 |
 |
Thanks Ed. Note that the URI without the trailing slash does a 303 (Linked Data) redirect to the URI with the slash. Doing a 301 back would create a redirect loop. VIAF uses the trailing slash URI to support the Linked Data Cool URIs pattern called "303 URIs forwarding to One Generic Document" <http://www.w3.org/TR/cooluris/#r303gendocument>. Unfortunately, I didn't heed the warning documented in the alternative pattern called "303 URIs forwarding to Different Documents" <http://www.w3.org/TR/cooluris/#r303uri>: "When the RDF and HTML representations of the resource differ substantially, the previous setup should not be used." Unfortunately, the die is probably cast. The HTML negotiated from the current generic resource is destined to deliver a user-interface experience that has very little correlation with the RDF. I have an idea for a solution, but it will require some effort to implement. :-( Jeff > -----Original Message----- > From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of > Ed Summers > Sent: Monday, April 11, 2011 3:25 PM > To: [log in to unmask] > Subject: Re: [CODE4LIB] FW: VIAF linked data and non-Latin searching > > Nice, Jeff. I really like the simplified VIAF RDF. In particular I > like how you've modeled the deprecation of resources. Are you planning > to use a 301, e.g. http://viaf.org/viaf/77390479/ -> > http://viaf.org/viaf/77390479 ? > > //Ed > > On Mon, Apr 11, 2011 at 3:18 PM, Young,Jeff (OR) <[log in to unmask]> > wrote: > > Here is some information about pending updates to the VIAF Linked > Data. > > > > > > > > I'm working on before/after diagrams to better explain the > differences > > and will share them soon. Questions and comments are welcome. > > > > > > > > Jeff > > > > > > > > From: Hickey,Thom > > Sent: Monday, April 11, 2011 12:57 PM > > To: [log in to unmask] > > Cc: Young,Jeff (OR) > > Subject: VIAF linked data and non-Latin searching > > > > > > > > Non-Latin searching: > > > > > > > > We believe we have resolved a reoccurring issue with non-Latin > searching > > failing (it had to do with restarting VIAF in different > environments). > > If anyone still has issues with this, please let us know. > > > > > > > > Linked Data: > > > > > > > > We have taken another look at the RDF generated for linked data. The > > attached files show a personal, corporate and geographic (there are > few > > pure geographic records in VIAF as of yet, but a mixed record such as > > Missouri's may be identified as geographic) record rendered in RDF. > > > > > > > > We think the new records are both simpler and easier to understand > and > > use. The biggest difference is that we have eliminated the > > viaf:NameAuthorityCluster that acted as a "record hub". Formerly, > this > > record hub was responsible for linking to the separately identified > > primary entity. In the new record structure, contributed authorities > > bypass this record hub and link directly to the primary entity > > themselves. The description of the primary entity appears first in > the > > record inside an rdf:Description element followed by skos:Concept > > entries, one for each source file, each of which links back to the > > primary entity via foaf:focus. > > > > > > > > We have included some deprecated identifiers matching those used in > > previous RDF, which may help those processing it as linked data. For > > those simply parsing it as XML and pulling information out of it, we > > have switched to fully qualified URIs which should make that easier. > > > > > > > > We will probably phase the new RDF in over the next two months. This > > month we will generate both for those getting full dumps of VIAF, > then > > next month switch both the online and offline versions to the new > > format. > > > > > > > > For those with suggestions about the new format, this would be an > ideal > > time to let us know. If we stay with the schedule outlined above we > > have until mid to late May before the new formats are in production. > > > > > > > > --Th > > > >