<!--
        html3.dtd

        Document Type Definition for the HyperText Markup Language (HTML 
DTD)

        Draft: Mon 13-Mar-95 09:51:25

        Author: Dave Raggett <dsr@hplb.hpl.hp.com>

        W3O is developing a testbed browser to provide practical
        experience with HTML 3.0 before it becomes a standard.
        See:  http://www.w3.org/hypertext/WWW/Arena/

        This is an open process and comments are welcomed on the
        www-html mailing list.

        While HTML 3.0 is being discussed, please use text/x-html3
        as a temporary MIME content type, as this will allow clients
        to distinguish html 3 from current html documents.

        The entity HTML.Recommended can be used to give a more rigorous
        version of the DTD suitable for use with SGML authoring tools.
        The default version of the DTD offers a laxer interpretation,
        e.g. allowing authors to omit leading <P> elements. You can
        switch on the more rigorous version of the DTD by including
        the following at the start of your HTML document.

         <!DOCTYPE HTML PUBLIC "-//W3O//DTD W3 HTML 3.0//EN//"
            [ <!ENTITY % HTML.Recommended "INCLUDE"> ] >


        Design Objectives:

        o   Backwards compatibility with 2.0

        o   Tightening up HTML.Recommended and
            moving more things to HTML.Deprecated

        o   Keep HTML - simple don't compete with CALS

        o   Make it practical for people to edit HTML 3.0
            documents directly, i.e. avoid long names.

        o   Tables, figures and math from HTML+
            with tweaks based on recent experience

        o   Client-side event handling for figures
            and graphical form selection menus

        o   Add limited presentational controls with
            a view to use of linked style sheets
            (style overrides are supported)

        o   Compatibility with ICADD as per Yuri's suggestions

        HTML 3.0 relies on linked style info to give authors
        control over the appearence of documents. Such info is
        placed in a linked style sheet, or as overrides in the
        HTML document head, using the STYLE element. The generic
        CLASS attribute can be used to subclass elements when
        you want to use a different style from normal, e.g. you
        might use <h2 class=3Dbigcaps> for headers with enlarged
        capital letters. Note that the class attribute has a
        wider scope than just style changes, e.g. browsers could
        provide the means for searching through documents,
        restricting search according to element class values.

        The DTD contains a small number of attributes for direct
        control of basic alignment parameters; column widths for
        tables; support for custom bullets, sequence numbering for
        lists and headers; and text flow. These attributes offer
        control over appearence which would be inconvenient to
        express exclusively via associated style sheets.

        The MD attribute for each hypertext or inline link specifies a
        message digest such as MD5 for the linked object and is needed
        to ensure someone hasn't tampered with a linked document.

  History:

    13th March '95
    

    Dropped MARK tag on advice from SGML Open
    Allowed spaces in table colspec attribute
    Changed ARRAY element
    Added CHOOSE tag to BOX element
    Cleaned up PRE content model
    Obsoleted tags incompatible with SGML
    

    6th March '95
    

    Added several tags to MATH: %mathface, %mathvec
    and improved ROOT, with new SQRT convience tag
    

    1st March '95
    

    Dropped align attribute from BR element
    Added indent attribute to TAB
    Added optional CREDIT to end of BQ
    Changed FIG to %body.content to allow headers
    

    22nd February '95
    

    Added align attribute, and dropped before, after, center and right
      attributes to clean up TAB element
    Added INS and DEL for legal documents    

    Added CREDIT to end of FIG element
    Dropped FN in favor of <NOTE ROLE=3DFootNote>
    

    9th Feburuary '95

    Dropped base attribute mechanism for scoping relative URLs
    Dropped nofold attribute for disabling whitespace folding
    Dropped border width attributes for FIG (-> style sheet)
    Dropped delims attribute from math BOX element
    Dropped stylistic attributes from OL such as inherit (-> stylesheet)
    Added baseline to list of valign attribute values for tables.
    Added DIV element for generic container class and static banners
    Added MARK element for marked range class
    Added closed set of LINK REL values for toolbars
    Added numbering attributes to headers
    Added bullet attributes to headers
    Added TERM element to math for style sheet control of term rendering
    Changed to imagemap=3DURI for server-side event handling for FIG/OVERLAY
    Changed delimiter attributes for math arrays
    Changed ROOT element for maths to allow an arbitrary radix
    Simplified numbering attributes for ordered lists
    Simplified STYLE element to leave binding to style language
-->
<!-- CL: ok lets get rid of this

<!DOCTYPE HTML [

because:
    sgmls: SGML error at /afs/mcc.ac.gb/users/cgu/cgu_sources/Dir9/html-check/
       lib/html-3.dtd, line 103 at "E":
     DOCTYPE markup declaration not permitted here; declaration ended
CL: -->

<!ENTITY % HTML.Version
        "-//W3O//DTD W3 HTML 3.0//EN//"

        -- Typical usage:

            <!DOCTYPE HTML PUBLIC "-//W3O//DTD W3 HTML 3.0//EN//">
            <html>
            ...
            </html>
        --
        >

<!ENTITY % HTML.Recommended "IGNORE"
        -- Certain features of the language are necessary for compatibility
           with widespread usage, but they may compromise the structural
           integrity of a document. This feature test entity enables
           a more prescriptive document type definition that eliminates
           the above features.
        -->

<![ %HTML.Recommended [
        <!ENTITY % HTML.Deprecated "IGNORE">
]]>

<!ENTITY % HTML.Deprecated "INCLUDE"
        -- Certain features of the language are necessary for compatibility
           with earlier versions of the specification, but they tend
           to be used an implemented inconsistently, and their use is
           deprecated. This feature test entity enables a document type
           definition that eliminates these features.
        -->

<!ENTITY % HTML.Obsoleted "IGNORE"
        -- The XMP, LISTING and PLAINTEXT tags are incompatible with SGML
           and derive from very early versions of HTML. They require non-
           standard parsers and will cause problems for processing
           documents with standard SGML tools.
        -->

<!ENTITY % Content-Type "CDATA"
        -- meaning a MIME content type, as per RFC1521
        -->

<!ENTITY % HTTP-Method "GET | POST"
        -- as per HTTP specification
        -->

<!ENTITY % URI "CDATA"
        -- The term URI means a CDATA attribute
           whose value is a Uniform Resource Identifier,
           as defined by
        "Uniform Resource Identifiers" by Tim Berners-Lee
        aka http://info.cern.ch/hypertext/WWW/Addressing/URL/URI_Overview2E.html
        aka RFC 1630

        Note that CDATA attributes are limited by the LITLEN
        capacity (1024 in the current version of html.decl),
        so that URIs in HTML have a bounded length.

