W3C home > Mailing lists > Public > public-openannotation@w3.org > June 2012

Re: Q: how to encode annotation "roles"

From: Rob Sanderson <azaroth42@gmail.com>
Date: Wed, 13 Jun 2012 09:50:51 -0400
Message-Id: <59B20232-77DF-4102-97AD-7EF8D97158CD@gmail.com>
Cc: "public-openannotation@w3.org" <public-openannotation@w3.org>
To: Jacco van Ossenbruggen <Jacco.van.Ossenbruggen@cwi.nl>

Hi Jacco,

This is a tricky issue, as we don't want to re-invent RDF, encoded in RDF! :)
In an early version of the Open Annotation Collaboration schema, we did have a "predicate" relationship to make the relationship explicit.  This was seen as very strange by the people who reviewed it, as the object of the triple was a predicate. Nothing prevents this, but it's not a common pattern.

At the moment, then, we would recommend using classes if it's necessary to explicitly model individual relationships, but with the caution that you might end up duplicating every single relationship in RDF into a subclass of annotation.

Any further thoughts from the group on this issue are appreciated, especially from a Linked Data perspective?


Sent from my iPad

On Jun 12, 2012, at 4:18 AM, Jacco van Ossenbruggen <Jacco.van.Ossenbruggen@cwi.nl> wrote:

> Dear Bob, dear Robert,
> Thanks for your elaborate answers.  
> I understand Bob's concerns on the body restrictions, but I do not think this is a problem in my application.
> I can live with the fact that multiple tags live in multiple Annotation instances, and I can deal with the multilingual issues be assigning different language labels to the skos concepts I use as semantic tags
> To Robert: sorry if my example was not clear.  The rdf:type of the tags (like Person versus Building) should of course be modeled on the semantic tags, as they are not restricted to this particular image.  What I'm looking for is a way to model in general the nature of the relationship between the body and the target, so facts like is the body depicted on the target image or associated with it in another way.
> I think this is an issue for all pre-OA annotations that currently model annotations by simple (Body, Relation, Target) or (Target,Relation,Body) triples in RDF.  If these applications want to move to using OA, it is clear where the Body and the Target end up in the OA model, but less clear where the Relation part fits in.  For interoperability reasons I think it would be good if there where some hints on how to convert these single triple models to OA.
> Jacco
Received on Wednesday, 13 June 2012 13:51:20 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:22:00 UTC