- From: John Walker <john.walker@semaku.com>
- Date: Thu, 8 Oct 2015 10:25:03 +0200 (CEST)
- To: Ruben Verborgh <ruben.verborgh@ugent.be>
- Cc: Hydra <public-hydra@w3.org>, Melvin Carvalho <melvincarvalho@gmail.com>
Hi Ruben > >> The document would be named </items/10268448>, > >> the resource you're after would be named </items/10268448#id>. > > > > Does it make sense to update examples from the post accordingly, just to > > please the purists ;) > > Thanks for the suggestion. Done for all non-information resources. Excellent. > > The results pages are information resources that can be directly represented, > so there's no need to distinguish there. Agree > > >>> So </items/10268448.ttl> and </items/10268448.jsonld> for the Turtle and > >>> JSON-LD > >>> respectively. > >> > >> That hinder interoperability between clients that use different media > >> types. > >> We need an identifier for the document as well, regardless of media type. > > > > Presumably to enable PUT/PATCH requests and so on? > > Yeah, and just so they know they're talking about the same thing. > If one client is talking about </items/10268448.ttl>, > and another is talking about </items/10268448.jsonld>, > are they talking about the same thing? > If they use the same identifiers, we know the answer immediately. Sure. As I see it the only reason to mint URIs for the representation would be if one want to provide additional metadata about it e.g. what version of software was used to generate that representation. John
Received on Thursday, 8 October 2015 08:25:39 UTC