        -->

<!ENTITY % Misc.Relations "stylesheet|node|path">

<!ENTITY % FLOAT "CDATA" -- floating point numbers (not in SGML) -->

<!ENTITY % SHAPE "CDATA"

    -- Shape of hotzone in image.

       All coordinates are assumed to be numbers in the range 0 to 1
       and interpreted as fractional width/height and measured from
       the top left corner of the associated image.

       The attribute value is a string taking one of the following forms:

       		"default"

       Used to define a default link for the figure background.

            "circle x, y, r" 

       (x, y) define the center and r the radius.

            "rect x, y, w, h"

       (x, y) defines upper left, and w and h the width and height.

            "polygon x1, y1, x2, y2, ..."

       Given n pairs of x, y coordinates, the polygon is closed by a
       line linking the n'th point to the first. Intersecting polygons
       use the non-zero winding number rule to determine if a point lies
       inside the polygon.I
    --
>

<!-- 3.0 Parameter Entities -->

<!ENTITY % heading "H1|H2|H3|H4|H5|H6">

<![ %HTML.Obsoleted [
    <!ENTITY % preformatted "PRE | XMP | LISTING">
]]>


<![ %HTML.Deprecated [
    <!ENTITY % list "UL | OL | DIR | MENU">
    <!ENTITY % blockquote "BLOCKQUOTE | BQ">
]]>

<!ENTITY % list "UL | OL">

<!ENTITY % blockquote "BQ">

<!ENTITY % preformatted "PRE">


<!-- The style attribute is used to subclass HTML elements for
     rendering purposes, when used with style sheets, e.g. DSSSL lite -->


<!ENTITY % attrs  -- common attributes for elements --
        'id       ID      #IMPLIED -- as target for hrefs (link ends) --
         lang     CDATA   "en.us"  -- ISO language, country code --
         class    NAMES   #IMPLIED -- for subclassing elements --'>

<!-- SGML standard forces different NAMES for all attribute values
     in the same element, regardless of the attribute name! As a result
     CDATA is used for CLEAR attribute to avoid clash with ALIGN attribute.-->

<!ENTITY % needs -- Attributes for controlling text flow. Used in headers
                    and other elements to guarantee sufficient room --
        'clear  CDATA "no"  -- (left|right|all|no) move down past figures --
         needs  CDATA #IMPLIED -- minimum width needed in em"s or pixels --
                                        -- e.g. "40 em" or "100 pixels"  --'>

<!--
    The following attribute may be included where ever a URL can be given:

    md      message digest e.g. md=3D"md5:jV2OfH+nnXHU8bnkPAad/mSQlTDZ"
            where the digest is base64 encoded and preceded by a prefix
            denoting the algorithm (in this case MD5).
-->

<!ENTITY % url.link -- Attributes associated with URL based links --
        "md     CDATA  #IMPLIED  -- message digest for linked object --">


<!-- The HTML list of Latin-1 entities includes the full range
     of characters in widely available Latin-1 fonts, and as such
     is a mixture of ISOlat1 and other ISO publishing symbols -->

<!ENTITY % HTMLlat1 PUBLIC
  "-//IETF//ENTITIES Added Latin 1 for HTML//EN">
%HTMLlat1;

<!ENTITY emsp   SDATA "[emsp  ]" -- em space -->
<!ENTITY ensp   SDATA "[ensp  ]" -- en space (1/2-em) -->
<!ENTITY mdash  SDATA "[ndash ]" -- em dash -->
<!ENTITY ndash  SDATA "[ndash ]" -- en dash (1/2-em) -->
<!ENTITY nbsp   SDATA "[nbsp  ]" -- non breaking space -->
<!ENTITY shy    SDATA "[shy   ]" -- soft hyphen -->
<!ENTITY copy   SDATA "[copy  ]" -- copyright sign -->
<!ENTITY trade  SDATA "[trade ]" -- trade mark sign -->
<!ENTITY reg    SDATA "[reg   ]" -- registered sign -->

<!-- a range of standard icons such as &folder; for use
     in speeding up display of directory listings etc. -->

<!ENTITY % HTMLicons PUBLIC
  "-//IETF//ENTITIES icons for HTML//EN">
%HTMLicons;

<!-- ISO subset chosen for use with the widely available Adobe math font 
-->

<!ENTITY % HTMLmath PUBLIC
  "-//IETF//ENTITIES Math and Greek for HTML//EN">
%HTMLmath;

<!ENTITY % font " U | S | TT | I | BIG | SMALL">

<!ENTITY % phrase "EM | STRONG | CODE | SAMP | KBD | VAR | CITE">

<!ENTITY % misc "Q | LANG | AU | DFN | PERSON | ACRONYM | ABBREV | INS | DEL">

<!ENTITY % special "TAB | MATH | A | IMG | BR">

<!ENTITY % notmath "%font | %phrase | %special | %misc">

<!ENTITY % text "#PCDATA | SUB | SUP | B | %notmath">


<!ENTITY % pre.exclusion "TAB|MATH|IMG|BIG|SMALL|SUB|SUP">

<!ELEMENT (%font|B|%phrase|%misc) - - (%text)+>
<!ATTLIST (%font|B|%phrase|%misc) %attrs;>

