Having examined some of the reasons for introducing a hypertext project, the groups' individual approaches are now briefly described. In addition to learning some specifics of HTML, each group was faced with many decisions about how the document would be organized and how the finished document could be read. [Horn1989] indicates that are are four key questions involved in the design of a hypertext document:
One key feature was true of the organization of all the groups' documents. Each architecture was presented as a unit with links connecting topics within an architecture (e.g., the type of knowledge representation) to a separate general discussion of the topic, independent of any single architecture. This dichotomy between architecture-specific and architecture-independent information facilitated both a high-level description of the particular topic and a low-level description of the way in which the topic was related to an architecture. For example, the PRODIGY architecture [Carbonell et al.1991] uses a knowledge representation scheme based on first-order, predicate logic. When browsing the PRODIGY description in any of the documents, this property is described in terms of how it is implemented in PRODIGY and some of the capabilities and limitations resulting from its use. Within this discussion, there is also a link to a node that presents a definition of first-order, predicate logic and its relation to cognitive architectures in general (normally including additional links to discussions of logic within architectures). This organization allows one familiar with logic to not be distracted by a definition in the middle of its presentation within PRODIGY while also allowing one unfamiliar with logic to seek out a definition and links to other architectures using such a representation.
Outside of this similar division, however, each group's document developed a unique organization and style, based on the ideas of the group members. In the following sections, these approaches are briefly outlined.