- From: Markus Lanthaler <markus.lanthaler@gmx.net>
- Date: Wed, 1 Oct 2014 17:56:40 +0200
- To: <public-hydra@w3.org>
On 27 Sep 2014 at 13:02, Dietrich Schulten wrote: > Am 17.09.2014 20:06, schrieb Kev Kirkland: >> I think the solution you mention below does this (if I interpret >> it correctly). That is we return a representation that has a single >> Hydra operation and the single operation is to POST some >> information using a SupportedClass? > > In fact, that is one possibility. Mike Amundsen once identified > exactly this as a possible way to tell a client what to do in the > hypermedia-web group. Yeah, it's certainly a viable option to do this. I, however, would typically prefer something a bit more explicit to guide clients through a unknown path to a certain goal. > Another is a rel the client looks for. The IANA > rel 'next' might be a good candidate for a wizard BTW :) Indeed... but obviously it depends on how you use it in detail > My interpretation of json-ld attributes having links as values is that > they are a special case of relation types. The attribute names are > always IRIs, if you resolve them. The attribute "event" on the > schema.org class "Organization" can be interpreted as an extension rel: > Link rel=http://schema.org/event href="http://api.example.com/events". > Only, they are not registered with IANA, but published in another > well-known place. Correct. Actually, for a while all IANA link relations had a canonical mapping to URLs [1] but that was lost with RFC 4287 [2]. I tried to reintroduce that a while ago but there wasn't much support on the IETF side, just a little from W3C (which would obviously like to use it in RDF-based specs). > So a generic client could also be parameterized to look for "event", > then for "offers" etc. Right > This means that clients follow sequences of rels. With the rel 'next' > it could also follow without knowing exactly how many times. Yeah, for navigational traversal (think a paged collection) that's a great pattern. For state changing interaction however I think you would typically want something more explicit. Cheers, Markus [1] http://tools.ietf.org/html/rfc4287#section-4.2.7.2 [2] https://tools.ietf.org/html/rfc5988 -- Markus Lanthaler @markuslanthaler
Received on Wednesday, 1 October 2014 15:57:12 UTC