<< Click to Display Table of Contents >> Navigation: DOI Applications > Redirection to Linked Data Services |
Linked Data is the general term for a set of best practices for exposing data in machine-readable form using the content-negotiation feature of the standard HTTP(S) web protocol. These best practices support the development of tools to link and make use of data from multiple web sources without the need to deal with many different proprietary and incompatible application programming interfaces (APIs), and use of HTTP(S) to request data in structured form meant for machines instead of human-readable displays.
In Linked Data applications, evaluation of the HTTP(S) request that comes in to the proxy service determines if it is a request for content of form application/rdf+xml, or one of a few similar types that are commonly understood to be a request for Linked Data. These requests for special content types would come from automated processes or special "linked data" browsers and would not normally come from end users. The utility of this, of course, is that it allows outside developers to query the extensive and reliable set of metadata records held by Registration Agencies (RA) to build value-added services.
Some RAs are using this approach for all of their DOI names, offering services that return metadata in a common machine-readable format. A significant advantage of applying Linked Data principles and technologies to DOI-registered material is that it is "data worth linking to": it is curated, value-added, data, which is managed, corrected, updated and consistently maintained by RAs. It is also persistent, so avoiding "bit-rot". In practice, the quality of Linked data implementations is only as good as the data you are linking to, and the meaning and contextualization of the link you use. The DOI System can offer "curated data", it means consistent, managed, linking so you can link to other "quality data" with confidence, while still using the standard Linked Data technologies.
The following figure shows a redirection to the metadata service managing the DOI name. The 10320/LOC element information can be stored at prefix level (see 10320/LOC at Prefix Level): should an RA at some point need to change their approach to linked data and point to a different service or use different parameters, the change could be made to a single DOI prefix and it would apply to all of the millions of DOI names automatically.
For more information, see Content Negotiation.