Daniel-Mietchen
commented
7 years ago Here's what I came up with for now:
-
Target audience: users and operators of research-related databases
-
Title: Ten simple rules for integrating domain databases with Wikidata
-
Principles:
- Use persistent identifiers
- within a database
- when referring to databases or their contents
- also for metadata
- if that's not possible, think about using Wikidata IDs instead
- Version your data and metadata publicly
- Map your data model to existing controlled vocabularies (including Wikidata's)
- if that's not possible, think about using Wikidata property IDs instead
- Make your database usable for machines
- open API
- dumps
- open standards
- open licensing
- Put your metadata into the public domain
- Put your data into the public domain if possible
- Make a list of queries you could imagine
- so as to drive development
- Engage with the Wikidata community
=====
Share your curatorial workflows
Make use of Wikidata's infrastructure in your database and to engage with its community
- multilingual
- identifier hub
- cross-disciplinary
- queriable
- public domain
- integrated with Wikipedia et al.
- customizable frontends
Beware of vandalism and spam
Review the data model for Wikidata entities related to entities in your database
Use FLOSS-licensed software for your database
Sustainability
- many research-related databases go bust when project ends
- Wikidata is community-supported and thus sustainable as long as it is useful to the community
Push vs. pull
It's a wiki
Test it
Andra had started a Gdoc on this last year, but it hasn't gone anywhere yet.
Making thus a note to self to give it another go.
Perhaps half a day to get started, so giving it the 1day and hackathon labels for now.