<!-- Subscripts and superscripts. The ALIGN attribute is only used for math -->

<!ELEMENT (SUB|SUP) - - (%text)+>
<!ATTLIST (SUB|SUP)
        %attrs;
        align (left|center|right) #IMPLIED
        >

<!-- Forced line break -->

<!ELEMENT BR    - O EMPTY>
<!ATTLIST BR
        %attrs;
        %needs; -- for control of text flow --
        >

<!-- Named left, center and right tab stops (independent of '\t' char) -->

<!ELEMENT TAB - O EMPTY>
<!ATTLIST TAB
        id       ID       #IMPLIED  -- defines named tab stop --
        indent   NUMBER   0         -- em units before new tab stop --
        to       IDREF    #IMPLIED  -- jump to named tab stop --
        align   (left|center|right) left -- alignment of following text --
        >

<!--
     With HTML 3.0 you can use ID attributes on most elements for named
     link ends. The use of the NAME attribute on anchors is deprecated.

     Do we want to support arbitrary elements for link starts? This would
     involve adding HREF and related attributes to most elements.
-->

<![ %HTML.Deprecated [
    <!ENTITY % linkName "name CDATA #IMPLIED -- named link end --">
]]>

<!ENTITY % linkName "">


<!ENTITY % ToolBar "home|toc|index|glossary|copyright|
                    up|previous|next|help|bookmark"
        -- LINK RELationship values which are used to create toolbar
           buttons or menu items for navigation, where toc stands
           for table of contents and bookmark provides for an open
           ended set of links, i.e. you can use multiple bookmarks
           for key entry points. Use the optional TITLE attribute
           to override default names.
        -->

<!ENTITY % linkType "NAME"
        -- A definitive list will be specified at a later date.

           They are used

                a) by stylesheets to control how collections of
                   html nodes are rendered into printed documents

                b) for document specific toolbars/menus when used
                   with the LINK element in document head:

                    "home|toc|index|glossary|copyright|
                           up|previous|next|help|bookmark"

                   where toc stands for table of contents and
                   bookmark provides for an open ended set of links,
                   i.e. you can use several bookmarks for key entry
                   points. Use the optional TITLE attribute to
                   override default names.

                c) for hypertext paths or guided tours,
                   with REL=3DNODE and REL=3DPATH. 

                d) to make a link to a style sheet, e.g. rel=3Dstyle
                   (used only with the LINK element).
        -->


<!ENTITY % linkExtraAttributes -- URN moved to %url.link --
       "rel %linkType #IMPLIED -- forward relationship type --
        rev %linkType #IMPLIED -- reversed relationship type
                                  to referent data --
        title   CDATA #IMPLIED -- advisory only --
        methods NAMES #IMPLIED -- supported public methods of the object:
                                        TEXTSEARCH, GET, HEAD, ... --
        ">

<![ %HTML.Deprecated [
    <!ENTITY % A.content "(%heading|%text)+">
]]>

<!ENTITY % A.content   "(%text)+">

<!ELEMENT A     - - %A.content -(A)>

<!ATTLIST A
        %attrs;
        href   %URI;  #IMPLIED
        %url.link;    -- standard link attributes --
        %linkName;     -- name attribute is deprecated; use ID instead --
        shape %SHAPE; #IMPLIED -- for shaped hotzones in FIGs --
        %linkExtraAttributes;
        >

<!-- Desired widths are used for negotiating image size
     with the module responsible for painting the image.
     align=3Dleft or right cause image to float to margin
     and for subsequent text to wrap around image -->

<!ELEMENT IMG    - O EMPTY --  Embedded image -->
<!ATTLIST IMG
        %attrs;
        src  %URI;  #REQUIRED  -- URI of image to embed --
        %url.link;             -- standard link attributes --
        alt CDATA   #IMPLIED   -- for display in place of image --
        align  (top|middle|bottomleft|right) top -- relative to baseline --
        width  NUMBER #IMPLIED -- desired width in em's or pixels --
        height NUMBER #IMPLIED -- desired height in em's or pixels --
        units  (em|pixels) pixels -- units for width and height --
        ismap (ismap) #IMPLIED -- pass clicks to server --
        >


<!ELEMENT P     - O (%text)+>
<!ATTLIST P
        %attrs;
        align  (left|center|right|justify) #IMPLIED
        %needs; -- for control of text flow --
        nowrap (nowrap) #IMPLIED -- disable wordwrap --
        >

<!ELEMENT HR    - O EMPTY -- customizable horizontal rule -->
<!ATTLIST HR
        %attrs;
        src     %URI;   #IMPLIED -- URI of custom rule graphic --
        %url.link;               -- standard link attributes --
        %needs; -- for control of text flow --
        >

<!--
  Headers can be numbered, although this is a matter for style sheets.
  The style sheet controls the numbering style:

     a) whether the parent numbering is inherited, e.g.  5.i.c
        where 5 is the current sequence number for H1 headers, and
        1 is the number for H2 headers and 3 for H3 headers.

     b) what style is used for current sequence number
        e.g. arabic, upperalpha, loweralpha, upperroman, lowerroman
        or a numbering scheme appropriate for the current language

  The skip attribute is used to skip over sequence numbers for items
  which have been left out of the list, e.g. skip=3D3 advances the
  sequence number past 3 omitted items. The seqnum sets the sequence
  number to a specified value. Note that the style sheet may take
  advantage of the sequence number for higher level headers.

