OT: DataCite Recommends "doi:" Scheme Prefix

(Posted here so I don't forget it and others can find it...)

In the latest DataCite Metadata Kernel document, section "2.2 Citation", DataCite recommends ("prefers" is their term) that original-format DOIs used in citations include the doi: scheme prefix. ("Original format" refers to non-URL DOIs.) Unfortunately, the DataCite Metadata Kernel document does not mention that the string doi: is a scheme prefix, making it hard to find.

An example DOI with the scheme prefix is doi:10.1006/ijhc.1996.0110.


What kind of scheme are you talking about? I cannot find "doi:" in the URI namespace registry. DOI is registered as scheme "info:doi/" under the info URI specified by RFC 4452. The official DOI registry, however, uses the URI prefix "http://dx.doi.org/" in its RDF representation of DOI data: try RDF::Trine, curl, ... to access the data:

curl -LH 'Accept: text/turtle' http://dx.doi.org/10.1006/ijhc.1996.0110

BTW it looks like your sample DOI refers to an arbitrary journal article instead of a DataCite URI. See this article for another example (also using the prefix "http://dx.doi.org/").

Mark -

I've been thinking lately about citing my Perl work in journal articles, specifically via a doi. Do you know if there is any doi provider for Perl modules? I strongly suspect not, but figured I'd check with somebody else thinking about these things. :-)

DOI is big business. There are some registration agencies that provide free DOI registration for recognised academic institutions, but outside that world you need to pay quite a bit to register DOIs.

DOIs are really just a form of persistent URI; and if you want that, then IMHO you're better off going the PURL route. A few years ago I set up a PURL identifier system for all CPAN distributions, with identifiers taking the form:


(where %s is the distribution name, such as Role-Tiny)

The system serves RDF data about the distribution using the DOAP vocabulary. It's in need of some overhaul - I'd like it to make use of the MetaCPAN API.

How do DOIs compare to ISBNs? I had to get one of those recently, and it was basically "bend over and accept Bowker's thorny shaft." Has some useless company with a mutex and an integer managed to corner the DOI market as well?

Leave a comment

About Mark Leighton Fisher

user-pic Perl/CPAN user since 1992.