caT: Adapting and updating the Trellis hypertext model

The Trellis project has investigated the structure and semantics of human-computer interaction in the context of hypertext systems. The Trellis model is defined formally using coloured timed Petri nets as the structure of a hyperprogram, and this gives the model an elegant structure that can be both programmed and analysed. In this paper, we review and reexamine the Trellis hypertext model and its implementations to identify ways in which the model can be applied in modern computing environments and to determine improvements that will enhance the usability of the implementations. We discuss issues with Trellis concerning the graphical complexity of specifications, the characteristics of mobile and Web-based environments, and the specifics of its prototypes' user interfaces and implementation platforms. Based on our analyses, we have enhanced the model and extended the earlier systems to produce a new prototype called caT (for context-aware Trellis) which provides a framework for authoring, browsing, and analysis of reasonably complex, dynamic hypertexts. We describe how caT addresses the issues identified in Trellis. We also describe the results of our small usability study of the caT system and discuss matters that we continue to investigate.