- From: Dan Brickley <danbri@google.com>
- Date: Tue, 2 Sep 2014 16:20:13 +0100
- To: "Wallis,Richard" <Richard.Wallis@oclc.org>
- Cc: "kcoyle@kcoyle.net" <kcoyle@kcoyle.net>, Chaals from Yandex <chaals@yandex-team.ru>, "<public-vocabs@w3.org>" <public-vocabs@w3.org>
On 2 September 2014 15:38, Wallis,Richard <Richard.Wallis@oclc.org> wrote: > Already in place: > > Schema:Book has an illustrator property > Schema:Movie has director, producer, productionCompany properties Glancing through http://www.loc.gov/marc/relators/relacode.html there's a lot in there. Are there any super-useful things we're missing that could be argued for inclusion as first class properties? Otherwise having a way to point from a Role to external lists seems a potentially useful convention (though perhaps overkill). >>>> On 1 Sep 2014, at 15:35, Karen Coyle <kcoyle@kcoyle.net> wrote: >>>>> I was under the impression that creative works would be able to use the "roles" pattern that was discussed [1]. Right now, CreativeWork has author, creator, and contributor. Although there are a handful of common creative roles that come to mind (editor, translator, illustrator), the actual number blossoms quickly when you move beyond books. Movies have a huge number of creative roles; music also has quite a few (librettist, composer, performer, lead singer...). The roles list used by the Library of Congress gives an idea of the magnitude of the problem. [2] >>>>> kc >>>>> >>>>> [1] http://lists.w3.org/Archives/Public/public-vocabs/2014May/0085.html >>>>> [2] http://id.loc.gov/vocabulary/relators.html Another option is to just use native (microdata/rdfa/json-ld) syntax for additional more detailed properties. If I remember right, multiple relation types between is a single pair of entities is most awkward in JSON-LD. In RDFa, this seems ok (needs @rel not @property): (where ANM is relator code for "A person contributing to a moving image work or computer program by giving apparent movement to inanimate objects or drawings"; perhaps not a perfect classification of Richard Williams' complex role here but enough for an example) <p vocab="http://schema.org/" typeof="Movie"> <span property="name">The Thief and the Cobbler</span> <link property="sameAs" href="http://en.wikipedia.org/wiki/The_Thief_and_the_Cobbler"/> <span rel="contributor http://www.loc.gov/loc.terms/relators/ANM"> <span typeof="Person"> <span property="name">Richard Williams</span> <link property="sameAs" href="http://en.wikipedia.org/wiki/Richard_Williams_(animator)"/> </span> </span> </p> >From a quick look, Microdata per http://www.whatwg.org/specs/web-apps/current-work/multipage/microdata.html#names:-the-itemprop-attribute should accept: <p itemscope itemtype="http://schema.org/Movie"> <span itemprop="name">The Thief and the Cobbler</span> <link itemprop="sameAs" href="http://en.wikipedia.org/wiki/The_Thief_and_the_Cobbler"/> <span itemprop="contributor http://www.loc.gov/loc.terms/relators/ANM"> <span itemscope itemtype="http://schema.org/Person"> <span itemprop="name">Richard Williams</span> <link itemprop="sameAs" href="http://en.wikipedia.org/wiki/Richard_Williams_(animator)"/> </span> </span> </p> If all we want is additional relationship types, then creating an intermediate Role entity could be rather heavyweight, since RDFa and Microdata seem to tolerate whitespace-separated lists of properties. Anyone care to sketch this out in JSON-LD? Dan ps. +1 for including translator alongside properties for indicating links between translations, and yes " (I think the sticking point on translation is about the difference between identifying the thing translated and the translation, or between noting that there *is* a translation without specifying which is the original...)" is the only sticking point I'm aware of
Received on Tuesday, 2 September 2014 15:20:41 UTC