WikiCite 2016/Report/Group 1

Group 1: Modeling bibliographic source metadata

Lounge, 4:00 - 6:00 pm • Etherpad: Lounge

Goal

Discuss and draft data models to represent different types of sources as Wikidata items

Participants

1. Andrea Zanni (Wikisource) / Aubrey
2. Chiara Storti, (Wikimedia Italia, Rete bibliotecaria di Romagna e San Marino)
3. Chris Keene (Jisc)
4. Chris Wilkinson (eLife Sciences)
5. Erika Herzog (Wikimedia New York City)
6. Jens Nauber (Die Sächsische Landesbibliothek – Staats- und Universitätsbibliothek Dresden (SLUB) (Saxon State and University Library Dresden (SLUB)))
7. John Kaye (Jisc)
8. Katie Filbert (Wikimedia Deutschland, Wikidata)
9. Lambert Heller (Technische Informationsbibliothek (TIB) (German National Library of Science and Technology))
10. Luca Martinelli (Wikimedia Italia)
11. Lydia Pintscher (Wikimedia Deutschland, Wikidata)
12. Mairelys Lemus-Rojas (University of Miami Libraries)
13. Marielle Volz (Wikimedia Foundation) (attending remotely)
14. Markus Kaindl (Springer Nature)
15. Merrilee Proffitt (OCLC Research)
16. Nettie Lagace (National Information Standards Organization (NISO))
17. Rachael Lammey (Crossref)
18. Susanna Giaccai (Wikimedia Italia)
19. Tobias Schönberg (Wikidata Community)

Summary

The group conducted in-depth discussions of various approaches of modeling different types of sources in Wikidata. Examples included blog posts, newspaper articles, journals, and different kind of books. The approaches have different ramifications and concerns, including problems of automation, accessibility for new users, reusability, and required maintenance.

The group discussed previously existing modeling attempts used by the Wikidata community (for example, regarding books), and reviewed the list of properties for different types of documents (articles, books, etc.).

Choosing a consistent approach is important both to the Wikidata community and third party users. The discussion focused on many kinds of sources and important edge cases. The result of the discussion was a list of workable best practices.

Introduction

If we want to structure references and bibliographies over Wikipedia, we need a place where to store the metadata. Wikidata is, right now, the best candidate for that. This means that Wikidata needs a way to model different kind of documents (used as sources in Wikipedia) such as books, scientific articles, newspaper articles, videos, etc.

There have been several attempts over the years to produce viable "models" for books and articles on Wikidata, but unfortunately the inherent complexity of the task never allowed the community to develop simple, stable workflows: How to describe a book/article on Wikidata? How many items do we need to create? Which properties do we have to use?

Recommendations

The following recommendations were made for the data working group (Wikidata community and 3rd party users)

• Further discussion should use existing WikiProject infrastructure (most importantly WikiProject Books on Wikidata) and the WikiCite mailing list to keep the different parties connected

The following recommendations were made for the editing of the data:

• Adding information about the source as a set of statements as a reference to a Wikidata claim is good, but creating an item reduces redundancy. If a set of statements is used a number of times as a reference, an item should be created for it.
• Regarding books: it's important to make always an item for a specific edition (if we need it), and an item for the work only if it's necessary.

Appendix: workgroup notes

Raw notes from group 1.