Internet Archaeol 15. Goodrick and Earl. Virtual reality development

3.3 Virtual reality development

No single virtual reality development package will meet all of our virtual reality development needs. The only package that approached this was Cosmo Worlds, unfortunately discontinued when SGI changed direction and abandoned development of VRML in 1997. The VRML export facilities in most CAD and modelling programs give greater flexibility than VRML-specific programs but produce very inefficient code, making poor use of the scene graph for instance. Conversely, the VRML development tools tend to assume rather limited modelling requirements in terms of both size and complexity of models required. It is therefore necessary to utilise a range of products to develop models, integrate existing data, and extensively optimise the finished model. If virtual archaeology is to be widely implemented the discipline requires a collated corpus of information on the availability and implementation of tools, most of which are free or of little cost, and to provide a forum for discussion. Again, the CVRO may offer a potential focus for this, coupled with other fora such as Arch3D.net or Vista. The emerging best practices would allow for the rapid development of virtual reality models by new users, avoiding extensive research into the availability of techniques and the inevitable wasted effort when it is discovered that an inappropriate choice has been made. Whatever the detail, those working with virtual archaeology need the kinds of centralised repositories described not only to discuss the technical aspects but, even more importantly, to open up our own endeavours to the critique of the wider community. This is perhaps not so much a fault of access to computing resources and so on, but to the way we as computing experts discuss and continue our work.

Similarly, many projects are working on methods for presenting uncertainty — frequently derived from those standards for archaeological and other illustrators (for example, the Association of Archaeological Illustrators and Surveyors) — but a centralised repository for these techniques would be extremely beneficial. This area also needs further thought and critique. Clearly there is much to be gained from new methodologies and examples of their application but much research outlined in the literature reiterates the need for authentication, untheorised as a given (see Frischer et al. 2000 and Frischer et al. 2001, 4). There is thus much work on making our models more realistic, in other words like the real, the original past, or more closely tied to archaeological fact, as if composed of tangible entities easily linked together within understood process. One interpretation might be that authentication has become so entrenched in the reconstruction discipline that the few critiques of its theoretical validity have been ignored. Wheatley and Gillings have been amongst a small group suggesting caution (Gillings 2000; Wheatley 2000). The thought underlying the methodological solutions is that the more convincing a model the more powerful (for which one may read useful) and potentially dangerous it will be.

The discursive language underpinning this methodological authentication of reconstruction is not surprising given its roots in the complex processual computing/archaeology interaction of the 1980s, in which the reconstruction graphics could be seen to approach a true representation of the archaeological past studied, rather than a regime of truth. It also fits within the environment of many research projects, tending to adopt a blue sky approach which is infrequently explored in further detail. If such a conclusion is considered harsh then it is perhaps so because the discipline has failed to collaborate on the scale required for rapid, useful adoption, critique and redevelopment of methodological approaches to authenticity. In fact, each new conference produces a raft of new methodologies and we must hope that the advent of the CVRO may direct these energies or focus their output.

In addition, there is a range of projects considering methods for integrating data through virtual reality or other types of visualisation interfaces. Again these need to be shared and, perhaps, standardised. Even more necessary is a detailed exploration of the usefulness of such approaches. We have talked about recontextualising data but how useful is this? If we are working towards different kinds of post-excavation tools, where one might see an excavated context — its volume, colour, texture even (with haptics) — and have artefact avatars representing the objects recovered (able to describe their appearance, their related types, other sources of information and so on) we need to know who is going to use it. How does this approach fit with the established post-excavation techniques? Is it more of a presentation tool? Through discussions with other researchers it has already become apparent that the very idea of recontextualising artefacts is problematic. Briefly, one might image that to recontextualise an artefact based on the excavation and context records serves only to widen the divide between the initial context and the record. Perhaps this has more to do with authenticity and archaeology than the specifics of archaeological visualisation, but requires investigation and discussion from both ends of the technical spectrum.

If we produce a visualisation incorporating missing data and then add a reconstruction based on excavated fragments — for example a mosaic — we stray into a further problematic. One mosaics expert consulted over a proposed project was horrified at the idea of recontextualising any form of art, partly on the grounds that, as western consumers of 'art', placing archaeological materials such as wall-paintings or mosaics, or indeed works of art in the modern sense, in authentic reconstructed environments was to obscure our different way of perceiving. Secondly it was suggested that this promotes the idea of authentic reconstructed environments allowing us to retrieve in some way the perception of the inhabitants of the past — something not suggested in the virtual reality or visualisation literature. On the other hand, colleagues working with gallery paintings have received the opposite reaction — namely that by recontextualising and authentically illuminating a computer environment the interpretation is directly benefited, with some of the choices in colour and composition making sense in some way due to this process (Chalmers pers. comm.)