Mapping Cyberspace

"The satellite technology used to beam images from Iraq to America and on to London can be thought of as a vector. This technology could link almost any three such sites, and relay video and audio information of a certain quality along those points at a given speed and at a certain cost. It could just as easily link Beijing to Berlin and Sydney, or quite a few other combinations of points. Yet in each case the speed of transmission and its quality would be exactly the same. This is the sense in which any particular media technology can be thought of as a vector. Media vectors have fixed properties, like the length of the line in the geometric concept of the vector. Yet that vector has no necessary position: it can link almost any points together."

'Virtual Geography: Living with Global Media Events McKenzie Wark

"Investigating structure, we discover isomorphism. In isomorphism we find different reflections of a deep structure. We find profound meaning when these isomorphisms reveal apects of the very deepest level of structure, that is, Brahman. Discovering such isomorphisms, these structures perform the role of mantras, mirroring the fundamental vibrations of the universe."

Steven R. Holtzman Digital Mantras: The Languages of Abstract and Virtual Worlds

In attempting to map cyberspace, a form must be created. Rather than impose a form on this space, it would be more desirable if the space itself could present itself in a self-organised fashion. This would best reflect the (hyper)reality, which is mutable, fluid, amorphous. Individual users could organise their own personal maps according to their needs and desires. Cyberspace is not one place, but many different things to many people.

The concept of the vector is a compelling one, a straight line in space with direction and magnitude (but no fixed position), it could be seen as a metaphor for a hyperleap from one site to another, and the thread which connects these sites together in their placeless space.

The vector, however, whilst having the property of multidimensionality, cannot easily accommodate the idea of the 'relatedness' of things, of the need to organise words, ideas, documents into meaningful groups or orders. For this we need the concept of the 'cluster'. Clustering techniques are used in information retrieval systems. A variety known as 'hierarchic agglomerative clustering' allows the 'document space' to impose its own form on the clusters. In terms of the web, a user could be asked to classify a particular document, or to assign it to a particular cluster. Documents could then exist in clouds of clusters interconnected by threadlike vectors.