4. Conclusion

The rapid development of computer graphics technology over the last twenty years has not been ignored in archaeology, and especially not in buildings archaeology, where there has been involvement since the early 1980s, with the Bath Temple Complex project. However, because of the speed of development, archaeologists have until recently not begun to consider the implications and potential of their work. The initial response was to use three-dimensional modelling to produce bigger and better reconstruction drawings, and to use whatever technology was offered to them to achieve this. This was also in the interests of the computer graphics industry, who were seeking unusual research projects to test and illustrate their software's capabilities.

As the cost of hardware and software has come down, and it is now possible to have a reasonable modelling package running on a normal desktop (or even laptop) computer for a few thousand pounds, archaeologists are now able to take greater control over their work with three-dimensional modelling. Solid modelling has become accessible also because of the integration of solid modellers into existing CAD packages, providing a more friendly graphical user interface, rather than relying on the programming languages of dedicated solid modellers for entering data. For large-scale projects, however, it would still be necessary to use a specialised solid modelling package.

There is a use for perception modelling in archaeology, both in research and in the presentation of results within archaeology and to the public. It is not clear, however, that an attempt at photorealism (with all the extra costs that this involves) should be considered as the default option. Rather, I would argue for the use of simple data visualization for most research and presentation purposes.

An area which has only recently begun to be developed in archaeology is that of structural modelling. Because it is a new field, we do not as yet have the wealth of experience which we have in perception modelling to help us to develop a strategy for future research; the work is still very experimental, and no doubt some mistakes will be made, and many new avenues opened up. The benefit of perception modelling to date appears to have been largely in presenting results, with some direct benefit to research in aiding the conceptualisation of space. The benefits of structural modelling to research should be much greater.

Archaeology would not be alone in concentrating on solid modelling, as this would appear to be the focus of current research, especially in engineering. In developing our interest in solid modelling, however, we should not ignore the fact that the mainstream of three-dimensional modelling is beginning to incorporate object-oriented techniques. In the past, archaeology has been used as the guinea pig of software developers. Now that computer applications are more widely understood in archaeology, we should be in a position to suggest projects to developers which are of immediate relevance to our research interests.

Technical Notes

The illustrations used in this paper are all three-dimensional models, developed using Autodesk's AutoCAD (release 12). This is the design industry-standard CAD package, as well as being the standard package used in archaeology.

AutoCAD was designed as a two-dimensional design and draughting package, to which the ability to produce three-dimensional drawings has been added. Solid modelling is possible through the Advanced Modelling Extension (AME). This is a b-rep modeller with a CSG GUI. The drawing can be linked to a relational database through the Data Extension package. Some data can also be attached directly to entities using extended entity data. Release 13 of the package allows for object-oriented modelling.

The two main surface modellers on the market at the moment are AutoCAD and Bentley's MicroStation. One of the few absolute limitations of MicroStation is that it has a restricted number of a layers in a drawing (although this can be accommodated by combining sets of layers) which are numbered rather than given names; this makes it unsuitable for use with a layering convention.

Solid modellers have tended to work with geometric shapes, which are more cumbersome in the design process (and for the archaeological record) rather than with free-form shapes. Surface modellers normally implement free-form curves, using at least the Bezier algorithm, and increasingly allowing NURBS (Non-uniform rational B-splines). Where this is the case, there may be an argument in certain applications for using a surface rather than a solid modeller. However, the problem is with the authors of the software and market demand rather than being intrinsic to the modelling algorithms; there is no reason why these algorithms could not be implemented in a solid modeller (Rappoport et al. 1996, 19). For example, a NURBS extension is available for surface and solid modelling in AutoCAD (Rhino for AutoCAD).


© Internet Archaeology
Last updated: Thu May 1 1997