Update: here is a more recent posting about this.
I promised in a recent comment to talk more about how the links work to WorldCat Identities. As I said before, the mechanism that WorldCat.org uses is an OpenURL, but complicated by the desire to put the new page within a frame. Here is the equivalent URL without the frame:
http://worldcat.org/identities/find ?url_ver=Z39.88-2004 &rft_val_fmt=info:ofi/fmt:kev:mtx:identity &rft.namelast=Austen &rft.namefirst=Jane &rft.id=info:oclcnum/908814 , slightly edited by adding some spaces so that the lines can break.
If you leave off the OCLC number, the system is no longer sure exactly which Jane Austen you are interested in and will give you a list to choose from. Here is the same example without the OCLC number: http://worldcat.org/identities/find ?url_ver=Z39.88-2004 &rft_val_fmt=info:ofi/fmt:kev:mtx:identity &rft.namelast=Austen &rft.namefirst=Jane, which should return something like this:
What is returned is really XML with a reference to an XSL stylesheet to transform the XML into the HTML displayed by the browser. The page has a brief explanation of the symbols and results formatting:
More than one person has suggested that the need for a key to the results indicates a problem. Although I think it useful (especially for the library-literate) to contrast between controlled and uncontrolled names, I'm less comfortable with the traditional personal/corporate split. In retrospect it would have been better to ignore the MARC coding on the names and merge the occasional Jane Austen that is coded as a corporate name in with the rest of those coded as a personal name.
Actually there is a much simpler URL which works too: http://worldcat.org/identities/find?fullName=jane+austen. This is the probably the format we will encourage people to use in external linking.
Although WorldCat Identities uses LCCNs when available there is no reason we can't add indexes to OCLC's Authority Record Numbers (ARNs). There is essentially a 1 to 1 correspondence between ARNs and LCCNs for names and it is ARNs that are carried along with controlled and linked headings in WorldCat.
Thanks to Ralph LeVan who designed, implemented and explained most of this.