Edinburgh EH8 9XP Scotland

In recent years, the development of Virtual Reality (VR) as a means of visualisation has been extremely rapid. However, the use of VR has often been in isolation from the many different elements that people use in the real world to give a meaningful context to their interaction with the environment. Models of reality extend from the descriptive (written text), through to the highly abstract (the tube map of the London Underground), and on to the pseudo-real (VR). Each model provides opportunities (and impediments) to the effective and efficient interpretation of geographic space. Through the development of a Computer Aided Learning (CAL) package for mountain navigation, we demonstrate how context was provided both to, and within, a 3D VR world. With reference to navigation through real world spaces, we illustrate the importance of giving the user sufficient information so that they can interpret both where they are and where they are going. The CAL described in this paper uses a mixture of HTML, Virtual Reality Modelling Language (VRML) and Java in order to provide the multi-layered contexts that are required for successful navigation through a VR world. In the paper we explain how the system was developed, implemented and evaluated. We then discuss the difficulties posed in adding context to a VR landscape and examine different techniques for enhancing the navigability of the landscape