> Is this related to server-side image maps?  Where the 
client 
> sends X/Y coordinates from an image and the server 
> determines what 'region' was clicked and presents the 
> appropriate page?
Yes - kind of nightmare to manage, because the client has 
to fetch a certain amount of "pages" randomly (8 pixels by 
8 pixels, for example), and the biggest problem is in fact 
to change the html code so that it can still work in the 
local repository..