DOI

Owned by James Myers
Oct 26, 2016
2 min read

Different DOI registration agencies support a number of metadata access variants using the basic Accept Header scheme that is one option for this pilot project. Documentation is available at http://citation.crosscite.org/docs.html

A simple example is the DOI:10.13012/J8MW2F2Q which can be resolved using http://doi.org/10.13012/J8MW2F2Q . This DOI refers to the streaming Illinois Climate Network: Monthly and Daily Data available from the Illinois State Water Survey. Using an Accept: text/turtle header with this GET request returns the following:

<https://doi.org/10.13012/J8MW2F2Q>
<http://purl.org/dc/terms/creator>
"Water and Atmospheric Resources Monitoring Program" ;
<http://purl.org/dc/terms/date>
"1998" ;
<http://purl.org/dc/terms/identifier>
"10.13012/J8MW2F2Q" ;
<http://purl.org/dc/terms/publisher>
"Illinois State Water Survey, Prairie Research Institute, University of Illinois at Urbana-Champaign" ;
<http://purl.org/dc/terms/title>
"Illinois Climate Network" ;
<http://www.w3.org/2002/07/owl#sameAs>
"info:doi/10.13012/J8MW2F2Q" , "doi:10.13012/J8MW2F2Q" .


Other Notes/Comments:


In an NDS Hackathon, a simple resolver service was created that allowed wrapping of this URL call with a type-independent resolver that produced json-ld. This resolver required a URL of the form /acr/resteasy/urls/doi%3A10.13012%2FJ8MW2F2Q/metadata and produced a json-ld output (pretty-printed below) that is the same (fairly generic) format SEAD would produce for its internal identifiers. (The simple resolver code read the turtle file and mapped all RDF statements about the dataset to json-ld and was not specific to the metadata sent)

{

}


This specific DOI is potentially a useful outlier example for our group in that it does not reference a fixed set of files and, like many datasets, has fairly minimal metadata available directly via the DOI. So our ability to do anything with this dataset across NDS might be limited to just being able to catalog and list it. Further metadata would be needed to discover any files in it, or to recognize that the set of files available could be different tomorrow (in this case it is probably just bew files being added as data continues to come in). One goal of the approach being pursued is to make these types of extensions simpler/more incremental - more metadata can be sent if/when/as the value in agreeing and adding it outweighs the incremental cost to do so.