- From: McBennett, Pat <McBennettP@DNB.com>
- Date: Sat, 2 Aug 2014 10:37:14 -0500
- To: Ruben Verborgh <ruben.verborgh@ugent.be>, "<public-hydra@w3.org>" <public-hydra@w3.org>
> -----Original Message----- > From: Ruben Verborgh [mailto:ruben.verborgh@ugent.be] > Sent: Friday, August 01, 2014 8:59 AM > To: <public-hydra@w3.org> > Cc: Markus Lanthaler > Subject: Re: Proposed resolution for ISSUE-64 (status codes) > > Dear all, > > I personally strongly agree with Markus' > > s/Remove all MUSTs about/Remove all normative statements about/ > and I find > > There was no error, neither on the client nor the server side. > > very convincing. > > That said, I would like to move forward as follows. > > The current specification for triple pattern fragments has: > - Section 3. A large normative part on what triple pattern fragments look like; > - Section 4. A small normative part on what HTTP servers of those fragments > look like. > > Given that the part on HTTP servers mostly consists of the status codes > discussion (the remainder is variable expansion, which will move to the Core > Vocabulary), this would allow us to reduce Section 4 to something very small > and non-normative. > > In other words, triple pattern fragments would be determined entirely by the > documents, not any other interface agreements, which nicely agrees with > the statement that "A REST API should spend almost all of its descriptive > effort in defining the media type(s) used for representing resources and > driving application state" [1]. > > Additionally, this would easily enable the deployment in non-HTTP > environments, because the spec would essentially be protocol-independent > (even though we would thoroughly explain how one can do it for HTTP). > > This would then also be a great example of the power of hypermedia and a > generic document type as enabled by the Hydra Core Vocabulary: > no URLs, no protocol, just the document type. > > Please let us know any comments, agreements, and/or disagreements. > +1 (just for the record, I was strongly against the '404 just means empty' approach too, for all the reasons already stated.) > Best, > > Ruben > > [1] http://roy.gbiv.com/untangled/2008/rest-apis-must-be-hypertext-driven
Received on Saturday, 2 August 2014 15:37:43 UTC