The question of why is the ARIADNE Portal using the Getty Art & Architecture Thesaurus (AAT) as its main vocabulary and not Wikidata came up as a topic of discussion recently in the ARIADNEplus project. Given that there are pros and cons with both, this is a topic worth exploring. In fact, the Getty AAT was adopted by the original ARIADNE project for the following reasons:
- the AAT is a widely used international standard with a broad coverage of different areas in archaeology,
- it has a freely available linked data implementation and API, maintained and promoted by Getty who have developed AAT as an intellectual resource with indexing/search in mind.
Consequently, an ARIADNE linked data resource with an AAT URI will potentially connect to many other resources with that URI.
Since the original ARIADNE metadata model has been updated to the AO-CAT in the present ARIADNEplus project and now uses a “spine” of (mandatory) high-level terms to classify all data items for which the AAT is still considered the best choice. This may also be referred to as a backbone and to which more detailed metadata can be added to facilitate better search results.
However, Wikidata has its merits too…
- It is open, general and flexible.
- Local vocabularies may be directly ingested into Wikidata and mapped to standard terms.
- Vocabularies made available via Wikidata enable exposure to all Wikimedia services.
- Wikidata offers a lot of existing tools for mappings, processing, data cleansing and so on.
- Wikidata is more wide-ranging and so may allow more precise mappings.
The disadvantages are that the scope of Wikidata is much wider and more general (since it was developed for all Wikimedia projects) and results can be random. It is also subject to change – not a desirable property if consistent terminology is required. Furthermore, Wikidata was never meant to be a primary source. As they state themselves: Wikidata is a secondary database and (ideally) statements do have references and connections to other databases are provided as well (https://www.wikidata.org/wiki/Wikidata:Introduction).
The Getty AAT was especially developed by people with a cultural heritage/humanities/arts background specifically for this field. Though the terms have a longer curation process, this results in a high quality vocabulary with good descriptions. Whilst it is not necessarily complete, it is possible to propose extensions to the Getty AAT vocabulary, and contributions from established research projects such as ARIADNEplus are welcomed. If terms are compared between the two, then Wikidata is likely to offer a wider, eclectic selection including those from outside of the CH/humanities/arts domain.
However, this does not mean that Wikidata (or other vocabularies) can’t be used for metadata included in the Portal – Wikidata (and other) links can be included in the indexing. Furthermore, Wikidata and the Getty AAT are connected via the Wikidata Property ‘Getty AAT ID’ (https://www.wikidata.org/wiki/Property:P1014), which links existing Wikidata items to its corresponding concept in the AAT.
The ARIADNE Portal will include search options on AAT concepts AND also search on all partner indexing terms. The AAT’s role is to allow the interlinking of different partner terminologies, i.e. Findable and Interoperable in terms of FAIR data management. Finally, as part of implementing a multilingual search functionality, work is in progress to include the relevant Wikidata language variants in the Portal in addition to any translations provided by the partners to existing AAT terms.
Getty Art & Architecture Thesaurus (AAT) – https://www.getty.edu/research/tools/vocabularies/aat/about.html
Wikidata – https://www.wikidata.org/wiki/Wikidata:Introduction
With thanks to the contributors to the discussion this article is based upon: David Novák (CAS), Bogdan Sandric (INP), Kate Fernie (CARARE), Julian Richards (ADS), Martin Doerr (Forth), Doug Tudhope (USW), Martina Trognitz (OEAW) and Peter Yanase (NARA).