VRML has a number of shortcomings with specific reference to archaeology — in particular, the inability to cope with cartographic projection systems and geo-referenced co-ordinate systems. VRML97 uses single precision (32 bit) floating point data and Cartesian local co-ordinate systems originating at 0, 0, 0. This limits the effectiveness of VRML for representing real world data, reducing resolution to 10-100 metres, and making the implementation of cartographic systems difficult. Fortunately, these problems have been addressed by the Geo VRML working group using Java extensions to standard VRML. It is now possible to use 64 bit precision co-ordinates, use earth-based co-ordinate systems, and define origins. A number of other enhancements improve support for terrain visualisation. Thus, Geo VRML export has now been integrated into the current release of ArcInfo, providing for a potentially very useful combination.
If virtual reality is to offer, in some circumstances, an alternative to Cartesian representations of geographic information there is a need to assimilate the varied research into geographic uses for virtual reality. Again there seems to be a process of re-inventing a wheel. Currently one of us is involved in the development of a direct interactive interface between GIS and VRML specific to one set of archaeological data, such that one might compare a landscape as presented in a GIS to one as modelled in virtual reality. In addition, this might be taken further by providing a dynamic link between other types of coverage — for example, reclassifying the virtual reality landscape to show a cumulative viewshed coverage or some other reclassification. Along similar lines we have begun to compare the experience of a virtual environment, given real geographic constraints (hence wandering at real speed and ideally at a speed dependent to some extent on topography), with that both of the real environment modelled and the environment presented through a GIS.
Further enhancements could be developed for specific archaeological data. It should be possible for users to interact with worlds changing weather, lighting, time or date, the position or types of objects. There should be straightforward and non-specific means for integrating objects in the world, linking both ways to databases and GIS systems. Extensions to VRML to perform these functions are all being developed but until we standardise and publicise them the effort of developing them will have to be repeated each time they are needed.
© Internet Archaeology
Last updated: Wed Jan 28 2004