Page 58 - Building Digital Libraries
P. 58

Acquiring, Processing, Classifying, and Describing Digital Content


                 The http://purl.oclc.org is the PURL resolver, and the /digitallibraries/book
                 is an arbitrary string that points to a record for this book. If it is desirable to
                 have the PURL point only to the latest edition, the address that the PURL
                 points to can be changed without affecting the original URL. The advantage
                 of using a PURL over a regular URL is that a PURL resolution service is
                 likely to be more stable. If a resource is migrated, the access URL will most
                 likely be different even if the same domain name is used, but the structure
                 of the PURL will remain the same.
                     Like PURLs, the Handle System provides a level of indirection to
                 resources. It is different than PURLs in that rather than consisting of arbi-
                 trary strings which are managed on a central service, handles consist of a
                 naming authority and a name managed on a distributed service. Consider
                 the following URI:
                        10.6083/M4NZ86RR

                 The 10.6083 is the naming authority, and the name of the object is
                 M4NZ86RR. To access it, the value needs to be routed through a proxy
                 handle server, for example:

                        http://hdl.handle.net/10.6083/M4NZ86RR

                 which points to exactly the same place as the PURL above.
                     A Digital Object Identifier is a specific implementation of the Handle
                 System—by definition, all DOIs are also handles. In fact, the handle above
                 resolves as a DOI because it was originally created as one:

                        http://doi.org/10.6083/M4NZ86RR
                        http://hdl.handle.net/10.6083/M4NZ86RR
                 resolve to the same Internet address. Notice that the two different URLs are
                 both URIs that point to an object with the same DOI—10.6083/M4NZ86RR
                 (which is also a URI). The DOI never changes, but the URL can even if the
                 doi.org and hdl.handle.net domains are very stable. For this reason, it is best
                 to store URIs that are resolved to URLs on demand rather than URLs that
                 may someday be inaccurate.
                     DOIs are different from handles in that they have metadata assigned
                 to them, can express relationships, and are specifically designed to identify
                 objects such as articles, reports, datasets, and other resources. The DOI tech-
                 nology platform is designed to be easy to use for large numbers of resources.
                     PURLs, handles, and DOIs are different from identifiers such as Inter-
                 national Standard Book Numbers (ISBNs) and International Standard Serial
                 Numbers (ISSNs) in that they are actionable and are designed to take users
                 directly to a resource rather than merely identify them. However, be aware
                 that redirected links receive lower rankings in search engines—and more-
                 over, semantic information in the URL cannot help the resource be found.
                     If a library simply assigns unique identifiers on its own without using
                 the PURLs, handles, or DOIs, the objects will still be findable, presuming



                                                                                                                      43
   53   54   55   56   57   58   59   60   61   62   63