Hi Karen, On 5 January 2012 12:36, Karen Coyle <[email protected]> wrote: > Quoting Ben Companjen <[email protected]>: > > The Work's >> record description, however, is not linked to the work description. >> Only to agents who interpret parts of the URI may the relationship >> between a frbr:Work and something identified by "http:// ... /about/" >> be clear. >> I think this may be solved by saying >> <rdf:Description rdf:about=""> >> <dcterms:created>...</dcterms:created> >> <dcterms:modified>...</dcterms:modified> >> </rdf:Desciption> >> since rdf:about="" means 'about this document'. > > That's a neat trick - I'll want to remember that. However, the > created/modified refer to the OL record, so the /about/ bit was > supposed to work like the /about/ of the 'cool uri' document. We use > the OL URI to identify the Work not the record about the work, so we > were struggling with a way to give information about the OL Work > *record.*
I understand that. In the current situation however, there is nothing saying that the "/about/" URI identifies the record. And there is no link between the <rdf:Description>s of the record and the work. <foaf:page> is an example property that links documents to the subject of the pages, but I'm not sure if records can/should be called documents (I think they're more abstract, but again: not a librarian :)). > > That said, if it's clearer I don't see a reason not to use 'about this > document'. I don't know of a case where it makes a difference. Only thing I can think of right now is that if the last modified property applies to the whole record, it may be implicit that all its representations have that property, instead of just the RDF document. > >> >> I committed an adaptation of the Work RDF template in my fork (this is >> my first day on Github, but I think I'm starting to get the hang of >> it), but I was not sure what other people think* about losing the >> trailing slash in Work and Edition URIs, so I left them in at first. I >> committed an additional change, that removes the trailing slashes. >> *: in the RDF improvement bug report, >> https://bugs.launchpad.net/openlibrary/+bug/703664, Karen says all >> URIs should end with "/". > > Yes, some folks on this list when we were first developing the RDF > felt strongly about that. What do people think now? Most example URIs I see do not have "/" at the end. It is of course possible to say of all <URIs without "/"> <owl:sameAs> <URIs with "/">, but I haven't seen any situation in which that is applied. The Talis Open Library dataset only contains URIs without slash. The other datasets in the LOD cloud that link the OL dataset are either unavailable or apparently do not contain OL URIs. > >> >> Please have a look (if you can) at >> https://github.com/bencomp/openlibrary/blob/patch-1/openlibrary/plugins/openlibrary/templates/type/work/rdf.html >> and >> https://github.com/bencomp/openlibrary/blob/patch-1/openlibrary/plugins/openlibrary/templates/type/edition/rdf.html >> and comment on it. If all is okay, I will send a pull request. > > Is there a way to test this without moving it into production? Also, > there may be people using the RDF API and it would be good to give > them notice before a change. I'll admit my proposed solution is quick & perhaps dirty, so testing may be a good thing. I don't know about test facilities or people using the RDF except the administrators of datasets listed on thedatahub.org. Ben > > kc > > > -- > Karen Coyle > [email protected] http://kcoyle.net > ph: 1-510-540-7596 > m: 1-510-435-8234 > skype: kcoylenet > > _______________________________________________ > Ol-tech mailing list > [email protected] > http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech > To unsubscribe from this mailing list, send email to > [email protected] _______________________________________________ Ol-tech mailing list [email protected] http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech To unsubscribe from this mailing list, send email to [email protected]
