RFC 1980
This document is obsolete. Please refer to RFC 2854.

Network Working Group                                         J. Seidman
Request for Comments: 1980                                Spyglass, Inc.
Category: Informational                                      August 1996

         A Proposed Extension to HTML : Client-Side Image Maps

Status of this Memo

   This memo provides information for the Internet community.  This memo
   does not specify an Internet standard of any kind.  Distribution of
   this memo is unlimited.


   The markup language known as "HTML/2.0" provides for image maps.
   Image maps are document elements which allow clicking different areas
   of an image to reference different network resources, as specified by
   Uniform Identifier (URIs).  The image map capability in HTML/2.0 is
   limited in several ways, such as the restriction that it only works
   with documents served via the "HTTP" protocol, and the lack of a
   viable fallback for users of text-only browsers.  This document
   specifies an extension to the HTML language, referred to as "Client-
   Side Image Maps," which resolves these limitations.

Table of Contents

   1.  Introduction ...............................................  1
       1.1  Purpose ...............................................  1
       1.2  Overall Operation .....................................  2
   2.  Client-Side Image Map Extension ............................  2
       2.1  Syntax ................................................  2
       2.2  Required Changes to HTML/2.0 DTD ......................  4
       2.3  Backwards Compatibility ...............................  5
       2.4  Examples ..............................................  5
   3.  Security Considerations ....................................  6
   4.  References .................................................  6
   5.  Author's Address ...........................................  7

1. Introduction

1.1  Purpose

   Image maps are an important feature of the point-and-click interface
   which makes the World Wide Web so popular. The most common use of
   image maps is to allow users to access different documents by
   clicking different areas in an image.

Seidman                      Informational                      [Page 1]

RFC 1980                 Client-Side Image Maps              August 1996

   There are several limitations of the current image map implementation
   as it applies to this use. First, it only works over the HTTP
   protocol, making it unusable for reading local files or files
   accessed via alternate protocols. Second, a server transaction is
   required merely to determine where the link is directed. This can
   degrade performance noticeably when accessing distant sites. Third,
   unlike for normal links, there is no way for a browser to provide
   visual feedback to the user showing where a portion of an image map
   leads before the user actually clicks it. Lastly, the method for
   specifying the active regions of image maps is server-dependent,
   compromising portability of documents.  This extension to support
   client-side image maps addresses these issues.

   It is proposed that this extension be included in a future revision
   of the HTML specification.

1.2  Overall Operation

   Client-side image maps work by placing a complete representation of
   the active areas of an image, including their shape, size, and
   destination (URI), into an SGML-compliant textual form.  This markup
   may also optionally include a textual description for each area for
   display on non-textual browsers.  This representation, or "map," is
   given a name to identify it.

   When an image is included in an HTML document, it may include an
   attribute specifying a map to use.  The map may be contained in the
   same file which references the image, but this it not required.  If
   the map is in a different file, a URI to that file must be provided.

   The browser will parse the map and remember the contents.  When the
   user clicks the map, the browser will match up the location with the
   specified destination for that location and access that URI.  In the
   case of a non-graphical browser, the browser could display the
   textual descriptions for each area instead of the image.  Clicking a
   given textual description would then go to the associated

2. Client-Side Image Map Extension

2.1 Syntax

   Adding a USEMAP attribute to an IMG element indicates that it is a
   client-side image map.  The USEMAP attribute can be used with the
   ISMAP attribute to indicate that the image can be processed as either
   a client-side or server-side image map.  The argument to USEMAP
   specifies which map to use with the image, by specifying the URI for
   the file containing the map, followed by a '#', followed by the name

Seidman                      Informational                      [Page 2]

RFC 1980                 Client-Side Image Maps              August 1996

   of the map.  If the argument to USEMAP starts with a '#', the map is
   assumed to be in the same document as the IMG tag.  The presence of a
   USEMAP attribute overrides the effect of an enclosing anchor (A)

   The different regions of the image are described using a MAP element.
   The map describes each region in the image and indicates where it
   links to. The basic format for the MAP element is as follows:

   <MAP NAME="name">
   <AREA [SHAPE="shape"] COORDS="x,y,..." [HREF="reference"]
         [NOHREF] [ALT="alt"]>

   The NAME attribute specifies the name of the map so that it can be
   referenced by an IMG element.  Each AREA element contained inside the
   map element specifies a single clickable area of the image.  The
   SHAPE attribute gives the shape of this area.  Possible shapes are
   "RECT", "CIRCLE", and "POLYGON", which specify rectangular, circular,
   and polygonal regions respectively.  If the SHAPE tag is omitted,
   SHAPE="RECT" is assumed.

   The COORDS tag describes the position of an area, using image pixels
   as the units with the origin at the upper-left corner of the image.
   For a rectangle, the coordinates are given as
   "left,top,right,bottom".  The rectangular region defined includes the
   lower-right corner specified, i.e. to specify the entire area of a
   100x100 image, the coordinates would be "0,0,99,99".

   For a circular region, the coordinates are given as
   "center_x,center_y,radius", specifying the center and radius of the
   ircle.  All points up to and including those at a distance of
   "radius" points from the center are included.  For example, the
   coordinates "4,4,2" would specify a circle which included the
   coordinates (2,4) (6,4) (4,2) and (4,6).

