Rod Page
@rdmpage.bsky.social
Expat Kiwi, Professor of Taxonomy at Glasgow University, inclined to say that something sucks at every available opportunity. Biodiversity informatics, phylogeny, knowledge graphs. I also run the @evoldir.bsky.social bot.
Reposted by Rod Page
A shorter snout is one of several physical traits linked to domestication. Using almost 20,000 raccoon photos from iNaturalist, researchers found that urban raccoons' snouts were 3.5% shorter than rural ones.
November 17, 2025 at 9:30 PM
A shorter snout is one of several physical traits linked to domestication. Using almost 20,000 raccoon photos from iNaturalist, researchers found that urban raccoons' snouts were 3.5% shorter than rural ones.
I guess one question is what is the entity? If it is a name, then string matching is all you need. If it’s a taxon then things get more complicated.
November 16, 2025 at 10:27 PM
I guess one question is what is the entity? If it is a name, then string matching is all you need. If it’s a taxon then things get more complicated.
Sounds a bit like doi.org/10.7717/peer..., but if I understand you correctly what you are after is not “here are some names and identifiers in different databases, match them” but “I have some names and some context, match them to a gold standard list of names using a model”?
20 GB in 10 minutes: a case for linking major biodiversity databases using an open socio-technical infrastructure and a pragmatic, cross-institutional collaboration
Biodiversity information is made available through numerous databases that each have their own data models, web services, and data types. Combining data across databases leads to new insights, but is ...
doi.org
November 16, 2025 at 10:21 PM
Sounds a bit like doi.org/10.7717/peer..., but if I understand you correctly what you are after is not “here are some names and identifiers in different databases, match them” but “I have some names and some context, match them to a gold standard list of names using a model”?
So the incentives are different. I’m still curious as to why you don’t think @wikidatacommunity.bsky.social isn’t the answer to the identifier mapping problem.
November 16, 2025 at 6:34 PM
So the incentives are different. I’m still curious as to why you don’t think @wikidatacommunity.bsky.social isn’t the answer to the identifier mapping problem.
Companies and banks have money, and a strong incentive to resolve entities. Biodiversity projects have little money and have an incentive to mint their own identifiers. It gives them control and novelty. Plus I’m not sure companies want to share their identifier mappings.
November 16, 2025 at 6:33 PM
Companies and banks have money, and a strong incentive to resolve entities. Biodiversity projects have little money and have an incentive to mint their own identifiers. It gives them control and novelty. Plus I’m not sure companies want to share their identifier mappings.
There is also the question of why every taxonomic project deems it necessary to create its own identifiers for taxa. Arguably that is the root problem. What if people settled on using, say, Catalogue of Life identifiers in their projects, or at least mapped their local identifiers to CoL?
November 16, 2025 at 4:00 PM
There is also the question of why every taxonomic project deems it necessary to create its own identifiers for taxa. Arguably that is the root problem. What if people settled on using, say, Catalogue of Life identifiers in their projects, or at least mapped their local identifiers to CoL?
Who would manage these mappings? Wikidata seems the obvious candidate, and is already used by @gbif.org in that role. If a particular identifier type is missing from Wikidata, then it can be added. There are issues that identifiers may be for names or taxa, hence the mapping may be 1 to many.
November 16, 2025 at 3:55 PM
Who would manage these mappings? Wikidata seems the obvious candidate, and is already used by @gbif.org in that role. If a particular identifier type is missing from Wikidata, then it can be added. There are issues that identifiers may be for names or taxa, hence the mapping may be 1 to many.
Reposted by Rod Page
I've interviewed *many* researchers in the past years, and few if any use ORCID to it's full potential (or want to), or speak about it with excitement because it's hard to use and doesn't really meet their needs. They have them because they're required to have them to publish. Their words, not mine.
November 14, 2025 at 10:48 PM
I've interviewed *many* researchers in the past years, and few if any use ORCID to it's full potential (or want to), or speak about it with excitement because it's hard to use and doesn't really meet their needs. They have them because they're required to have them to publish. Their words, not mine.
I think this is just Kingsley’s writing style, he has been relentlessly advocating for linked data for years now. He runs a company @openlink.bsky.social so has skin in the game.
November 13, 2025 at 5:33 PM
I think this is just Kingsley’s writing style, he has been relentlessly advocating for linked data for years now. He runs a company @openlink.bsky.social so has skin in the game.