- From: Martynas Jusevicius <martynas@graphity.org>
- Date: Fri, 22 Nov 2013 15:24:36 +0100
- To: Ruben Verborgh <ruben.verborgh@ugent.be>
- Cc: public-lod <public-lod@w3.org>
Hey Ruben, regarding RFC6570, I'm not planning to adopt it, since the specification is better suited for building URIs, not matching them (1.4 Limitations): "In general, regular expression languages are better suited for variable matching" I'm using JAX-RS syntax since it can be used for matching and building, and has utility classes that help you do that. I don't like how Linked Data API uses endpoints either. I reuse api:ListEndpoint in some of Graphity ontologies, but give it no special meaning. If I understand right what you want, I am using SIOC to define the parent/child relationship between resources (which is usually represented by a slash in the URI): http://rdfs.org/sioc/spec/#term_Container In your case, it could be: </topics> a sioc:Container . </topics/Global_Warming> sioc:has_space </topics> . Cheers, Martynas graphityhq.com On Wed, Nov 20, 2013 at 4:52 PM, Ruben Verborgh <ruben.verborgh@ugent.be> wrote: > Hi Martynas, > >> - URI templates: Linked Data API vocabulary >> https://code.google.com/p/linked-data-api/wiki/API_Vocabulary > > Cool, I do like that. Have you thought about extending to RFC6570? > Do you know about usage of this vocabulary? > > The one thing that I like less is the notion of endpoints. > While this is perfect for SPARQL, which is indeed an endpoint > or “data-handling process” that expects a “block of data” [1], > it does not work well in resource-oriented environments. > > I’m looking for predicates that work with groups of resource, such as: > </topics/Global_Warming> :belongsTo </topics>. > </topics> a :ResourceList; > api:itemTemplate "/topics/{topicID}". > That is, I don't consider there to be a "topics" endpoint; > instead, there is a "topics" resource which lists several topics, > and individual topics can be accessed by ID. > The reason I would still need the template is because /topics/ is not exhaustive, > and new elements can be created by following the template. > This would be equivalent to a HTML GET form. > >> - HTML forms: RDF/POST encoding http://www.lsrn.org/semweb/rdfpost.html > > Interesting, thanks! > > Best, > > Ruben > > [1] http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html#sec9.5
Received on Friday, 22 November 2013 14:25:06 UTC