The problem with image maps
This document attempts to describe in detail the accessibility problems that image maps present as viewed by Vision Australia accessibility consultants (comment submitted for the 17 May 2007 Working Draft of WCAG 2.0).
If you would like background information
about what an image map is, the different types of image maps and how
they are implemented, please see an article such as Jim Thatcher’s: http://www.jimthatcher.com
The problem with image maps as discussed here is that for some user groups there is no method available to access the information conveyed by the image map regions. The information that needs to be conveyed includes:
The affected user groups considered in this document are users:
Issues with the identification of the presence of an image map region (item 1 above) can be just as hard as it is for visual users without particular visual presentation requirements, except where the user is browsing without images. In this circumstance, the difficulty is increased greatly. The current version of WCAG 2.0 (draft) does not prohibit the use of images as a means of presenting important text. Since the text within images can not be resized without loss of clarity (and therefore readability) and can not have the foreground and background colours changed, this places a requirement for these user groups to browse without images so that they can access the text alternatives for those images in the visual presentation format they require. See the section on User Group 1 for examples of the image map issues encountered by users browsing without images. For simplicity, no other alterations of visual presentation are included in the User Group 1 examples.
The ability to determine the result of selecting an image map region (item 2 above) requires access to the region’s text alternative. The browsers in popular use only provide visual access to the region’s text alternative through a “tooltip” where a box containing the alternative text is displayed for a short time in response to mouse-hover over a region. This presents some potential accessibility issues for users with low vision or keyboard users, including that the text is:
See the section on User Groups 2, 3 and 4 for examples of the above issues. For simplicity, the pages are rendered with images loaded in the examples for these user groups.
Note: only client-side image maps are discussed in this document.
Note: the rendering of image map regions discussed in this document is the browser-implemented “tooltip” and is not referring to a script-implemented show and hide of boxes containing text.
For each of the user groups described in the introduction, the following popular graphical browsers (and platform) were tested:
The examples presented in this document demonstrate that, of the tested popular graphical browsers above:
Currently WCAG 2 requires that users who need a different text size or colour combination to turn off images and access information through the associated text alternatives which can be rendered in the required visual presentation.
All of the user groups discussed in the subsequent sections and their accessibility issues with image maps are relevant here. This section is focussing on the particular issues raised by the need to access content with images off.
When images are not loaded, a user utilising a current graphical browser:
Current graphical browsers do not render region outlines, region text alternatives or provide an alternate method of accessing the target of the regions, when images are not loaded. In addition, some browsers do not allow keyboard or mouse interaction with the image map regions when images are not loaded.
Most of the browsers will display a tooltip containing the image map region text alternative when the region has mouse-over focus. This requires the user to search for the clickable areas because of the lack of visual cues normally provided by the image. The display of the tooltip is subject to the issues discussed in the user groups below and is displayed for a limited time.
None of the browsers display the image map region text alternative when the region has keyboard focus.
These results show that users who have particular visual presentation requirements for content in order to be able to access the content will have difficulties or be unable to access information conveyed by the image map regions. This can hinder the user’s navigation of the site, unless another accessible navigation method is provided.
Internet Explorer requires the user to uncheck ‘show pictures’ (an advanced menu setting) and check ‘Always expand ALT text for images’ to visually access the image’s text alternative. This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
The following figures (Figure 1, Figure 2
and Figure
3) show that these browser versions
do not display image map region outlines or region text alternatives
when images are not loaded. A region’s text alternative is displayed
for a limited time in response to mouse-over focus for a region but
it is not shown in response to keyboard focus. The region’s outline
is not displayed in response to mouse-over focus but is displayed when
it has keyboard focus.
A capture showing how the image map renders when images are not loaded. There is no visible indication that the image is actually an image map (other than what is suggested by the image’s text alternative in this particular example). The user has no visual cues as to where the image map regions are or what they target.
A capture showing how the image map renders when images are not loaded, with mouse focus over one of the regions. For users interacting with a mouse, they could mouse-over areas of the image to hunt for the regions and access the region’s target information by the text alternative information in the tooltip.
