The Dexter Hypertext Reference Model

The Dexter reference model resulted from workshops on hypertext that were organized by Jan Walker and John Leggett. They brought together a group of experienced hypertext system designers. The first workshop was held in October 1988, at the Dexter Inn in New Hampshire. These workshops were primarily funded by Digital Equipment Corporation and Texas A&M University.

The goal was to hold discussion that might lead to a reference hypertext system model. After several meetings, they successfully developed a reference model, which is called Dexter Hypertext Reference Model.

The Dexter Hypertext Reference Model is "an attempt to capture, both formally and informally, the important abstractions found in a wide range of existing hypertext systems and future hypertext systems" [Halasz and Schwartz, 1994].

The Dexter reference model can serves as a starting point for the design of hypertext system, as well as a principle basis for comparing system and for developing standards in interchange and interoperability. Instead of starting from scratch, a designer can start building a hypertext system by following the constructs shown in the Dexter model.

Many models have evolved to refine, formalize and find a common abstraction in hypertext systems. However, the Dexter model is the most complete and general formal model of hypertext systems. The model defines standard hypertext terminology and specifies common abstractions in different hypertext systems. It tries to get the important abstractions found in the wide range of existing and future hypertext systems.

The model is divided in three layers, as shown in the figure below:

Source: Halasz, F., Schwartz, M., The Dexter Hypertext Reference Model, NIST Hypertext Standardization Workshop, February 1990.

The Dexter model focuses on the storage layer. This layer describes how the hypertext components and links are connected together. The basic unit in this layer is component which is used as a container for data storage. The storage layer acts as a "database" that is composed of a hierarchy of data-containing components/nodes which are interconnected by links. The components contain the chunks of text, graphics, images, animations, etc. that form the basic content in the hypertext network.

The storage layer focuses on the mechanisms by which the components and links are "glued together" to form hypertext networks. The components are treated in this layer as generic containers of data. Each component contains sets of attributes, presentation specification, anchors, and content specification.

The within-component layer is responsible for the content selection of individual component through anchors and structure within the components of the hypertext network. The range of possible content/structure that can be included in a component is open-ended. Content selection means the selection of a portion of the content. If the content of a hypertext component is a file, the ability to locate and display only the second line of this file is an example of content selection.

The within component layer is purposefully not elaborated within the Dexter model. The Dexter model treats within-component structure as being outside of the hypertext model intrinsically. A mechanism for referring to locations or items within the content of an individual component is called anchoring.

The functionality of the layer is acting as tools for the user to access, view and manipulate the network structure, because the storage and within-component layers treat hypertext as an essentially passive data structure. The model doesn't say much in this layer because there are so many possibilities for handling user interface.

The Dexter model uses presentation specifications as the interface between the storage layer and the run-time layer. Presentation specifications contain instructions on how the component is displayed. The specifications are a mechanism by which information about how a component/network is to be presented to the user can be encoded into the hypertext network at the storage layer.