Back to RDA Plenary 17 main page
Registration
PARSEC: Vocabularies need to be trusted and usable
Monday 19 April, 22:00 UTC
The goal of this co-located event is to share the current status of the PARSEC project and hear from others in the RDA community on efforts to connect to researchers about data sharing, FAIR, and citation best practices.
Registration information to follow.
To maximize semantic interoperability of our data the use of shared or harmonized terminology is essential, particularly where the data comes from multiple communities and from many countries. The term ‘vocabulary’ is used to denote any semantic asset containing terms and information about those terms, including value sets (i.e., bag of terms, term list), controlled-vocabularies, glossaries, thesauri, and taxonomies, and potentially even concept maps, ontologies, and knowledge graphs. With the increasing uptake of the semantic web, vocabularies and other semantic resources are proliferating. Researchers are becoming confused as to which one they should use, and more importantly, which ones they can trust. And where it can be shown that a vocabulary does not exist, what is best practice in developing, maintaining and sustaining one.
Agenda:
Trust:
- (10 min) Simon Hodson: What is good practice in the maintenance, governance and sustainability of vocabularies?
- (10 min) Lesley Wyborn: the 4 key dimensions of quality in a vocabulary: scientific validity, technical structure, stewardship and access.
Repository Delivery:
- (12 min) Siddeswara Guru: Use of vocabularies in repositories
- (12 min) Margaret O’Brien: Repositories role in helping to select the most useful vocabularies.
Usable by Researchers:
- (12 min) Romain David: Researcher efforts to develop data dictionaries that align to vocabularies
- (12 min) Alison Specht: Systematic approach for researchers to create and use robust and community-standard vocabularies.
Report
The session, curated by the PARSEC project (www.parsecproject.org), a Belmont Forum funded project with funding partners in France, the USA, Brazil and Japan, and associates in Australia, had good attendance, and the presentations and notes are posted on Zenodo:
Wyborn, Lesley, Hodson, Simon, Guru, Siddeswara, O'Brien, Margaret, David, Romain, Specht, Alison, & Stall, Shelley. (2021, April). PARSEC: Vocabularies need to be trusted and usable. Zenodo. http://doi.org/10.5281/zenodo.4705129
The speakers came from largely earth and environmental science backgrounds, experiencing vocabularies as theoreticians, managers of repositories, or working with researchers. We hoped this would create a ‘reality check’ of how vocabularies are perceived across these spheres, as creating them without them being adopted or systematic, is worth little if they are not usable or communicable.
Discussion was wide ranging, from concerns about semantics related to vocabulary management to potential links with other RDA groups (such as the FAIRsharing WG). The importance of a well-founded community-accepted vocabulary that is well-curated was highlighted, and an equivalent of a core trust seal for repositories for vocabularies was suggested. Providing information about vocabularies as they are updated or replaced was considered vital and will enable the user to judge better which to use (assuming they ‘discover’ the vocabulary in the first place, another issue). An active relationship with the community of practice was also suggested as integral to the success and durability of a vocabulary, but this is not always easy to achieve. In the situation of the repositories, a new term may be hosted as an RDF concept if it does not already exist in a published vocabulary. The link with and role of ontologies was mooted by several as important and helpful, but not pursued further in the session.
The group has submitted a session to SciDataCon to continue the discussion.