Figure 3 – keyboard focus user
A capture showing how the image map renders when images are not loaded, with keyboard focus on one of the regions. The outline of the region is shown but the region’s target is not displayed.
This browser requires the user to uncheck ‘Load images automatically’ in the Content area of the Options panel (from the Tools menu) to visually access the image’s text alternative. This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
The following figures (Figure 4, Figure 5
and Figure
6) show that this browser does
not display image map region outlines or region text alternatives when
images are not loaded. A region’s text alternative is displayed for
a limited time in response to mouse-over focus for a region but it is
not shown in response to keyboard focus. The region’s outline is not
displayed in response to mouse-over focus or keyboard focus.
A capture showing how the image map renders when images are not loaded. There is no visible indication that the image is actually an image map (other than what is suggested by the image’s text alternative in this particular example). The user has no visual cues as to where the image map regions are or what they target.
A capture showing how the image map renders when images are not loaded, with mouse focus over one of the regions. For users interacting with a mouse, they could mouse-over areas of the image to hunt for the regions and access the region’s target information by the text alternative information in the tooltip.
A capture showing how the image map renders when images are not loaded, with keyboard focus on one of the regions (the region for WA). The outline of the region is not shown and the region’s target is not displayed.
Opera has a menu option to control the display of images (under the View menu). This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
Opera displays the image area as per the size required to render the image’s text alternative (or the word “Image” if the image does not have a text alternative) rather than the size specified by the image’s width and height attributes. It is not possible to mouse-over any image map regions and the regions can not attain keyboard focus.
The following figure (Figure 7)
shows that this browser does not display image map region outlines or
region text alternatives when images are not loaded. The user is unable
to access the image map regions with a mouse or a keyboard.
Figure 7 – no focus (user or author mode)
A capture showing how the image map renders when images are not loaded in Opera 9. Opera displays the image area as per the size required to render the image’s text alternative. The image text alternative is rendered over other text in the page. The image map region text alternatives are not displayed.
This browser requires the user to select the type of site from the ‘Site Controls’ panel (from the Options area of the Tools menu) and then uncheck ‘Allow images to be displayed’. This browser version does not display the image’s ‘alt’ attribute value in the image’s place. This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
The following figures (Figure 8, Figure 9
and Figure
10) show that this browser does
not display the image’s text alternative, image map region outlines
or region text alternatives when images are not loaded. A region’s
text alternative is displayed for a limited time in response to mouse-over
focus for a region but it is not shown in response to keyboard focus.
The region’s outline is not displayed in response to mouse-over focus
or keyboard focus.
A capture showing how the image map renders when images are not loaded in Netscape 8.1. There is no visible indication that the image is actually an image map (even the image’s text alternative is not displayed). The user has no visual cues as to where the image map regions are or what they target.
A capture showing how the image map renders when images are not loaded in Netscape 8.1, with mouse focus over one of the regions. A tooltip is displayed for a short time when the mouse focus is over the image or an active region. For users interacting with a mouse, they could mouse-over areas of the image to hunt for the regions and access the region’s target information by the text alternative information in the tooltip.
A capture showing how the image map renders when images are not loaded, with keyboard focus on one of the regions (the region for WA). The outline of the region is not shown and the region’s target is not displayed.
This browser requires the user to uncheck ‘load images automatically’ (Content area of the Options panel from the Tools menu). This browser version does display the image’s ‘alt’ attribute value in the image’s place. This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
The following figures (Figure 11, Figure 12
and Figure
13) show that this browser does
not display image map region outlines or region text alternatives when
images are not loaded. A region’s text alternative is displayed for
a limited time in response to mouse-over focus for a region but it is
not shown in response to keyboard focus. The region’s outline is not
displayed in response to mouse-over focus or keyboard focus.
A capture showing how the image map renders when images are not loaded in Netscape 9.0. There is no visible indication that the image is actually an image map (other than what is suggested by the image’s text alternative in this particular example). The user has no visual cues as to where the image map regions are or what they target.
A capture showing how the image map renders when images are not loaded in Netscape 9.0, with mouse focus over one of the regions. A tooltip is displayed for a short time when the mouse focus is over the image or an active region. For users interacting with a mouse, they could mouse-over areas of the image to hunt for the regions and access the region’s target information by the text alternative information in the tooltip.
