During the BHL Day 2026 a lightning talk was delivered by representatives of the University of Cambridge’s Darwin Correspondence Project. In that presentation the presenters mentioned that they had a database of people related to that project but that they had yet to create authority files or structured data about those people. I immediately thought of Wikidata and asked the question whether they had thought of using Wikidata as a question at the end of their presentation. Yes they said but that this was on their “to do” list.
As luck would have it @Pigsonthewing was watching these presentations online and left messages on Bluesky suggesting a Wikidata property be proposed for the Darwin Correspondence Project person identifier. Below is a link to the property proposal he has subsequently created!
Thanks for letting us know Rod! @Pigsonthewing - just to let you know I’ve emailed Liz Smith at the Cambridge University Library and Archives to tell her the property has been accepted.
Data (9144 rows of it) now loaded into Mix’n’Match - maybe I’m impatient, but I’m not seeing any automated or preliminary matches, after an hour or so.
It turns out that there are also a small number of entries for organisations, and publications.
Not sure a restriction to people is always best approach. For example, searching for Oliver Alexander Ainslie in Wikidata the first hit is a record in a database that looks like it is for the same person, but person themselves isn’t in Wikidata.
I didn’t realise that you could add parameters to the search in the actual search box. One reason this feels hacky is that if you find nothing Wikidata invites you to create an item whose name includes those parameters!
There are a couple of reasons for not having those parameters. For “Oliver Alexander Ainslie” I get no hits in Mix n Match, but if I trim the parameters I get:
Understood, but to a casual user clicking on individual records, finding Q98840143 might then be inspired to create a Wikidata item for Oliver Alexander Ainslie, using Men-at-the-Bar as a source.
Understood, but if you know that some entities aren’t people, wouldn’t make sense not to use the default Q5 for those records?
Cool!
Anyway @Pigsonthewing thanks for your patience with my ignorance, I was simply curious about the matches I was (or wasn’t) seeing by idly clicking on the Wikidata button.
if you know that some entities aren’t people, wouldn’t make sense not to use the default Q5 for those records?
Yes, if there was a 50/50 (or even, say 70/30) split. But when its <300 vs. >9K, it reduces the efficacy of <300 searches, while greatly increasing the efficacy of >8700. In my view (as the person setting up and wanting to use the catalogue) that’s a price worth paying.