  The dingbat or src attributes may be used to specify a bullet like
  image to be placed adjacent to the header. Defining this in the
  header element simplifies the document markup and avoids the need
  to use the clear or needs attribute in the following element to
  prevent it flowing around this image.
-->

<!ELEMENT ( %heading )  - -  (%text;)+>
<!ATTLIST ( %heading )
        %attrs;
        align  (left|center|right|justify) #IMPLIED
        %needs; -- for control of text flow --
        seqnum   NUMBER   #IMPLIED  -- starting sequence number --
        skip     NUMBER   0         -- skip seq nums for missing items --
        dingbat  ENTITY   #IMPLIED  -- dingbat entity from HTMLicons --
        src      (%URI;)  #IMPLIED  -- bullet defined by graphic --
        %url.link;                  -- standard link attributes --
        nowrap   (nowrap)  #IMPLIED -- disable wordwrap --
        >

<!ELEMENT TITLE - -  (#PCDATA)
          -- The TITLE element is not considered part of the flow of text=2E
             It should be displayed, for example as the page header or
             window title.
          -->

<!ENTITY % block
     "P | %list | DL
        | %preformatted
        | %blockquote
        | FORM | ISINDEX
        | TABLE | FIG | NOTE">

<!--
  ((%block)* | (%text)*) would be much nicer as it would avoid the need
  for a <P> tag when all you want is a few words of text. The problem
  is that it also prevents: "<LI> <P>some text" since it forbids PCDATA
  and hence the white space between the <LI> and the <P>.
-->

<![ %HTML.Recommended [
    <!ENTITY % flow "(%block)*">
]]>

<!ENTITY % flow "(%text | %block)*">

<!-- CL: altered on advice of Terry Allen from:
<!ELEMENT PRE - - text* -(%pre.exclusion)> to:
-->
<!ELEMENT PRE - - (%text;)* -(%pre.exclusion)>

<!ATTLIST PRE
        %attrs;
        width NUMBER #implied
        %needs; -- for control of text flow --
        >

<![ %HTML.Obsoleted [

<!ENTITY % literal "CDATA"
        -- special non-conforming parsing mode where
           the only markup signal is the end tag
           in full. This will cause problems for
           standard SGML tools!
        -->

<!ELEMENT XMP - -  %literal>
<!ELEMENT LISTING - -  %literal>
<!ELEMENT PLAINTEXT - O %literal>

]]>

<!ELEMENT DL    - -  (LH?, (DT|DD)+) -- this is perhaps too lax? -->
<!ATTLIST DL
        %attrs;
        %needs; -- for control of text flow --
        compact (compact) #IMPLIED -- more compact style --
        >

<!ELEMENT DT    - O  (%text)+>
<!ELEMENT DD    - O  %flow;>
<!ATTLIST (DT|DD)
        %attrs;
        %needs; -- for control of text flow --
        >


<!ELEMENT (OL|UL) - -  (LH?, LI+) -- should we allow a list header ? -->

<!-- style sheet controls numbering style
    a) whether the parent numbering is inherited, e.g.  5.1.c
    b) what style is used for current sequence number
        e.g. arabic, upperalpha, loweralpha, upperroman, lowerroman
        or a numbering scheme for the current language
-->
<!ATTLIST OL
        %attrs;
        %needs; -- for control of text flow --
        continue (continue)  #IMPLIED   -- don't restart sequence number --
        seqnum    NUMBER     #IMPLIED   -- starting sequence number --
        compact (compact)    #IMPLIED   -- reduced interitem spacing --
        >

<!-- Unordered lists:

    o   single or multicolumn with
        horizontal or vertical wrapping

    o   plain or bulletted list items

    o   bullets can be customised via:
            - entities (dingbats in HTMLicons)
            - external graphic via URL
            - individual attributes on LI tags
-->
<!ATTLIST UL
        %attrs;
        %needs; -- for control of text flow --
        wrap (vert|horiz|none) none -- multicolumn list style --
        plain   (plain)  #IMPLIED   -- suppress bullets --
        dingbat  ENTITY  #IMPLIED   -- dingbat entity from HTMLicons --
        src     (%URI;)  #IMPLIED   -- bullet defined by graphic --
        %url.link;                  -- standard link attributes --
        compact (compact) #IMPLIED  -- reduced interitem spacing --
        >

<!ELEMENT LH - O (%text;)+ -- list header -->
<!ATTLIST LH %attrs;>

<!--
    For unordered lists, you can override the standard bullet with
    a custom graphic specified via a URI e.g. src=3D"splash.gif" or
    a reference to one of the HTMLicons graphics e.g. dingbat=3Dfolder

    The skip attribute is used with ordered lists to skip over sequence
    numbers for items which have been left out of the list, e.g. skip=3D3
    advances the sequence number past 3 omitted items.
-->

<!ELEMENT LI - O %flow; -- list item -->
<!ATTLIST LI
        %attrs;
        %needs; -- for control of text flow --
        dingbat ENTITY #IMPLIED -- dingbat entity from HTMLicons --
        src  (%URI;) #IMPLIED   -- custom bullet graphic --
        %url.link;              -- standard link attributes --
        skip NUMBER 0           -- skip seq nums for missing items --
        >

<!-- DIR and MENU are now subsumed by UL with type=3Dplain. Use the
     wrap attribute to control wrapping style for multicolumn lists -->

<![ %HTML.Deprecated [
    <!ELEMENT (DIR|MENU) - - (LI)+ -(%block)>
    <!ATTLIST (DIR|MENU)
        compact (compact) #IMPLIED>
]]>

<![ %HTML.Recommended [
        <!ENTITY % body.content "(DIV|%heading|%block|HR|ADDRESS)*"
        -- <h1>Heading</h1>
           <p>Text ...
                is preferred to
           <h1>Heading</h1>
           Text ...
        -->
]]>

<!ENTITY % body.content "(DIV | %heading | %text | %block | HR | ADDRESS)*">

<!ELEMENT BODY O O  %body.content>
<!ATTLIST BODY
        %attrs;
        background %URI; #IMPLIED  -- texture tile for document background --
        >

