W3C home > Mailing lists > Public > public-openannotation@w3.org > August 2015

Re: Interlink annotations

From: Jacco van Ossenbruggen <Jacco.van.Ossenbruggen@cwi.nl>
Date: Thu, 06 Aug 2015 18:11:17 +0200
Message-ID: <55C38725.40505@cwi.nl>
To: public-openannotation@w3.org
On 08/05/2015 11:46 AM, Andre Z. wrote:
> Hello everyone,
> I wonder whether it is possible to establish a directed connection 
> between annotation objects. The scenario is as follows: One has 
> identified two concepts in a document by semantic concept tagging. Now 
> you want to describe the relation between both by a property. I have 
> thought about this a bit and find two approaches useful:
> Approach (1): Annotation A is referencing the annotation B (and 
> probably vice versa)
> Approach (2): A third annotation C describes the relation between A and B
> Does someone could imagine how to model this in OA?
> Kinds regards
> Andre

Hi Andre,

Good question.  My 2 cents:

Assuming both annotations A and B and the property are identified by 
URIs A, B and P resp., you can easily do your approach (1): Just create 
a triple

     A P B .

(or vice versa: B P A. ).   You need to write your own application logic 
to do something useful with that triple, but it should not break any 
existing OA software. I would probably take this route myself, because I 
would need only one triple extra and have the full flexibility of using 
RDF to define or reuse whatever property I want.

or: do your approach (2): just create annotation C with A as the body 
and B as the target (or vice versa).  Downside is that there is no 
logical place to model the property other than in the oa:Motivation 
bits, which is as general as the "real" RDF property in the first 
approach.  Old-timers on this list might remember I ranted in the past 
that not being able to use a plain RDF property to model the relation 
between body and target is a missed opportunity for OA.  But I never got 
enough support to do something about this... oh well... you cannot get 
everything in life ...  :-)

Hope this helps,

Jacco
Received on Thursday, 6 August 2015 16:11:47 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 6 August 2015 16:11:48 UTC