Hello all,
I have a question regarding Crossref test account prefixes such as those found here: https://api.crossref.org/members?query=test%20accounts
and https://api.crossref.org/members?query=Crossref
Some of these, such as 10.5555, have been ingested into OpenAlex, and I’ve posted a question about this in their users forum. Are there any of the Crossref prefixes that are valid for publications? Or are they all used for various testing purposes? Here is the list of prefixes from both of the API calls:
test_prefixes = [“10.18810”,
“10.5555”,
“10.88888”,
“10.30444”,
“10.30446”,
“10.30447”,
“10.30448”,
“10.30449”,
“10.50505”,
“10.13003”,
“10.30443”]
Some of these prefixes return a large number of publications:
10.18810 = 29
10.5555 = 317771
10.88888 = 24
10.30444 = 0
10.30446-9 = 0
10.50505 = 4346
10.13003 = 54
10.30443 = 0
10.13003, under the “Crossref” name, seems to be valid Crossref publications based on their titles. As well, 10.18810 seems to contain trials registries, though it is under the “Test accounts” name.
Is it safe to assume that everything in 10.5555, 10.88888, and 10.50505 are all invalid and just for testing?
Many thanks for your help and I hope you all are doing well!
-Poppy
For the most part, those prefixes either for testing or used as a kind of placeholder container for DOIs that have been (sort of, but not really) “deleted”.
10.13003 is used for Crossref’s own publications. For example, we’ll register slide decks or records from our events. Those are valid and real publications.
10.18810 is the prefix for Linked Clinical Trials registry. I’m not sure why it’s attached to the “Test accounts” member. The DOIs under it are legitimate identifiers, but not for publications. It’s similar to how the 10.13039 prefix is used for the Funder Registry identifiers.
Is it safe to assume that everything in 10.5555, 10.88888, and 10.50505 are all invalid and just for testing?
It’s safe to assume those aren’t valid publications. 10.5555 in particular has been used both for testing and as a holding place for DOIs that were created inadvertently and aren’t being actively stewarded. We also overwrite their metadata with placeholder data, but that wasn’t always true in the past, so you might occasionally find real metadata on 10.5555-owned records. 10.88888 and 10.50505 are used strictly for testing.
It would be ideal if OpenAlex didn’t index those prefixes, other than 10.13003 - if they need any conformation from us directly, we’ll be happy to reach out.
Wonderful - thank you for sharing and confirming about 10.13003 and 10.18810. I’ll post this to my discussion in the google user groups for openalex as it may be helpful for others.
Thank you!
-poppy