<!ELEMENT (%blockquote) - - (%body.content, CREDIT?)>
<!ATTLIST (%blockquote)
        %attrs;
        %needs; -- for control of text flow --
        nowrap   (nowrap)  #IMPLIED -- disable wordwrap --
        >

<!ENTITY % address.content "((%text;)* | P*)">

<!ELEMENT ADDRESS - - %address.content>
<!ATTLIST ADDRESS
        %attrs;
        %needs; -- for control of text flow --
        >
        
<!--
  DIV can be used with the CLASS attribute to represent different
  kinds of container, e.g. chapter, section, abstract, appendix.
  CLASS=3DBanner is reserved for a banner section which appears at
  the top of the window and doesn't scroll with window contents.
  This can be used for corporate logos, copyright statements and
  disclaimers, as well as customized navigation/search controls.
-->

<!ELEMENT DIV - - %body.content>
<!ATTLIST DIV
        %attrs;
        align   (left|center|right) left -- alignment of following text --
        nowrap   (nowrap)  #IMPLIED -- disable wordwrap --
        >

<!--
    As HTML 2.0 plus a few extensions:

        a)  A RANGE control which varies between pair of values
            specified with the size attribute, e.g. SIZE=3D"1, 10"

        b)  FILE widget for uploading one or more files to a server

        c)  SCRIBBLE on image widget that sends the "ink" to the server

        d)  AUDIO widget for playing and recording audio samples

        e)  SUBMIT/RESET buttons can now be customised with an image.
            This subsumes the IMAGE type which is now deprecated.

        f)  Graphical SELECTion menus are now supported, using
            the new SHAPE attribute on OPTION elements.

    Further extensions are in the pipeline (e.g. table entry,
    multiple data formats for textarea fields and client-side
    scripts with custom widgets) but will have to wait until
    the backlog of implementation work diminishes.
-->

<!ELEMENT FORM - - %body.content -(FORM) +(INPUT|SELECT|TEXTAREA)>
<!ATTLIST FORM
        action %URI #REQUIRED -- server-side form handler --
        method (%HTTP-Method) GET -- see HTTP specification --
        enctype %Content-Type; "application/x-www-form-urlencoded"
        script %URI #IMPLIED -- link to client-side script --
        >

<![ %HTML.Deprecated [
    <!ENTITY % InputType "(TEXT | PASSWORD | CHECKBOX | RADIO | SUBMIT | RESET
                   | RANGE | AUDIO | FILE | SCRIBBLE | HIDDEN | IMAGE)">
]]>

<!ENTITY % InputType "(TEXT | PASSWORD | CHECKBOX | RADIO | SUBMIT | RESET
                   | RANGE | AUDIO | FILE | SCRIBBLE | HIDDEN)">

<!ELEMENT INPUT - O EMPTY>
<!ATTLIST INPUT
        %attrs;
        type %InputType TEXT
        name  CDATA #IMPLIED     -- required for all but submit and reset --
        value CDATA #IMPLIED
        src   %URI  #IMPLIED     -- for fields with background images --
        %url.link;               -- standard link attributes --
        checked (checked) #IMPLIED -- for radio buttons and check boxes --
        size CDATA #IMPLIED    -- like NUMBERS,
                                  but delimited with comma, not space --
        maxlength NUMBER #IMPLIED
        align  (top|middle|bottom) top -- relative to baseline --
        >

<!-- SRC attribute added for graphical selection menus -->

<!ELEMENT SELECT - - (OPTION+) -(INPUT|TEXTAREA|SELECT)>
<!ATTLIST SELECT
        %attrs;
        name CDATA #REQUIRED
        size NUMBER #IMPLIED
        multiple (multiple) #IMPLIED
        src  %URI  #IMPLIED   -- for graphical selection menus --
        %url.link;            -- standard link attributes --
        >

<!ELEMENT OPTION - O (#PCDATA)>
<!ATTLIST OPTION
        %attrs;
        selected (selected) #IMPLIED
        value  CDATA  #IMPLIED -- default to element content --
        shape %SHAPE; #IMPLIED -- for graphical selection menus --
        >

<!ELEMENT TEXTAREA - - (#PCDATA) -(INPUT|TEXTAREA|SELECT)>
<!ATTLIST TEXTAREA
        %attrs;
        name CDATA #REQUIRED
        rows NUMBER #REQUIRED
        cols NUMBER #REQUIRED
        >

<!ELEMENT CAPTION - - (%text;)+ -- table or figure caption -->
<!ATTLIST CAPTION
        %attrs;
        align (top|bottom|left|right) #IMPLIED
        >
<!--
    Tables and figures can be aligned in several ways:

    bleedleft   flush left with the left (window) border
    left        flush left with the left text margin
    center      centered (text flow is disabled for this mode)
    right       flush right with the right text margin
    bleedright  flush right with the right (window) border
    justify     when applicable the table/figure should stretch
                to fill space between the text margins

    Note: text will flow around the table or figure if the browser
    judges there is enough room and the alignment is not centered
    or justified. The table or figure may itself be part of the
    text flow around some earlier figure. You can in this case use
    the clear or needs attributes to move the new table or figure
    down the page beyond the obstructing earlier figure. Similarly,
    you can use the clear or needs attributes with other elements
    such as headers and lists to move them further down the page.
-->

<!ENTITY % block.align
        "align  (bleedleft|left|center|right|bleedright|justify) center">


<!--
    The HTML 3.0 table model has been chosen for its simplicity
    and the ease in writing filters from common DTP packages.

    By default the table is automatically sized according to the
    cell contents and the current window size. Specifying the columns
    widths using the colspec attribute allows browsers to start
    displaying the table without having to wait for last row.

    The colspec attribute is a list of column widths and alignment
    specifications. The columns are listed from left to right with
    a capital letter followed by a number, e.g. COLSPEC=3D"L20 C8 L40".
    The letter is L for left, C for center, R for right alignment of
    cell contents. J is for justification, when feasible, otherwise
    this is treated in the same way as L for left alignment.
    Column entries are delimited by one or more space characters.
    
