http://web.media.mit.edu/~amber/cogmac/cogmac1.html
Software designed to allow the user to search intuitively through content of seven million text documents by reconfiguring histograms used by common search engines.
A. We begin with a histogram of words and their corresponding weights relative to a given set of documents. We order the histogram to form loosely defined concepts.
B. We wrap the histogram data around a point and draw bounding boxes to define salient concept areas. We then create a shape which becomes the documents profile showing comparing concept areas within the document.
C. The interface shows the document icons as they would be used to compare between 30-100 documents in a single frame.