Try to visualize the whole World Wide Web at once. The first image that comes to mind is a globe covered with dots and lines.
But, of course, cyberspace doesn't work like that - physical location is irrelevant.
Why not construct a spatial image of websites that are located relative to the number of times the hyperlinks between sites are used? In this way, "clusters" of sites would develop around topics of interest to users of the Web.
But this shouldn't occur in a 3-D space. Instead, it will be a multidimensional space, using the statistical technique of multidimensional analysis for its construction.
It would be Gibson's cyberspace done right.
