- From: Markus Lanthaler <markus.lanthaler@gmx.net>
- Date: Wed, 4 Mar 2015 22:13:34 +0100
- To: <public-hydra@w3.org>
Hi François-Paul, Sorry for the delay. I have been busy and unfortunately no one else jumped in answering your question. On 24 Feb 2015 at 15:20, François-Paul Servant wrote: > if I understand correctly, the first goal of the "manages" bloc is to tell clients what a > collection represents. As it is, it allows to define a simple triple pattern, eg. > </alice> foaf:knows ?x Yes, kind of. It works around the issue you would have with vocabularies that use RDFS range. Assume you would have a triple </alice> foaf:knows </my-collection> In principle, that would make it very clear to the client what can be expected to be found in that collection. The issue, though, is that the semantics of foaf:knows are defined in such a way that </my-collection> is interpreted to be a person. So we need to indirectly reference the collection but nevertheless tell the client enough to be able to find the right one (there could be multiple) </alice> hydra:collection </my-collection> </my-collection> hydra:manages [ hydra:subject </alice> hydra:property foaf:knows ] > In many cases however, a collection could be much more complex than that. Certainly. But if that's the case, I wouldn't generally use the manages block but introduce a special property pointing to it (something replacing foaf:knows in the first example). > For instance, it could be the list of people who share an interest with Alice (a server > may well want to publicize its capability to return such interesting information about persons). > What could we do for such collections? One solution of course would be to define an ad- > hoc property. We can think of other ways, including some ways based on well established > standards. For instance, we could think of using SPARQL to define the collection, or > OWL. Shouldn't hydra tries to open such possibilities ? Yeah. We can certainly look at them as well. But I'd say first we need to get the basics right. We still haven't consensus on paging for instance. That being said, if you have concrete proposals, don't hesitate to share them. I'm more than happy to discuss them. I would just like to avoid getting sidetracked with too abstract discussions for advanced use cases. -- Markus Lanthaler @markuslanthaler
Received on Wednesday, 4 March 2015 21:14:17 UTC