A Week in Williamsburg

The Computer Applications in Archaeology conference hopped across the pond again this year and was hosted in sunny Williamsburg, VA. I hadn’t been to Virginia before but was struck by its combination of charming woodland and swampy, well, swamps. Williamsburg is part of the ‘Historic Triangle‘ of Jamestown, Williamsburg and Yorktown, which were the first colony, first capital and last colony (insofar as it was where we Brits surrendered) of Virginia. Williamsburg itself is more or less a themepark (complete with period dress actors) and was heavily reconstructed by Rockefeller Jr. from the thirties onwards. It’s pretty though. Jamestown, on the other hand, is just archaeological remains on a marshy island, but has a couple of interesting museums that make a reasonable stab of portraying the conflicting goals and world views of both the original settlers and their Powhatan hosts.

Thanks for the postcard, I hear you cry, but what of the conference? Well, I’m happy to report that it was, for me at least, one of the most stimulating CAAs I can remember. Without intending to blow my own trumpet, that was at least in part due the success of the ‘Semantic Web: 2nd Generation Applications’ session I co-chaired. By a stroke of bad luck, my original partner in the venture, Tom Elliott, was unable to attend, but Keith May of the EH/Uni of Glamorgan STAR Project made an equally stalwart companion on the day.

We had a large number of submissions for the session, and in the end were asked to move some into a separate one on Data Management the following day but this may in the end have been to our benefit, with two opportunities for discussion. This was something of a milestone at CAA, which has seen a number of presentations on SemWeb aproaches in the past but never this number or all brought together. The end product was interesting not just for the sum of its parts but also in showing just how healthy the SemWeb community in Archaeology is.

This seems to be for two reasons: the first is that there is an increasing acceptance that these technologies are just a little too big for any one research group to manage. In turn they’ve increasingly started to focus on different aspects rather than complete systems. In communities where there is no need to co-operate with one another (for example those working in Network Analysis) it has seemed harder to get collaborative ventures going, so this may be a promising indicator. Only time will tell whether it’s true but there is certainly a lot of enthusiasm at present. The second factor is that both ‘Bottom-up’ and ‘Top-down’ approaches slowly seem to be converging. This is most apparent in the increased use of small sections of big ontologies (like the CIDOC CRM), as well as acceptance by the ‘just give your own stuff URIs’ community that stable, independent resources that provide canonical URIs for shared concepts are invaluable for linking data.

All the papers were interesting, but some highlights for me included:

– Sebastian Heath, who’s been one of the pioneers of making data available as URIs saying that, for him, the great moment is when you get ‘unexpected value’, i.e. when someone else uses your stuff in a way you had never even imagined. To me, that’s that’s the true spirit of scholarship.

– Ian Johnson‘s team at University of Sydney Archaeological Computing Lab have been working on system called Heurist that seems to be effectively triples-based. Even more exciting to me was the theoretical work done by Cathy Campbell on describing time periods. If they could create a GeoNames-like service out of it I’d want it in my stocking this Christmas.

– Kate Byrne reminding us that it is ‘better to be correct than complete’. I’m increasingly of the view that ‘correctness’ is over-rated too 😉 but it certainly is vital to recall that all Semantic data is inherently incomplete (as is all archaeological data). The less of it that is inaccurate however, the less fuzzy our aggregated view of the past will be.

– Robert Kummer‘s point that ‘users don’t want a record, they want to understand a historical topic’. If the SemWeb approach is going to work at all, we still need to do a lot of work in this area.

– Although it wasn’t actually in either of the sessions above, my vote for best idea of the conference goes to Julian Richards‘ suggestion that there should be an annual prize for best reuse of archaeological data. Currently the discipline still seems too often focussed on new results and this would be a great way to encourage both data contributors and and analysts to come back to the ‘data mountain’ we continue to accumulate.

I guess it’s not too vain to add that the importance of making it quick and easy (as in ‘one rainy afternoon’) for archaeologists to create Linked Data from their own datasets was the key take-home message of my presentation (whether any one wanted to take it home is for them to say :-D)

There was plenty more besides but we’re hoping to publish as many of the papers collectively as possible so updates will follow. Please also note that we intend to keep the conversation going over at the Graph of Ancient World Data (GAWD) Google Group so come on over and join us. Other than the SemWeb sessions, I was also very interested in developments over at Digital Antiquity, and of course ArchCamp 7 rocked the house as well. I also joined the CAA Steering Committee so as Student and Low Income rep so if you want any points raised, drop me a line.

Happy Easter all!