Re: Hydra Design Goals: How important is RDF?

Hey Karol,

I can show you how RDF hypermedia looks like. It can be implemented as
a simple response filter that augments RDF query result with state
transitions, depending on the requested resource:
https://github.com/Graphity/graphity-processor/blob/master/src/main/java/org/graphity/processor/filter/response/HypermediaFilter.java

The basic transitions look like this (URIs relativized to base):

# page resource

<?offset=20&limit=20>
        a          gp:Page , foaf:Document ;
        gp:limit   "20"^^xsd:long ;
        gp:offset  "20"^^xsd:long ;
        gp:pageOf  <> ;
        <http://www.w3.org/1999/xhtml/vocab#prev>
                <?limit=20&offset=0> .

# resource constructor

<queries?mode=http://graphity.org/gp%23ConstructMode&forClass=http://graphity.org/gp%23Item>
        a                 gp:Constructor , foaf:Document ;
        gp:constructorOf  <queries> ;
        gp:forClass       gp:Item ;
        gp:mode           gp:ConstructMode .

As you can see, RDF is essential here.


Martynas
graphityhq.com

On Thu, Oct 1, 2015 at 9:56 PM, Karol Szczepański
<karol.szczepanski@gmail.com> wrote:
>>>The Hydra Core Vocabulary is a vocabulary to describe hypermedia controls.
>
>>>So Hydra Core Vocabulary constructs will be inside of API responses.
>
>>>That's not the case with service description stuff like WSDL, Swagger,
>>> etc.
>
>
>
> Well, I’m somehow dissappointed and I hope Hydra to be more than this.
>
> While hypermedia is something what we seek I’m not that keen to throw
> everyhing into API responses altogether.
>
> I like to have separated raw data and hypermedia controls. I feel that my
> point of view is not that abandoned.
>
> The spec points to API documentation in very first paragraphs, thus I hope
> we’ll not close ourselves only to having embedded hypermedia controls.
>
> I think this will greatly limit the use cases for Hydra itself and will
> become it’s first nail to the coffin.
>
>
>
> Best,
>
> Karol Szczepański
>
>
>
>
> Od: Ruben Verborgh
> Wysłano: czwartek, 1 października 2015 11:50
> Do: Asbjørn Ulsberg
> DW: Hydra
> Temat: Re: Hydra Design Goals: How important is RDF?
>
>
>
>
>
>> As with JSON-LD, I feel the RDF part of Hydra is both
>
>> under-communicated and more of a nice-to-have than the core value of
>
>> the technologies. I think this is a good thing. While RDF and the
>
>> Semantic Web is awesome in its prospects, I highly doubt most people
>
>> getting their hands dirty with JSON-LD or Hydra will have a Semantic
>
>> Web perspective or problems related to RDF to solve.
>
>
>
> Yes, and it that sense Hydra is a great bridge.
>
>
>
> However, let's be very clear here:
>
> – JSON-LD without RDF is just JSON
>
> – just JSON is not self-descriptive
>
> – without self-descriptiveness, Hydra becomes a spec like all others
>
>
>
> For the last point, we build intelligent clients
>
> that strongly rely on the self-descriptiveness of interfaces.
>
>
>
>> Related and relevant: "JSON-LD and Why I Hate the Semantic Web"
>
>> http://manu.sporny.org/2014/json-ld-origins-2/
>
>
>
> I also recommend people to read this,
>
> but for different reasons: it explains well how to write a spec.
>
>
>
>> I expect JSON-LD to be used as a way to express URIs and hypermedia in
>
>> JSON and I expect Hydra to be used as "The WSDL of HTTP / REST".
>
>> Please arrest me if I'm wrong in these assumptions.
>
>
>
> I don't agree.
>
> The Hydra Core Vocabulary is a vocabulary to describe hypermedia controls.
>
> So Hydra Core Vocabulary constructs will be inside of API responses.
>
> That's not the case with service description stuff like WSDL, Swagger, etc.
>
>
>
> Additionally, it allows to describe the structure of an API.
>
>
>
>> Because of this, I think it's important to state this nice-to-have
>
>> status of RDF as a design goal, since from what I've gathered so far
>
>> from the discussions on this list, this isn't necessarily something
>
>> everyone is in agreement with.
>
>
>
> RDF is a means, not a goal.
>
> RDF enables self-descriptiveness, which is a goal.
>
>
>
> Best,
>
>
>
> Ruben
>
>
>
>

Received on Thursday, 1 October 2015 22:01:43 UTC