A capture showing how the image map renders when images are not loaded in Netscape 9.0, with keyboard focus on one of the regions (the region for WA). The outline of the region is not shown and the region’s target is not displayed.
This browser allows the user to prevent images from loading by unchecking ‘Display images when the page opens’ in the Appearance area of the Preferences panel. This browser does not appear to offer any settings that will render the location or target of the image map regions for this user group.
The following figures (Figure 14
and Figure
15) show that this browser does
not display the image’s text alternative, image map region outlines
or region text alternatives when images are not loaded. A region’s
text alternative is displayed for a limited time in response to mouse-over
focus for a region but the region’s outline is not displayed. This
browser does not allow the image map regions to attain keyboard focus.
A capture showing how the image map renders when images are not loaded in Safari (Mac). There is no visible indication that the image is actually an image map (even the image’s text alternative is not displayed). The user has no visual cues as to where the image map regions are or what they target.
A capture showing how the image map renders when images are not loaded in Safari (Mac). A tooltip is displayed for a short time when the mouse focus is over an active region. For users interacting with a mouse, they could mouse-over areas of the image to hunt for the regions and access the region’s target information by the text alternative information in the tooltip.
A user in this group would typically access the image map regions by shifting the mouse-focus to an area of the image that is of interest and select it by clicking the mouse button. Alternatively, if the image is not familiar the purpose of the regions can not be gleaned from the image itself, the user may shift the focus of the pointer to all areas sequentially in an attempt to determine the purpose of each area.
This is because the popular graphical browsers will display the area’s text alternative when the area has mouse focus.
The following examples capture the result for users requiring a large font size. A portion of a web page containing both an image map and the adjacent text content is captured to allow a comparison of the tooltip text with other text in the page.
When text size is increased though the browser settings, the text within the tooltip (when displayed) does not increase accordingly. This means that users who require a larger font size in order to be able to access the text content, will miss out on the information present in the region’s text alternative. This can hinder the user’s navigation of the site, unless another accessible navigation method is provided.
In this browser, the user has the option of ignoring the author-specified text size and/or changing the browser setting for text size.
The following figures (Figure 16, Figure 17 and Figure 18) show that the text within the tooltip (that is, the region’s text alternative shown in response to mouse hover) does not render at the same size as text within the page (whether the text size being applied is the author’s specification or the browser settings). The size of the text within the tooltip does not change.
Figure 16 – author’s text size
A capture showing the mouse pointer over an image map region in Internet Explorer 6 with Internet Explorer’s tooltip for the area displayed. The in-page text size rendered is the content author’s set size.
Figure 17 – browser default text size
A capture showing the mouse pointer over an image map region in Internet Explorer 6 with Internet Explorer’s tooltip for the area displayed. The in-page text size rendered is the browser setting at “medium” (that is, the text size specified by the author is being ignored). The text within the tooltip is the same size as in Figure 16.
Figure 18 – browser larger text size
A capture showing the mouse pointer over an image map region in Internet Explorer 6 with Internet Explorer’s tooltip for the area displayed. The in-page text size rendered is the browser setting at “largest” (that is, the text size specified by the author is being ignored and the browser setting has been changed). The text within the tooltip is the same size as in Figure 16.
In this browser, the user has the option of ignoring the author-specified text size and/or changing the browser setting for text size, as per IE 6.
In addition, the user has the option of using the “zoom” function to effectively magnify what would have rendered as a default (that is, in theory, all aspects of the page are magnified, not just text).
The same test conditions were undertaken as per the IE 6 example above and the same results were obtained. That is, the changes in text size (using the text-size features) had no effect on the text size displayed within the tooltip.
The following figures (Figure 19, Figure 20
and Figure
21) show that the tooltip (that
is, the region’s text alternative shown in response to mouse hover)
is only displayed at 100% zoom and is not displayed at all at 150% or
200% zoom. An additional figure has been included (Figure 7) which shows
the image’s text alternative (not an active region) displayed in the
tooltip. The text within the tooltip is the same size as that displayed
in Figure 4 even though the zoom setting is 200%. This suggests that
even if the the region text alternatives were displayed at higher zoom,
the size of the text in the tooltip would not be in line with the size
of the adjacent text.
