In other news, we spent last weekend up in York at the Computer Applications in Archaeology (UK chapter) conference. I remember cabbage and I several years ago talking about using a game engine to create virtual reconstructions of ancient buildings... it seems that several people are now actually doing this. I'm kind of interested in the potential for using 'authentic' reconstructions within an educational game context.
One of the other issues that frequently comes up when discussing virtual reconstructions is how to distinguish fact from varying levels of conjecture. For instance, you may have the lower level courses of a building still existing; you can make a good guess about the upper levels, but the roof may be pure guesswork. You could have a 'reality slider' that altered the transparency of different parts of the building according to how certain you were about them. So 100% reality would just show you what was still there, whereas 0% would give you a complete (if fanciful) representation of how the building may have looked. It may be useful to provide several alternative reconstructions to reinforce the idea that these are possibilities, not facts.
For almost as long as I can remember, people have been talking about IT-based on-site recording (getting your data straight into a database, rather than recording it on paper then having to enter it onto computer at a separate stage). It seems that the practical difficulties in doing this have yet to be resolved; cost, finding suitably rugged hardware, and user acceptability are all significant barriers. One of the guys from Oxford Archaeology gave an enthusiastic presentation looking at various hardware options, such as the laptop designed for the 'One Laptop per Child' programme (on the basis that it's designed to be rugged) and coming to the conclusion that the ideal on-site device does not have to be a 'computer' as such - they are now looking at the openmoko linux-based phone, and are planning to become distributors for the new version.