The content-structure associations map together elements of the structure and elements of the abstract content. In a graph-based structure, one natural association is to map the content elements to the nodes of the graph. The r-model imposes no restriction on the kinds of mappings that are permissible. It may be useful to map a single content element to multiple locations in the structure, or to map multiple content elements to a single location. Separating the abstract contents, structure and the content-structure associations has the advantage that one can retain the structure while replacing the contents of a hyperdocument, for instance by a translation into a foreign language.
The button-structure associations map the structure's relationship and abstract buttons. A natural association in a graph-based structure is to map the abstract buttons to arcs in the graph. In a Petri net based hypertext model, such as Trellis, the abstract buttons are associated to transitions of the net. Again, a one-to-one mapping is not enforced, although it may be the most useful mapping.
The container-structure associations describe the association of the structure, or portions of the structure, to one or more abstract containers. One use of this association is to permit grouping of elements of the structure, which might in turn be displayed to the reader in a single physical window (like the way graphics are mixed with text in this course). Different kinds of composite displays would be represented as associations with different types of abstract containers. In general, the container-structure associations allow the partitioning of the subsequent display of the hypertext into one or more possibly overlapping pieces.