berners-Lee and Connolly 5 severely limits the amount of nesting which is allowed: most things cannot be nested, in fact. No elements may be recursively nested. Anchors and character highlighting may be put inside other constructs. TAGS Every element starts with a tag, and every non-empty element ends with a tag. Start tags are delimited by < and >, and end tags are delimited by . Names The element name immediately follows the tag open delimiter. Names consist of a letter followed by up to 33 letters, digits, periods, or hyphens. Names are not case sensitive. Attributes In a start tag, whitespace and attributes are allowed between the element name and the closing delimiter. An attribute consists of a name, an equal sign, and a value. Whitespace is allowed around the equal sign. The value is specified in a string surrounded by single quotes or a string surrounded by double quotes. (See: other tolerated forms @@) The string is parsed like RCDATA (see below ) to determine the attribute value. This allows, for example, quote characters in attribute values to be represented by character references. The length of an attribute value (after parsing) is limited to 1024 characters. ELEMENT TYPES The name of a tag refers to an element type declaration in the HTML DTD. An element type declaration associates an element name with A list of attributes and their types and statuses A content type (one of EMPTY, CDATA, RCDATA, ELEMENT, or MIXED) which determines the syntax of the element's content A content model, which specifies the pattern of nested elements and data Empty Elements Empty elements have the keyword EMPTY in their declaration. For example: Berners-Lee and Connolly 6 This means that the following: is legal, but these others are not: Character Data The keyword CDATA indicates that the content of an element is character data. Character data is all the text up to the next end tag open delimiter-in-context. For example: specifies that the following text is a legal XMP element: Here's an example. It looks like it has <tags> and <!--comments--> in it, but it does not. Even this </ is data. The string There is no way to represent tags in CDATA Replaceable Character Data Elements with RCDATA content behave much like those with CDATA, except for character references and entity references. Elements declared like: can have any sequence of characters in their content. Character References To represent a character that would otherwise be recognized as markup, use a character reference. The string &# signals a character reference when it is followed by a letter or a digit. The delimiter is followed by the decimal character number and a semicolon. For example: Berners-Lee and Connolly 7 You can even represent </end> tags in RCDATA Entity References The HTML DTD declares entities for the less than, greater than, and ampersand characters and each of the ISO Latin 1 characters so that you can reference them by name rather than by number. The string & signals an entity reference when it is followed by a letter or a digit. The delimiter is followed by the entity name and a semicolon. For example: Kurt Gödel was a famous logician and mathematician. Note: To be sure that a string of characters has no markup, HTML writers should represent all occurrences of <, >, and & by character or entity references. Element Content Some elements have, in stead of a keyword that states the type of content, a content model, which tells what patterns of data and nested elements are allowed. If the content model of an element does not include the symbol #PCDATA , the content is element content. Whitespace in element content is considered markup and ignored. Any characters that are not markup, that is, data characters, are illegal. For example: declares an element that may be used as follows: Head Example But the following are illegal: no data allowed! Two isindex tags Mixed Content If the content model includes the symbol #PCDATA, the content of the element is parsed as mixed content. For example: Berners-Lee and Connolly 8 This says that the PRE element contains one or more A, B, I, U, or P elements or data characters. Here's an example of a PRE element: 
NAME
    cat -- concatenatefiles
EXAMPLE
    cat 

   The content of the above PRE element is: 
   
      A B element 
      
      The string ``   cat -- concatenate'' 
      
      An A element 
      
      The string ``\n'' 
      
      Another B element 
      
      The string ``\n   cat . After the comment
   delimiter, all text up to the next occurrence of -- is ignored.
   Hence comments cannot be nested. Whitespace is allowed between the
   closing -- and >. (But not between the opening 
HTML Guide: Recommended Usage


 

   There are a few other SGML markup constructs that are deprecated or
   illegal. 
   
  Delimiter               Signals... 
                         
  .  
                         
  . 
                         
  LINE BREAKS
  
   A line break character is considered markup (and ignored) if it is
   the first or last piece of content in an element. This allows you
   to write either 
   

some example text


   or 
   

some example text



   and these will be processed identically.
   
   Also, a line that's not empty but contains no content will be
   ignored altogether. For example, the element 
   


first line
 
third line
fourth line



   contains only the strings 
   

 first line

 third line
 fourth line.

  SPACES AND TABS
  
   Space characters must be rendered as horizontal white space.  In
   HTML, multiple spaces should be rendered as single spaces.
   