Figure 19 – default zoom (100%)
A capture showing the mouse pointer over
an image map region in Internet Explorer 7 with Internet Explorer’s
tooltip for the area displayed. The in-page text size rendered is at
the browser’s zoom setting at 100% (the default).
A capture showing the mouse pointer over an image map region in Internet Explorer 7 but a tooltip is not displayed. The in-page text size rendered is at the browser’s zoom setting at 150%.
A capture showing the mouse pointer over an image map region in Internet Explorer 7 but a tooltip is not displayed. The in-page text size rendered is at the browser’s zoom setting at 200%.
Figure 22 – 200% zoom (image text alternative comparison)
A capture showing the mouse pointer over a non-map region of the image in Internet Explorer 7 where a tooltip is displayed for the image’s text alternative. The in-page text size rendered is at the browser’s zoom setting at 200%. The text in the tooltip is at the same size as the image map region text alternative in Figure 19.
In this graphical browser, the user has the option of changing the browser setting for text size.
The following figures (Figure 23
and Figure
24) show that the text within
the tooltip (mouse hover) does not render at the same size as text within
the page when the text size is changed by settings in the browser.
A capture showing the mouse pointer over an image map region in Firefox 2 with Firefox’s tooltip for the area displayed. The in-page text size rendered is the author’s set size.
A capture showing the mouse pointer over an image map region in Firefox 2 with Firefox’s tooltip for the area displayed. The in-page text size rendered is the browser setting at two increments larger than the default. The text within the tooltip is at the same size as in Figure 23.
In this graphical browser, the user has the choice of zooming the content. This typically results in a reflowing of the increased size of the content.
This browser does not show the region’s text alternative in the tooltip (mouse hover). By default it shows the value of the ‘title’ attribute, if one has been implemented, and the URL of the target. If a ‘title’ attribute value has not been provided then it only displays the URL of the target. I don’t believe it is possible to specify that it displays the ‘alt’ attribute value.
However, it is worthy to note that the
text within the tooltip that is displayed does not resize in line with
changes in the zoom setting.
Figure 25 – 200% zoom
A capture showing the mouse pointer over an image map region in Opera 9 with Opera’s tooltip for the area displayed. The in-page text size rendered is the browser setting at 200% zoom. The text within the tooltip is at the same size as the default (at 100% zoom) and does not display the region’s ‘alt’ attribute value.
In this graphical browser, the user has the option of changing the browser setting for text size.
The following figures (Figure 26
and Figure
27) show that the text within
the tooltip (mouse hover) does not render at the same size as text within
the page when the text size is changed by settings in the browser.
Figure 26 – browser default text size
A capture showing the mouse pointer over an image map region in Netscape (8.1 and 9.0b3) with Netscape’s tooltip for the area displayed. The in-page text size rendered is the browser text size setting at default.
Figure 27 – browser larger text size
A capture showing the mouse pointer over an image map region in Netscape (8.1 and 9.03b) with Netscape’s tooltip for the area displayed. The in-page text size rendered is the browser setting at two increments larger than the default. The text within the tooltip is the same size as in Figure 26.
Figure 28 – browser minimum text size
A capture showing the mouse pointer over an image map region in Netscape (8.1 and 9.03b) with Netscape’s tooltip for the area displayed. The in-page text size rendered is the browser setting at a minimum font size of “20”. The text within the tooltip is the same size as in Figure 26.
This browser allows the user to change the text size by selecting the icons for bigger or smaller text (these can be displayed by selecting Text Size from the View menu) or by [apple symbol] + or by setting the preference ‘never use font size smaller than’ and specifying a minimum font size.
Mac OS X also provides a “zoom” feature
that magnifies all content. In this case, of course, the tooltip is
also magnified. However, it appears keyboard users can not shift the
viewport without emulating the mouse through the keyboard.
Figure 29 – author’s text size
A capture showing the mouse pointer over an image map region in Safari with Safari’s tooltip for the region displayed. The in-page text size rendered is the author specified size.
