Viewing the network structure

Navigating through a hyperdocument is like solving a maze. When you are in the labyrinth, it is very difficult to find the exit. However, when you have an overview of the labyrinth the correct path is easily found, unless the labyrinth is very large. Much the same is true for hyperdocuments: an overview of the overall structure of the network can be very helpful in navigating through the link structure.

Few hypertext systems offer a graphical view of the network. NoteCards, Intermedia and StorySpace are examples of such systems. If the hyperdocument is large, the overview will not fit on the screen. Some systems provide zoom-in and zoom-out operations and/or scrolling windows. However, scrolling windows are a nuisance, while an overview of a large network becomes cluttered and unreadable.

What is needed is a dynamically changing fish-eye view of the network. The details of the network are only important in the immediate neighborhood of the node that is currently displayed. The remainder of the network should be filtered in some way to hide the details. The lack of insight as to how the network can be filtered is the main reason why most systems do not provide a network fish-eye view. If the hyperdocument has a (partially) hierarchical structure, a filter is much easier to design. Such hierarchical structure can be provided by using composite objects, as they exist in the Tower Model, or by using contexts, as in HAM.