link
Publishers of electronic scholarly content join CrossRef as members and are assigned a DOI prefix. For each content item the publisher creates a unique DOI name by appending a suffix of their choosing to the prefix they've been assigned by CrossRef (e.g. "10.1006" is an assigned prefix and "jmbi.1995.0238" is a suffix composed by the publisher) . The publisher then creates an XML file which associates metadata (journal title, author, article title ... etc) with the DOI. The publisher then submits (uploads) the file to the CrossRef system.
The metadata associated with a DOI can include:
When a DOI is deposited CrossRef performs two actions, 1) we register the DOI name and URL in a central DOI directory and 2) we store the metadata in our lookup service. The central directory is what makes DOI names work when someone clicks on a DOI based link. The directory works by looking up the current URL for the specific DOI and sending the end user to that location. If the item moves to a new location the URL link seen by users will still work thereby establishing a permanent link. When a publisher changes the location of the content they need to update the URL in only one place: with CrossRef.
When you click on a DOI name link online, you are directed to the URL submitted by the publisher. If you know a DOI name for an item, you can turn it into a URL by appending the DOI to http:/dx.doi.org and using it in a browser's address bar. For example:
CrossRef's lookup service allows users (including publishers, libraries, researchers, authors or the public) who know some amount of information about an item to retrieve the DOI which allows them to build a link to the item as described above. Users submit queries to CrossRef which include as much metadata as they have (e.g. journal title, author, article title ... etc) and CrossRef looks for the corresponding DOI and returns it to the user. Unlike general purpose search engines the CrossRef system is designed to (typically) return only one DOI, the one that best fits the query supplied metadata. This behaviour is essential since the majority of queries sent to CrossRef come from publisher and library automated systems which are not equipped to deal with a multiple choice result.
Although most transactions with CrossRef are made by automated systems we have put in place several services designed for use by people. For smaller publishers who have limited technical resources or for end users like authors, researchers and librarians these services are intended for low volume activity. For detailed instructions on these services please see The CrossRef user interface.
Note: For a more extensive discussion of DOI names, see The DOI Handbook (pdf) on the Digital Object Identifier System web site.