Figure 30 – browser larger text size
A capture showing the mouse pointer over an image map region in Safari with Safari’s tooltip for the region displayed. The in-page text size rendered is the browser setting at three increments larger than the default. The text within the tooltip is not rendered in a larger size.
A user in this group would typically access the active regions by shifting the focus of the mouse-controlled pointer to an area of the image that is of interest (assuming the colours used allow the user to identify areas of interest) and select it by clicking the mouse button. Alternatively, if the image is not familiar, the user may shift the focus of the pointer to other areas sequentially in an attempt to determine the purpose of each area.
This is because the popular graphical browsers will display the area’s text alternative when the area has mouse focus. This is typically referred to as a “tooltip”.
The following examples show the results when the browser settings are set up for a particular combination of foreground and background colour.
When text and background colours are changed through the browser settings, the text within the tooltip (when displayed) is not rendered in the set colours. This means that users, who require a particular colour combination in order to be able to access the text content, will miss out on the information present in the region’s text alternative. This can hinder the user’s navigation of the site, unless another accessible navigation method is provided.
In this browser, the user has the option of ignoring the author-specified colours and changing the browser setting for text (including link text) and background colours. However, there appears to be a bug (at least in this combination of browser version and OS) that means the background colour is not rendered.
The following figure (Figure 31)
shows that changes in foreground and background colour in the browser
will not result in the tooltip text being rendered in the browser-specified
colours.
Figure 31 – Browser colour settings
A capture showing the mouse pointer over an image map region in Internet Explorer 6 with Internet Explorer’s tooltip for an area displayed. The in-page text colour combination is not rendered appropriately. The browser settings are for yellow text, orange links and black background. The text colours have been rendered appropriately but the background colour that is rendered is white (instead of the specified black). The text within the tooltip does not render in the foreground and background colours specified by the browser settings.
In this browser, the user has the option of ignoring the author-specified colours and changing the browser setting for text (including link text) and background colours. However, there appears to be a bug (at least in this combination of browser version and OS) that means the background colour is not rendered in some areas.
The following figure (Figure 32)
shows that changes in foreground and background colour in the browser
will not result in the tooltip text being rendered in the set colours.
Figure 32 – Browser colour settings
A capture showing the mouse pointer over an image map region in Internet Explorer 6 with Internet Explorer’s tooltip for an area displayed. The in-page text colour combination is not rendered appropriately. The browser settings are for yellow text, orange links and black background. The text colours have been rendered appropriately but the background colour that is rendered for some areas of the page is white. The text within the tooltip does not render in the foreground and background colours specified by the browser settings.
In this graphical browser, the user has the option of changing the browser setting for text (including link text) and background colours. This is done through the Content area of the Options panel (select the Colors button).
The following figure (Figure 33)
shows that the text within the tooltip (mouse hover) does not display
in the appropriate colours when the colours are specified by the browser.
Figure 33 – browser colour settings
A capture showing the mouse pointer over an image map region in Firefox 2 with Firefox’s tooltip for an area displayed. The in-page text colour combination rendered is as set by the browser (yellow text, orange links and black background). The text within the tooltip does not render in the foreground and background colours in the browser settings.
In this graphical browser, the user must choose the ‘User mode’ style for it to render colours other than the author’s specified colours.
This browser provides for setting the background colour, unvisited link and visited link colours (in the Web Pages tab of the Preferences panel).
This browser does not show the region’s
text alternative in the tooltip (that is, the region’s text alternative
shown in response to mouse hover). By default it shows the value of
the ‘title’ attribute, if one has been implemented, and the URL
of the target. If a ‘title’ attribute value has not been provided
then it only displays the URL of the target. I don’t believe it is
possible to specify that the tooltip displays the ‘alt’ attribute
value.
Figure 34 – Browser colour settings (in user mode)
A capture showing the mouse pointer over an image map region in Opera 9 with Opera’s tooltip for the area displayed. The in-page colours rendered are the background and link colours set in the browser (yellow text, orange links and black background). However, the text within the tooltip is not rendered in these colours.
In this graphical browser, the user has the option of changing the browser setting for text colour, link colour and background colour.