    The number specifies the width in em's, pixels or as a
    fractional value of the table width, as according to the
    associated units attribute. This approach is more compact
    than used with most SGML table models and chosen to simplify
    hand entry. The width attribute allows you to specify the
    width of the table in pixels, em units or as a percentage
    of the space between the current left and right margins.

    To assist with rendering to speech, row and column headers
    can be given short names using the AXIS attribute. The AXES
    attribute is used to explicitly specify the row and column
    names for use with each cell. Otherwise browsers can follow
    up columns and left along rows (right for some languages)
    to find the corresponding header cells.

    Table content model: Braille limits the width of tables,
    placing severe limits on column widths. User agents need
    to render big cells by moving the content to a note placed
    before the table. The cell is then rendered as a link to
    the corresponding note.

    To assist with formatting tables to paged media, authors
    can differentiate leading and trailing rows that are to
    be duplicated when splitting tables across page boundaries.
    The recommended way is to subclass rows with the CLASS attribute
    For example: <TR CLASS=3DHeader>, <TR CLASS=3DFooter> are used for
    header and footer rows. Paged browsers insert footer rows at
    the bottom of the current page and header rows at the top of
    the new page, followed by the remaining body rows.
-->

<!ELEMENT TABLE - - (CAPTION?, TR*) -- mixed headers and data -->
<!ATTLIST TABLE
        %attrs;
        %needs; -- for control of text flow --
        border (border) #IMPLIED -- draw borders --
        colspec CDATA   #IMPLIED -- column widths and alignment --
        units  (em|pixels|relative) em -- units for column widths --
        width NUMBER    #IMPLIED -- absolute or percentage width --
        %block.align;  -- horizontal alignment --
        nowrap (nowrap) #IMPLIED -- don't wrap words --
        >

<!ENTITY % cell "TH | TD">
<!ENTITY % vertical.align "top|middle|bottom|baseline">

<!--
    Browsers should tolerate an omission of the first <TR>
    tag as it is implied by the context. Missing trailing
    <TR>s implied by rowspans should be ignored.

    The alignment attributes act as defaults for rows
    overriding the colspec attribute and being in turn
    overridden by alignment attributes on cell elements.
    Use valign=3Dbaseline when you want to ensure that text
    in different cells on the same row is aligned on the
    same baseline regardless of fonts. It only applies
    when the cells contain a single line of text.
-->

<!ELEMENT TR - O (%cell)* -- row container -->
<!ATTLIST TR
        %attrs;
        align  (left|center|right|justify) #IMPLIED
        valign (%vertical.align) top -- vertical alignment --
        nowrap (nowrap) #IMPLIED -- don't wrap words --
        >

<!--
    Note that table cells can include nested tables.
    Missing cells are considered to be empty, while
    missing rows should be ignored, i.e. if a cell
    spans a row and there are no further TR elements
    then the implied row should be ignored.
-->

<!ELEMENT (%cell) - O %body.content>
<!ATTLIST (%cell)
        %attrs;
        colspan NUMBER    1      -- columns spanned --
        rowspan NUMBER    1      -- rows spanned --
        align  (left|center|right|justify) #IMPLIED
        valign (%vertical.align) top -- vertical alignment --
        nowrap (nowrap) #IMPLIED -- don't wrap words --
        axis CDATA #IMPLIED -- axis name, defaults to element content --
        axes CDATA #IMPLIED -- comma separated list of axis names --
        >

<!--
  The element contains text for use in non-graphical displays. Note that
  you can use the shape attribute in anchors to specify hotzones on images.
  This provides for local processing of pointer clicks and a unified method
  for dealing with graphical and non-graphical displays.

  Text is flowed around figures when the figure is left or right aligned.
  You can request the browser to move down until there is enough room for
  the next element, see the CLEAR and NEED attributes (in %needs)

  Figures offer a path towards embedding arbitrary information formats
  via some kind of OLE/OpenDoc mechanism.
-->

<!ELEMENT FIG - - (OVERLAY*, CAPTION?, %body.content;, CREDIT?) -(FIG|IMG)>
<!ATTLIST FIG
        %attrs;
        %needs;                  -- for control of text flow --
        src  %URI;  #REQUIRED    -- URI of document to embed --
        %url.link;               -- standard link attributes --
        %block.align;            -- horizontal alignment --
        width  NUMBER #IMPLIED   -- desired width in units --
        height NUMBER #IMPLIED   -- desired height in units --
        units (em|pixels) pixels -- specifies units as em's or pixels --
        imagemap (%URI) #IMPLIED -- pass background clicks to server --
        >

<!--
    Figure overlays. When combined with local caching, overlays
    provide a cheap way of modifying a larger base image sent as
    part of a previous page.
-->

<!ELEMENT OVERLAY - O EMPTY -- image overlay -->
<!ATTLIST OVERLAY
        src  %URI;  #REQUIRED    -- URI of image overlay --
        %url.link;               -- standard link attributes --
        units (em|pixels) pixels -- specifies units as em's or pixels --
        x      NUMBER   0        -- offset from left in units --
        y      NUMBER   0        -- offset from top in units --
        width  NUMBER #IMPLIED   -- desired width in units --
        height NUMBER #IMPLIED   -- desired height in units --
        imagemap (%URI) #IMPLIED -- pass background clicks to server --
        >

<!ELEMENT CREDIT - - (%text;)* -- source of image -->
<!ATTLIST CREDIT
        %attrs;
        >

<!ENTITY % note.roles "Simple|Note|Caution|Warning|Footnote">
<!ELEMENT NOTE - - %body.content; -- admonishment -->
<!ATTLIST NOTE
        %attrs;
        src %URI;   #IMPLIED  -- URI of custom graphic --
        %url.link;            -- standard link attributes --
        role (%note.roles) Simple -- kind of note --
        %needs; -- for control of text flow --
        >

<!-- Use &thinsp; &emsp; etc for greater control of spacing. -->

<!-- Subscripts and Superscripts

  <SUB> and <SUP> are used for subscripts and superscripts.