   The rendering of a horizontal tab (HT) character is not defined,
   and HT should therefore not be used, except within a PRE (or
   obsolete XMP, LISTING or PLAINTEXT) element.
   
   Neither spaces nor tabs should be used to make SGML source layout
   more attractive or easier to read. 



Berners-Lee and Connolly                                             10

  SUMMARY OF MARKUP SIGNALS
  
   The following delimiters may signal markup, depending on context. 
   
  Delimiter               Signals 
                         
  




























 






Berners-Lee and Connolly                                             35


        






























 






Berners-Lee and Connolly                                             36


















































Berners-Lee and Connolly                                             37















































]>









Berners-Lee and Connolly                                             38

                       LINK RELATIONSHIP VALUES
                                   
   Status: This list is not part of the standard.  It is intended to
   illustrate the use of link relationships and to provide a framework
   for further development.
   
   Additions to this list will be controlled by the HTML registration
   authority . Experimental values may be used on the condition that
   they begin with "X-".
   
   Link relationship valies are NOT case sensitive. That is, "Made"
   and "made" have th esame meaning.
   
   These values of the REL attribute of hypertext links have a
   significance defined here, and may be treated in special ways by
   HTML applications.
   
   These relationships relate whole documents (objects), rather than
   particular anchors within them. If the relationship value is used
   with a link between anchors rather than whole documents, the
   semantics are considered to apply to the documents.
   
   In the explanations which follows, A is the source document of the
   link and B is the destination document specified by the HREF
   attribute.
   
   A relationship marked "Acyclic" has the property that no sequence
   of links with that relationship may be followed from any document
   back to itself. These types of links may therefore be used to
   define trees. 
   
   Most relationships (except where noted) are between the objecs
   themselves rather than the subjects of the objects. Objects may be
   documents, images, people (with mailto: URIs for example.) 
   
  USEINDEX
  
   B is a related index for a search by a user reading this document
   who asks for an index search function.
   
   A document may have any number of index links, causing several
   indexes top be searched in a client-defined manner.
   
   B must support SEARCH operations under its access protocol. 
   
  USEGLOSSARY
  
   B is an index which should be used to resolve glossary queries in
   the document. (Typically, a double-click on a word which is not
   within an anchor).
   
   A document may have any number of glossary links. 
   



Berners-Lee and Connolly                                             39

  ANNOTATION
  
   The information in B is additional to and subsidiary to that in A. 
   
   Annotation is used by one person to write the equivalent of "margin
   notes" or other criticism on another's document, for example.
   
   Example: The relationship between a newsgroup and its articles.
   
   Acyclic. 
   
  REPLY
  
   Similar to Annotation, but there is no suggestion that B is
   subsidiary to A: A and B are on equal footings.
   
   Example: The relationship between a mail message and its reply, a
   news article and its reply.
   
   Acyclic. 
   
  EMBED
  
   If this link is followed, the node at the end of it is embedded
   into the display of the source document.
   
   Acyclic. 
   
  PRECEDES
  
   In an ordered structure defined by the author, A precedes B, B is
   followed by A.
   
   Acyclic. 
   
   Any document may only have one link of this relationship, and/or
   one link of the reverse relationship.
   
   Note: May be used to control navigational aids, generate printed
   material, etc.  In conjunction with " subdocument ", may be used to
   define a tree such as a printed book made of hypertext document. 
   The document can only have one such tree. 
   
  SUBDOCUMENT
  
   B is a lower part in the author's hierarchy to A.  Acyclic. See
   also Precedes . 
   
  PRESENT
  
   Whenever A is presented, B must also be presented.  This implies
   that whenever A is retrieved, B must also be retrieved. 
   



Berners-Lee and Connolly                                             40

  SEARCH
  
   When the link is followed, the node B should be searched rather
   than presented. That is, where the client software allows it, the
   user should immediately be presented with a search panel and
   prompted for text. The search is then performed without an
   intermediate retrieval or presentation of the node B 
   
  SUPERSEDES
  
   B is a previous version of  A. 
   
   Acyclic. 
   
  HISTORY
  
   B is a list of versions of A
   
   A link reverse link must exist from B to A and to all other known
   versions of A.  
   
  MADE
  
   Person (etc) described by node A is author of B
   
   This information can be used for protection, and informing authors
   of interest, for sending mail to authors, etc. 
   
  OWNS
  
   The owner of an object carried resposabiliy for and authority over
   the object.
   
