Today I was up bright and early to attend this Archaeology ontology workshop, organised by staff and students from universities in the Midlands3Cities (M3C) Doctoral Training Partnership. As someone who studies how archaeologists work, rather than really being an archaeologist myself (albeit with some experience in this area), I wasn’t sure how useful it would be for me, but the programme looked interesting, and I’ve been keen for a while to see how ontologies have been adopted in Archaeology, and the effect they have on researchers’ work. This workshop certainly did not disappoint. I learned a huge amount about the fundamentals and rationale behind ontology development, found out about some projects where ontologies have been successfully implemented, and gained some practical experience of building ontologies myself in Protégé.
One talk I found particularly useful was presented by Emilio Tuosto and Alessandro Quercia, which combined information about why ontologies are a good approach to managing and structuring archaeological data, with practical information about how this technology has been applied to a collection of loom weights. Emilio’s sections especially outlined the specific advantages that ontologies have over relational databases, which articulated explicitly what I have been thinking for some time. These particular advantages include not being limited to a particular dataset from the outset, with the flexibility to add new concepts at any time; interoperability with other datasets without having to make significant structural changes, and the potential for a much greater degree of inferencing than is possible with relational databases. This presentation gave such a good case for structuring data in this way, and gave me ideas to incorporate into my own research.
Another talk I enjoyed was by Charlotte Tupman, who spoke first about the development of an ontology for the Sharing Ancient Wisdoms (SAWS) project, which was interesting as it emphasised the importance of starting from the fundamental definitions of terms to be used in an ontology, and reaching a consensus among all the collaborators about these definitions (which was more difficult than expected). Charlotte then went on to talk about the Pelagios and SNAP:DRGN projects, which I spoke about at my Digital Classicist seminar last week. I was particularly interested to gain new insights into SNAP:DRGN and its ‘cookbook’ that allows researchers to connect their own data to that held in SNAP. It was also really encouraging to see mockups of the proposed future search interface for the data – I hope they are successful in gaining funding to make this interface a reality as it would make such a difference to the usability and uptake of the resource.
In the afternoon, we had a practical workshop in one of the computer labs, where we followed a tutorial in building ontologies using Protégé. Although I had never used an ontology-building tool before, I could immediately see its value, as it made the process so easy and intuitive (although this might have been different if I was not already aware of ontology and Linked Data structures). I am not sure if my PhD will involve ontology creation, although I really like the idea – I guess this is the librarian in me!
One thing that did concern me, however, was the emphasis on building a new, project-specific ontology, rather than linking out to existing terms. Charlotte’s talk stood out in that she discussed the research that went into identifying a suitable ontology that would meet most of the terms required by the SAWS project, and how the team decided to use FRBRoo, which incorporates terms from both FRBR (a bibliographic ontology) and the CIDOC CRM (a cultural heritage ontology). Additional terms that could not be found in an existing ontology were then used to build the new SAWS ontology, which has since been adopted by later projects. I feel that this is what projects should be aiming for – reusing and connecting to what is out there, and ensuring that any new ontologies are interoperable, rather than reinventing the wheel. I can absolutely see the benefits of ontology building as an exercise, as it helps you to engage with your data on a much more meaningful level than you would normally, but I do feel that perhaps more emphasis could have been given to ensuring that the terms from this ontology are aligned with those of existing ontologies, to promote data interoperability.
That said, I really enjoyed this workshop, and think it was probably one of the most worthwhile events I have attended on my PhD so far. I gained far greater insights into concepts I had not previously been able to articulate, and found yet more convincing arguments for adopting a Linked Data approach to Humanities projects – it can mean a lot more work in the short term than using a familiar relational database structure, but the rewards are so much greater – deeper engagement with data, increased flexibility of queries, inferencing capabilities, flexibility and interoperability. This is absolutely an approach which should be much more widely used in Humanities disciplines, and I hope that the research area to which my PhD is contributing will facilitate engagement with these technologies. I would like to thank the staff and students from the University of Leicester and M3C for organising what was a truly stimulating and inspiring day.