The setting is in the ‘Fonts & Colors’ panel, accessed from the General panel of the Options menu (under Tools). The user is required to uncheck ‘Use system colors’ and to select ‘Always use my colors’ for the browser set colours to be rendered instead of the author-specified colours.
The following figure (Figure 35)
shows that the text within the tooltip (mouse hover) will not render
in user-specified colours when specified through the browser.
Figure 35 – browser colour settings
A capture showing the mouse pointer over an image map region in Netscape (8.1) with Netscape’s tooltip for the area displayed. The in-page colours rendered are the background and link colours set in the browser (yellow text, orange links and black background). However, the text within the tooltip is rendered as black text on a white background.
In this graphical browser, the user has the option of changing the browser setting for text colour, link colour and background colour. The setting is in the Content tab of the Options menu (under Tools). The user is required to uncheck the ‘use windows colors’ and ‘Allow pages to choose their own colors, instead of my selections above’ checkboxes for the browser set colours to be rendered instead of the author-specified colours.
The following figure (Figure 36)
shows that the text within the tooltip (mouse hover) will not render
in user-specified colours when specified through the browser.
Figure 36 – browser colour settings
A capture showing the mouse pointer over an image map region in Netscape 9.0 with Netscape’s tooltip for the area displayed. The in-page colours rendered are the background and link colours set in the browser (yellow text, orange links and black background). However, the text within the tooltip is rendered as black text on a white background.
This browser does not provide user-friendly options for setting text, link and background colours. A user would most likely be required to write a style sheet for the browser to use.
Mac OS X provides extremely limited options for users requiring a different colour scheme.
For the above reasons, examples were not captured in this browser.
A user in this group would typically access the active regions by using the keyboard to shift the focus progressively through the interactive elements in the page. The shift of focus is typically achieved by using the keyboard’s ‘tab’ key. When the keyboard focus is on the active region of choice the user will select ‘enter’ to select this interactive element.
When a user must interact with content via the keyboard, the image map region’s tooltip is not displayed when the region has keyboard focus. This means that users who interact with the content by using a keyboard will miss out on the information present in the region’s text alternative. This can hinder the user’s navigation of the site, unless another accessible navigation method is provided.
Both versions of this browser render
a light grey (often dotted) outline to indicate the interactive content
in the page which has current keyboard focus. This outline is hard to
see for users with low vision and can effectively “disappear” when
the interactive content has colours that are dark or patterned. These
browsers does not appear to offer the option to show image map region
text alternatives (or otherwise access the regions) for this user group.
Figure 37
A capture showing keyboard focus on an image map region in Internet Explorer 6 with Internet Explorer’s outline displayed around the state of WA. The text alternative for the region is not displayed. The same result is obtained in IE 7.
This browser renders a black dotted outline
to indicate the interactive content in the page which has current keyboard
focus. This outline can be hard to see for users with low vision and
can effectively “disappear” when the interactive content has colours
that are dark or patterned. This browser does not appear to offer the
option to show image map region text alternatives (or otherwise access
the regions) for this user group.
Figure 38
A capture showing keyboard focus on an image map region in Firefox 2 with Firefox’s dotted outline displayed around the state of WA. The text alternative for the region is not displayed.
This browser renders a grey outline to
indicate regions in image maps (and a blue highlight for other interactive
content in the page) when it has current keyboard focus. This outline
can be hard to see for users with low vision and can effectively “disappear”
when the interactive content has colours that are dark or patterned.
This browser does not appear to offer the option to show image map region
text alternatives (or other access to the regions) for this user group.
Figure 39
A capture showing keyboard focus on an image map region in Opera 9 with Opera’s outline displayed around the state of WA. The text alternative for the region is not displayed.
Both versions of this browser render
a black dotted outline to indicate regions in image maps when it has
current keyboard focus. This outline can be hard to see for users with
low vision and can effectively “disappear” when the interactive
content has colours that are dark or patterned. This browser does not
appear to offer the option to show image map region text alternatives
(or other access to the regions) for this user group.
Figure 40
A capture showing keyboard focus on an image map region in Netscape (8.1 and 9.0) with Netscape’s outline displayed around the state of WA. The text alternative for the region is not displayed.
This browser does not allow image map
regions to attain keyboard focus.