Spatial Representation of Information
Charts and Graphs
In most cases, charts and graphs are pictorial representations of data stored in tables. As such, they can be stored on the Web as table data, and still be usable. Pie charts, for example, use table data and graphically point out which item has the biggest chunk of the total, by showing that chunk as a larger section of a "pie". This can be represented in a table simply by sorting the items in descending order of percentage of the total.
Bar charts do similar things and can be represented similarly.
Graphs that show how something varies over can be displayed in a table sorted by starting to ending time.
If any of the above are stored as images, then they would need to be described with appropriate descriptive information (see Implementing Rule A for more information on how to include this).
Maps
Maps pose different problems. They imply, in some cases, visual recognition of a geographic area for selection. This is not possible to render in tables. To convert maps into huge data tables for selection by the visually-challenged is to make them nearly useless.
New technology may solve that problem. NIST is working on a spatial map-generating device that will use "pins" to change the surface texture. Visually, this can be seen, and physically, this can be felt. Read the article from Federal Computer Week, NIST device lets visually impaired users 'feel' computer pictures and graphics.
Maps have one other important feature: the ability to determine relative location of two objects.
Cognitive Issues
The translation of a chart or graph or map into a Web page, while retaining accessibility, requires a careful analysis of what the user is supposed to do with the information. One of the most common maps used is a map of the United States, with each state designated as a clickable area. The purpose of this usage of a map is to select a state. There is no need for a two-dimensional full-color representation of the United States to do this. It is eye candy. The same result can be achieved by using a list of states, or a select box. If the map is a server-side map, this is required, anyway (see Implementing Rule E). For a client-side image map, if each clickable region is given a title containing the name of the state, the map would be accessible (see Implementing Rule F).
On the other hand, if maps are being used to display geographic topography, a three-dimensional model would be extremely useful - one could be able to feel mountains and valleys.
If a map is used to show routes to help identify how to get to some place, then an accessible alternative would be a list of directions from any given point on the map to any other point. This requires creating an alternative version of the map, and would be useful for sighted users, as well. Sites such as MapQuest already do this. (Note: I make no assertion as to the accessibility and usability of MapQuest.)
Using "D" Links
A "D" link can be created to direct the user to alternate descriptive information of charts and graphs. The "D" link can either go to an external file, or to another area on the same page. If the link is to an external source, it needs to be maintained concurrently with the page with the chart or graph. It might benefit both sighted and non-sighted users to have the descriptive information included along with the chart or graph. It might also be beneficial to provide skipovers to enable the user to skip reading the associated table data, if the description is sufficient. Plus, the designer can decide whether to include the information and allow the information to be visible or hidden.