                                          i j
      X <SUP>i</SUP>Y<SUP>j</SUP>  is   X  Y

  i.e. the space following the X disambiguates the binding.
  The align attribute can be used for horizontal alignment,
  e.g. to explicitly place an index above an element:
                                              i
        X<sup align=3Dcenter>i</sup>  produces  X

Short references are defined for superscripts, subscripts and boxes
to save typing when manually editing HTML math, e.g.

      x^2^    is mapped to   x<sup>2</sup>
      y_z_    is mapped to   y<sub>z</sub>
      {a+b}   is mapped to   <box>a + b</box>

Note that these only apply within the MATH element and can't be
used in normal text!
-->
<!ENTITY REF1   STARTTAG   "SUP">
<!ENTITY REF2   ENDTAG     "SUP">
<!ENTITY REF3   STARTTAG   "SUB">
<!ENTITY REF4   ENDTAG     "SUB">
<!ENTITY REF5   STARTTAG   "BOX">
<!ENTITY REF6   ENDTAG     "BOX">

<!USEMAP MAP1   MATH>
<!USEMAP MAP2   SUP>
<!USEMAP MAP3   SUB>
<!USEMAP MAP4   BOX>

<!SHORTREF MAP1 "^" REF1
                "_" REF3
                "{" REF5 >

<!SHORTREF MAP2 "^" REF2
                "_" REF3
                "{" REF5 >

<!SHORTREF MAP3 "_" REF4
                "^" REF1
                "{" REF5 >

<!SHORTREF MAP4 "}" REF6
                "^" REF1
                "_" REF3
                "{" REF5 >

<!--
 The inclusion of %math and exclusion of %notmath is used here
 to alter the content model for the B, SUB and SUP elements,
 to limit them to formulae rather than general text elements.
-->

<!ENTITY % mathvec "VEC|BAR|DOT|DDOT|HAT|TILDE" -- common accents -->
<!ENTITY % mathface "B|T|BT" -- control of font face -->
<!ENTITY % math "BOX|ABOVE|BELOW|%mathvec|ROOT|SQRT|ARRAY|SUB|SUP|%mathface">
<!ENTITY % formula "#PCDATA|%math">

<!ELEMENT MATH - - (#PCDATA)* -(%notmath) +(%math)>
<!ATTLIST MATH
        id      ID      #IMPLIED
        model   CDATA   #IMPLIED>

<!-- The BOX element acts as brackets. Delimiters are optional and
     stretch to match the height of the box. The OVER element is used
     when you want a line between numerator and denominator. This line
     is suppressed with the alternative ATOP element. CHOOSE acts like
     ATOP but adds enclosing round brackets as a convenience for binomial
     coefficients. Note the use of { and } as shorthand for <BOX> and
     </BOX> respectively:

                           1 + X
     {1 + X<OVER>Y}  is   _______
                             Y

                                 a + b
     {a + b<ATOP>c - d} is  
                                 c - d

     The delimiters are represented using the LEFT and RIGHT
     elements as in:

     {[<LEFT>x + y<RIGHT>]}   is   [ x + y ]
     {(<LEFT>a<RIGHT>]}       is   (a]
     {||<LEFT>a<RIGHT>||}     is   || a ||

     Use &lbrace; and &rbrace; for "{" and "}" respectively as
     these symbols are used as shorthand for BOX, e.g.

     {&lbrace;<LEFT>a+b<RIGHT>&rbrace;}  is  {a+b}

     You can stretch definite integrals to match the integrand, e.g.

     {&int;<SUB>a</SUB><SUP>b</SUP><LEFT>{f(x)<over>1+x} dx}

            b
            /  f(x)
            | ----- dx
            / 1 + x
            a

     Note the complex content model for BOX is a work around
     for the absence of support for infix operators in SGML.

     You can get oversize delimiters with the SIZE attribute,
     for example <BOX SIZE=3Dlarge>(<LEFT>...<RIGHT>)</BOX>

     Note that the names of common functions are recognized
     by the parser without the need to use "&" and ";" around
     them, e.g. int, sum, sin, cos, tan, ...
-->

<!ELEMENT BOX - - ((%formula)*, (LEFT, (%formula)*)?,
                   ((OVER|ATOP|CHOOSE), (%formula)*)?,
                   (RIGHT, (%formula)*)?)>
<!ATTLIST BOX
        size  (normal|medium|large|huge) normal -- oversize delims -->

<!ELEMENT (OVER|ATOP|CHOOSE|LEFT|RIGHT) - O EMPTY>

<!-- Horizontal line drawn ABOVE contents
     The symbol attribute allows authors to supply
     an entity name for an accent, arrow symbol etc.
     Generalisation of LaTeX's overline command.
 -->

<!ELEMENT ABOVE - - (%formula)+>
<!ATTLIST ABOVE symbol ENTITY #IMPLIED>

<!-- Horizontal line drawn BELOW contents
     The symbol attribute allows authors to 
     supply an entity name for an arrow symbol etc.
     Generalisation of LaTeX's underline command.
 -->

<!ELEMENT BELOW - - (%formula)+>
<!ATTLIST BELOW symbol ENTITY #IMPLIED>

<!-- Convenience tags for common accents:
     vec, bar, dot, ddot, hat and tilde
-->

<!ELEMENT (%mathvec) - - (%formula)+>

<!--
  T and BT are used to designate terms which should
  be rendered in an upright font (& bold face for BT)
-->

<!ELEMENT (T|BT) - - (%formula)+>
<!ATTLIST (T|BT) class NAMES #IMPLIED>

<!-- Roots  e.g. <ROOT>3<OF>1+x</ROOT> -->

<!ELEMENT ROOT - - ((%formula)+, OF, (%formula)+)>
<!ELEMENT OF - O (%formula)* -- what the root applies to -->

<!ELEMENT SQRT - - (%formula)* -- square root convenience tag -->

<!-- LaTeX like arrays. The COLDEF attribute specifies
     a single capital letter for each column determining
     how the column should be aligned, e.g. coldef=3D"CCC"

        "L"     left
        "C"     center
        "R"     right
        
     An optional separator letter can occur between columns
     and should be one of + - or =3D, e.g. "C+C+C+C=3DC".
     Whitespace within coldef is ignored. By default, the
     columns are all centered.
     
