How it came about
EngD students get a group project to complete in their first year. Naturally I had a number of ideas I wanted to pursue. I’ve always been interested in visualisation of very large data-sets, but that raised the question of what to visualise. I’ve also been interested in trying to use 3D more creatively as a part of that visualisation process.  Currently all the attempts at 3D visualisation come from the ‘its 3d it must be better’ school of programming. We where also interested in two technologies one called ‘adaptive’ rendering - altering what is seen to keep the frame rate a (near) constant and level of detail (more else where). The other is the three dimensional use of the wheel mouse ( or scroll mouse). You would expect it would be more natural to use a 3d program than a 2d or 1d ( text), after all we live in the real 3d world so it must be natural to occupy a computer version of one. In fact using 3D programs like C.A.D.D is highly unnatural and difficult. Perhaps one solution is to use the scroll wheel button to let the mouse become ‘three dimensional’ again. I’ve used a bat(three D mouse) on you lift and wave in the air and believe me you hand gets tried using it all day.
One conversation I had played with some of the notations of what a building did and software could not. This raised the concept of serendipity and the way buildings ‘generate’ accidents. My oft repeated example is using automation( via a computer normally) as a away of investigating an object or system. For example I once asked ‘what is a library for?’ ‘getting the books and journals you need came the authoritative reply’. ‘OK’ I say ‘lets automate that. You go in type the name of the book and it comes out of a slot’
‘Ah’ says the proponent ‘that wouldn’t work’
‘why not?’
‘well when I walk into a library I check the returns out, I look at the books around the one I am looking for.’
So in fact we can draw  the conclusion the purpose of a library is to  partly generate random interaction with books.
Saying ‘how would I automate this’ is just a way of doing what is known technically as a ‘social probe’.
Walking into the university for me keeps me bumping into people which generate interesting flows of information. The key observation was that of a library. The conversation turned to open and closed stack libraries.  Closed stack libraries ( like the British Library) you walk in and order the book you want and that’s what you get. Open stack libraries you get the books your self and you get to see the books near the ones your looking for. This can lead to wonderful accidents.
The software for UCL’s library is closed stack - it tells you about what you asked for and nothing else. We noted that on Amazon acts like a closed stack ( you ask for books it finds them ) and tries to act like a (open shelf) library or a book shop. It does this by presenting similar books to the one you take notice of. We called this ‘slit shelf’ ( between open and closed ). At first we like the notion of producing an interface to all of the UCL library. Trying to get the data would be FAR too difficult for a short project. A little research showed that Amazon had and a way to let developers access all the data (if a little slowly).So the project became to visualise all of Amazon (books).
This was good as Amazon is distinctly pre-Google. The primary web-page is stuffed with information noise. Each page is also overwhelms the reader with considerable excess and irrelevant information. From a design point of view it would be good to bring some clarity to the browsing process. What would you do?
Mapping Amazon
Point and scroll
Follow your desire but stay alert and get distracted.
Keep going and eventually you see the paper work start to appear.
Try using the interactive map here  [ index.php ]