   This information may be used for finding people responsible for
   incorrect information, etc.  The creator (Made) and owner (Owns) of
   an object may be different. 
   
  APPROVES
  
   Approval of objects is a method of attributing value, or fiability,
   to objects.  One determination of a value of an object is a
   function of the  set of objects which approve it.
   
   A reviewed journal, for example, may operate by approving articles.
    This could be expressed by an approval link from the journal
   itself to the article.   In the view of the web as an
   encyclopaedia, approval links one to filter information which has a
   certain quality according to some standard. 
   
  SUPPORTS
  
   A and B are objects representing assertions.  The assertion A
   supports the assertion B.   This may be used to overlay a weak



Berners-Lee and Connolly                                             41

   semantics of argument onto the web. For example, giving such
   relaionships within discussions  in this way will allow  arguments
   to be analysed by machine and followed by people with greater ease.
   See also: refutes. 
   
  REFUTES
  
   This is the opposite of "Supports", indicating that A is a
   proposition which refutes proposition B. 
   
  INCLUDES
  
   A includes B, B is part of A.  For example, a person described by
   document A is a part of the group described by document B.
   
   Note: This relationship conveys semantics about objects described
   by objects, rather than the documents themselves.
   
   Acyclic. 
   
  INTERESTED
  
   Person (etc) described by A is interested in node B. 
   
   This information can be used for notification of changes.
   
   Typically, this is a request that, when object B changes in some
   way, a new link is made to object A. 
   
   The phrase "object B changes" may be interpreted narrowly (as "B
   itself changes") or widely (as "B or anythink linked to it or
   related to it closely changes").   The amount of  change considered
   worth notifying people about is also subject to interpretation,
   varying from bit changes in the source to a "new edition" statement
   by the publisher.
   
                        REGISTRATION AUTHORITY
                                   
   The HTML Registration Authority is responsible for maintaining
   lists of: 
   
      Relationship names for link and anchor elements 
      
   It is proposed that a WWW consortium, the Internet Assigned Numbers
   Authority,  or their successors take this role.
   
   Unregistered values may be used for experimental purposes if they
   are start with "X-".
   
                           ACKNOWLEDGEMENTS
                                   
   The HTML document type was designed initially at CERN in 1990 for
   the World-Wide Web project.     The DTD was written, and the



Berners-Lee and Connolly                                             42

   specification tightened up, by Dan Connolly After much discussion
   on the network and some enhancement in particular the addition of
   inline images introduced by the NCSA "Mosaic" software for WWW, it
   was released as an Internet draft in 1993.
   
   This version of the specification follows from certain minor
   changes made at the WWW Wizard's Workshop in Cambridge, Mass., in
   July 1993, in particular the introduction of 
, 
,  . 
   
                                                                Tim BL
                                                                      
                              REFERENCES
                                   
  SGML                    ISO 8879:1986, Information Processing Text
                         and Office Systems Standard Generalized
                         Markup Language (SGML). 
                         
  sgmls                   an SGML parser by James Clark
                          derived from the ARCSGML
                         parser materials which were written by
                         Charles F. Goldfarb. The source is available
                         on the ifi.uio.no FTP server in the directory
                         /pub/SGML/SGMLS . 
                         
  WWW                     The World-Wide Web , a global information
                         initiative. For bootstrap information, telnet
                         info.cern.ch or find documents by
                         ftp://info.cern.ch/pub/www/doc 
                         
  URL                     Universal Resource Locators. RFCxxx. 
                         Currently  available by anonymous FTP from
                         info.cern.ch in /pub/www/doc/url*.{ps,txt}
                         
                          AUTHOR'S ADDRESSES
                                   
   This document was prepared with the help and advice of many people
   across the net.  Dan Connolly prepared the DTD and the section on
   HTML and SGML whilst with Convex Computer Corporation of 3000
   Waterview Parkway Richardson, TX 75083. He is now with Atrium
   Technology Inc., and is not a current editor of the document. 
   
                                Tim Berners-Lee
                Address         CERN
                                1211 Geneva 23
                                Switzerland
                Telephone:      +41(22)767 3755
                Fax:            +41(22)767 7155
                email:          timbl@info.cern.ch

                                Daniel Connolly
                Address:        Atrium Technologies, Inc.
                                5000 Plaza on the Lake, Suite 275
                                Austin, TX 78746



Berners-Lee and Connolly                                             43

                                USA
                email:          connolly@atrium.com


   




















































Berners-Lee and Connolly                                             44