- From: Ruben Verborgh <ruben.verborgh@ugent.be>
- Date: Wed, 10 Sep 2014 20:12:23 +0200
- To: Erik Wilde <dret@berkeley.edu>
- Cc: László Lajos Jánszky <laszlo.janszky@gmail.com>, public-hydra@w3.org
Hi Erik,
> that's assuming that the "property" property (sorry for that) is defined to be a link (that you have to follow to get type information) instead of an identifier. that's possible, but makes things more chatty. if that's what you want to do, the model has to be explicit that this is a link and not an identifier. right now i think this is not mentioned in the spec.
hydra:property has range rdf:Property in the spec [1], so that makes it at least an IRI.
And since we're talking about Linked Data here, probably a URL as well.
Not necessarily more chatty: you can also include this information in the message,
i.e., just add the following to the response.
<http://example.com/feedpaging/page> rdfs:range xsd:positiveInteger.
The benefit of this is reusing existing stuff.
>>> and value ranges.
>> That's not supported at the moment.
>>> "xs:minInclusive": "1",
>>> "xs:maxInclusive": "42",
>> Something like this might maybe be part of Hydra if there is demand for it. Opinions?
>
> just fyi: this is runtime information instead of design time
True, we should be careful separating the property-specific and control-specific information;
i.e., the property is always a positive interface, but only in this control limited to 142.
(The OWL way would be to create a subproperty for this; we don't wanna go there ;-)
Best,
Ruben
[1] http://www.hydra-cg.com/spec/latest/core/#the-hydra-core-vocabulary-in-json-ld
Received on Wednesday, 10 September 2014 18:12:58 UTC