“Wikidata is just a matter of facts”
On February 13rd and 14th, I joined the Cost MOBILISE Wikidata Workshop to develop the data model for taxonomic and nomenclatural data in Wikidata.
Modelling taxonomies and people in Wikidata is hard. Partly because even within ones own community consensus can be hard to reach and in the (rare) cases where consensus is reached, it turns out to be rather difficult or even impossible to reach that consensus within the Wikidata community.
This workshop was primarily intended to explore the usefulness of Wikidata for modelling taxonomic and nomenclature data. The value as a public commons for data is clearly recognised, but some question on how taxonomy and people nomenclature on wikidata is being modelled and more so how difficult it is change those models.
Personally, I have never seen Wikidata as a single community. A community of communities, where communities can overlap in a more venn-diagramian way might be a more accurate description. However, “the” wikidata community does not always sees it that way, meaning that some views can’t make it to wikidata. At the same time, consensus within one community can be hard to achieve. If views within one community differ, how can one expect “the” wikidata community to reach a consensus.
As long as references to primary sources and appropriate qualifiers consistantlty maintain the provenance of concepts captured in Wikidata, IMHO there is room for different or even colliding views. However if you are relative new to the Wikidata game, getting your opposing/colliding view across to the other commmunities in your domain, can be hard, really really hard.
Recent developments with respect to Wikibase and the even more recent developments in WBStack, might be just the missing link here. Currently changing an existing model on a given topic is done property by property, where the disagreeing model is more a collection of views. Here Wikibase can be instrumental as a staging server. Instead of arguing for every single property, one could first setup a Wikibase and draft the overall model needed to capture the essence of a data source.
The workshop was one with a mix of plenary and parallel sesssions. The plenary sessions were used to introduce Wikidata, Wikidata workflows and Wikibase, and also to report back from the parallel sessions.
We also explored how to use lexemes in this overal picture. Daniel Mietchen who participated remotely, prepared a stub for a lexeme on the “Loxodonta african” as an example on how a lexeme could be included in a taxonomy model on wikidata
The parallel workshop were distributed across the themes modelling 1. Taxon and Taxon names, 2. Modelling people, 3. Discussing the big picture. For a detailed description on the workshop I refer to its etherpad.
It was a very interesting workshop, a lot of the complexity involved surfaced, but we only scratched the surface there. However, I think the potential of Wikidata and Wikibase as a public infrastructure for taxonomy data was definitely acknowledged. I was specifically pleased with two concluding remarks (quoted below), during the wrap-up, which I think nicely wraps up the total event.