   For a polygonal region, the coordinates specify successive vertices
   of the region in the format "x1,y1,x2,y2,...,xn,yn".  If the first
   and last coordinates are not the same then a segment is inferred to
   close the polygon.  The region includes the boundary lines of the
   polygon.  For example, "20,20,30,40,10,40" would specify a triangle
   with vertices at (20,20) (30,40) and (10,40).  No explicit limit is
   placed on the number of vertices, but a practical limit is imposed by
   the fact that HTML limits an attribute value to 1024 characters.

   The NOHREF attribute indicates that clicks in this region should
   perform no action.  An HREF attribute specifies where a click in that
   area should lead.  A relative anchor specification will be expanded

Seidman                      Informational                      [Page 3]

RFC 1980                 Client-Side Image Maps              August 1996

   using the URI of the map description as a base, rather than using the
   URI of the document from which the map description is referenced. If
   a BASE tag is present in the document containing the map description,
   that URI will be used as the base.

   An arbitrary number of AREA tags may be specified.  If two areas
   intersect, the one which appears first in the map definition takes
   precedence in the overlapping region.  Multiple areas may share the
   same destination to create composite shapes.  Any portion of an image
   which is not described by an AREA tag defaults to having no action.

   The ALT attribute specifies optional text which describes a given
   area.  A text-only browser can display the textual contents for each
   area as a substitute for the image.

2.2  Required Changes to HTML/2.0 DTD

   The required changes to the HTML/2.0 DTD to support this syntax would
   be as follows:

  Change the IMG element definition to be:
            ALT CDATA #IMPLIED
            ALIGN (top|middle|bottom) #IMPLIED
            ISMAP (ISMAP) #IMPLIED
            USEMAP %URI; #IMPLIED
            %SDAPREF; "<Fig><?SDATrans Img: #AttList>#AttVal(Alt)</Fig>"

  Add the following new definitions:
    <!ELEMENT MAP    - - +(AREA)>
            NAME %linkName; #REQUIRED

            HREF %URI; #IMPLIED
            ALT CDATA #IMPLIED

Seidman                      Informational                      [Page 4]

RFC 1980                 Client-Side Image Maps              August 1996

2.3  Backwards Compatibility

   This extension is specifically designed to provide a variety of
   fallback options for browsers which do not support it.  These options
   are based on the assumption that browsers will ignore any attributes
   or elements which are not present in the HTML/2.0 DTD.

   An document can be written so that a client-side image map can have
   three different fallback behaviors.  First, the document can use the
   server-side image map capability, by specifying the ISMAP attribute
   as well as USEMAP.  In situations where this is possible, the image
   map will work whether or not the browser supports the client-side

   Second, clicking the image can direct the user to a single URI,
   regardless of where on the image he clicks.  This is accomplished by
   placing the image inside an anchor (A) element.  The fallback
   destination could provide the user with an error or a textual list of

   Lastly, the image can appear to not be a link at all (i.e. missing
   whatever visual cues a browser provides to indicate a hyperlink).
   This will be the result if the image element neither contains an
   ISMAP attribute nor is inside an anchor.

2.4  Examples

   The following three examples show markup demonstrating the three
   fallback mechanisms described in section 2.3:

      This image map will work with any graphical browser:
      <A HREF="/cgi-bin/imagemap/pic1">
      <IMG SRC="pic1.jpg" USEMAP="maps.html#map1" ISMAP></A>

      Clicking here will take you to a page with an error message if
      you don't have client-side image map support:
      <A HREF="no_csim.html">
      <IMG SRC="pic2.jpg" USEMAP="maps.html#map2"></A>

      You can only click here if your browser supports client-side
      image maps: <IMG SRC="pic3.jpg" USEMAP="maps.html#map3">

Seidman                      Informational                      [Page 5]

RFC 1980                 Client-Side Image Maps              August 1996

   The following example shows the use of a map in the same file as the

      <IMG SRC="picture.jpg" USEMAP="#mymap">

   The following example defines a simple map which describes an image
   with a circle in the middle overlapping two large rectangles:

      <MAP NAME="welcomemap">
      <AREA SHAPE=CIRCLE COORDS="50,50,40" HREF="about_us.html"
            ALT="About our company">
      <AREA SHAPE=RECT COORDS="0,0,100,50" HREF="products.html"
            ALT="Our products">
      <AREA SHAPE=RECT COORDS="0,51,100,100 HREF="technology.html"
            ALT="Technology for the next century">

3. Security Considerations

   Clicking a portion of a client-side image map may cause a URI to be
   dereferenced.  In this case, the security considerations related to
   URLs [5] apply.

4. References

   [1] Berners-Lee, T., and D. Connolly, "HyperText Markup Language
       Specification - 2.0", RFC 1866, November 1995.

   [2] Seidman, J., "An HTML Extension to Support Client-Side Image
       Maps", The Second Internation WWW Conference '94 Advance
       Proceedings, pp 927-930.

   [3] "Standard Generalized Markup Language"  ISO Standard 8879:1986
       Information Processing Text and Office Systems.

   [4] Berners-Lee, T., Fielding, R., and H. Frystyk Nielsen,
       "Hypertext Transfer Protocol -- HTTP/1.0", Work in

   [5] Berners-Lee, T., Masinter, L., and M. McCahill, "Uniform
       Resource Locators (URL)", RFC 1738, December 1994.

Seidman                      Informational                      [Page 6]

RFC 1980                 Client-Side Image Maps              August 1996

5. Author's Address

James L. Seidman
Senior Software Engineer
Spyglass, Inc.
1230 East Diehl Road
Naperville, IL 60563

EMail: jim@spyglass.com

Seidman                      Informational                      [Page 7]