     The ALIGN attribute alters the vertical position of the
     array as compared with preceding and following expressions.

     Use LDELIM and RDELIM attributes for delimiter entities.
     When the LABELS attribute is present, the array is
     displayed with the first row and the first column as
     labels displaced from the other elements. In this case,
     the first element of the first row should normally be
     left blank.

     Use &vdots; &cdots; and &ddots; for vertical, horizontal
     and diagonal ellipsis dots. Use &dotfill; to fill an array
     cell with horizontal dots (e.g. for a full row).
     Note &ldots; places the dots on the baseline, while &cdots;
     places them higher up.
-->

<!ELEMENT ARRAY - - (ROW)+>
<!ATTLIST ARRAY
        align (top|middle|bottom) middle -- vertical alignment --
        coldef  CDATA   #IMPLIED  -- column alignment and separator --
        ldelim  NAMES   #IMPLIED  -- stretchy left delimiter --
        rdelim  NAMES   #IMPLIED  -- stretchy right delimiter --
        labels (labels) #IMPLIED  -- TeX's \bordermatrix style -->

<!ELEMENT ROW - O (ITEM)*>
<!ELEMENT ITEM - O (%formula)*>
<!ATTLIST ITEM
        align   CDATA  #IMPLIED  -- override coldef alignment --
        colspan NUMBER 1         -- merge columns as per TABLE --
        rowspan NUMBER 1         -- merge rows as per TABLE -->

<![ %HTML.Deprecated [
    <!ENTITY % head.content "TITLE & ISINDEX? & BASE? & STYLE?
                              & META* & LINK* & NEXTID?">
]]>

<!ENTITY % head.nextid "">

<!ENTITY % head.content "TITLE & ISINDEX? & BASE? & STYLE?
                              & META* & LINK*">

<!ELEMENT HEAD O O  (%head.content)>

<!ELEMENT LINK - O EMPTY>
<!ATTLIST LINK
        href %URI #REQUIRED
        %linkExtraAttributes; >

<!ELEMENT ISINDEX - O EMPTY>
<!ATTLIST ISINDEX
        href   %URI  #IMPLIED -- server handling queries --
        prompt CDATA #IMPLIED -- prompt message -->

<!--
    The BASE element gives the base URL for
    dereferencing relative URLs, e.g.

         <BASE href=3D"http://foo.com/images">
         ...
         <IMG SRC=3D"bar.gif">

    The image is deferenced to

         http://foo.com/images/bar.gif
-->

<!ELEMENT BASE - O EMPTY>
<!ATTLIST BASE
        id   ID    #IMPLIED
        href %URI; #REQUIRED
        >

<![ %HTML.Deprecated [
    <!ELEMENT NEXTID - O EMPTY>
    <!ATTLIST NEXTID N CDATA #REQUIRED>
]]>

<!ELEMENT META - O EMPTY    -- Generic Metainformation -->
<!ATTLIST META
        http-equiv  NAME    #IMPLIED  -- HTTP response header name  --
        name        NAME    #IMPLIED  -- metainformation name       --
        content     CDATA   #REQUIRED -- associated information     --
        >

<!--
    A style sheet can be associated with the document using the
    LINK element, e.g. <LINK rel=3Dstyle href=3D"housestyle.dsssl">.
    Style overrides can be placed in the document head using the
    STYLE element, e.g.

        <style notation=3Ddsssl-lite>
           dsss-lite stuff
        </style>

    Later on in the document you can use:

        <h2 class=3Dbigcaps>Header with bigger than normal capitals</h2>
        <p class=3Dabstract>A paragraph with a unique style of its own
        ...
    
    Statements in the given style notation 
        
    The tag names, class and id attributes are used in the style sheet
    notation to describe how to render matching elements.
-->

<!ENTITY % style-notations "dsssl-lite">

<!--  CL: I have temporarily commented out the whole DSSL-Lite stuff because
      I want to use this DTD to validate some HTML 3 documents and I get this
      error:

      sgmls: SGML error at
      /afs/mcc.ac.gb/users/cgu/cgu_sources/Dir9/html-check/lib/html-3.dtd,
      line 1259 in declaration parameter 4:
      Formal public identifier "-//ISO//DSSSL Lite ??//EN//" invalid; treated
      as informal
       
<!NOTATION dsssl-lite PUBLIC "-//ISO//DSSSL Lite ??//EN//">

CL: -->

<!ELEMENT STYLE - O (#PCDATA)>
<!ATTLIST STYLE
    notation NOTATION (%style-notations;) #REQUIRED
>

<!ENTITY % html.content "HEAD, BODY">

<!ELEMENT HTML O O  (%html.content)>
<!ENTITY % version.attr 'VERSION CDATA #FIXED "%HTML.Version;"'>

<!-- suggested roles are: TOC, DOC, DOCPART, HITLIST, DIALOG -->

<!ATTLIST HTML
        %version.attr;         -- report DTD version to application --
        urn  CDATA   #IMPLIED  -- universal resource name for this document --
        role NAMES   #IMPLIED  -- role of this document, eg table of contents --
        >

<!-- The END -->


