We study how adaptive hypertext may improve the utilization of large online documents. We put forth the inter-related concepts of shattered documents, and renoding:
splitting a document into components smaller than the page, called noogramicles, and
creating each page as a new assemblage of noogramicles each time it is accessed. The
adaptation comes from learning the navigation patterns of the usors (authors and
readers), and is manifested in the assemblage of pages. Another essential trait of our
work is the utilization of user simulation for testing our hypotheses. We have created
software simulators and conducted experiments with them to compare several adaptive
and non-adaptive configurations. Yet another important aspect of our work was the
study and adoption of the technique of spreading activation to explore the network
database of the learnt model of travels. We have realised a quantitative evaluation
based on utilization quality measures adapted to the problem: session size, session