W3C home > Mailing lists > Public > public-openannotation@w3.org > January 2013

Re: New drafts - general comments and intro

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Wed, 9 Jan 2013 00:01:08 +0100
Message-ID: <50ECA534.4020902@few.vu.nl>
To: public-openannotation <public-openannotation@w3.org>
Hi Stian, Paolo,

I agree the directionality is a useful notion to keep. But if it's at the cost of having such a statement as
:ann1 a oa:Annotation ;
   oa:hasBody <http://en.wikipedia.org/wiki/List_of_Presidents_of_the_United_States> ;
   oa:hasTarget <http://dbpedia.org/resource/Bill_Clinton> .
for representing that someone tagged the wikipedia page with the resource representing Bill Clinton, then we're doomed!

This is indeed really counter-intuitive. To me the "tag" or "subject" (it could be a SKOS concept, a test-as-body, whatever) is the thing that annotates, not the thing that is annotated (NB: in fact while reading the previous versions of the spec, I had understood oa:semanticTag was playing the same functional role as oa:hasBody; I realize now I might have missed a big part of the motivation for keeping them separate then!)

One of the reasons for having the annotated resource always in the same "end" of the directed body-target relation is that in many scenarios there will be an annotation which is not specified.
Consider a system (and there are many) that allows me to annotate a picture such as
by any of the following (using strings as shortcuts):
"Leonardo da Vinci"
"16th century"
Should the system try to analyze the content of my tag to infer whether the picture should be the target or the body of the annotation?

We really should focus on the functional side of things. Maybe the problem with "about" is that it's too much loaded with intuitive semantics that have in the end only little to do with the technical aspect of annotation. We should rather aim at finding a word that expresses a function (which carries directionality indeed).
I was thinking about "the body applies to the target" with 'apply' means 'to be relevant to'. To me this is directed: the tag is necessary relevant to the page, but the other way round is clearly not necessary true.
Note that I agree that '"somewhat" about' goes in the right way. I'd just need this 'somewhat' to be muscled up!



> On Sun, Jan 6, 2013 at 5:05 PM, Paolo Ciccarese
> <paolo.ciccarese@gmail.com>  wrote:
>>> So unless there's a strong motivation I'm overlooking, I'd recommend a
>>> more neutral expression like "the body relates with the target". Granted,
>>> it's less informative, but at least it's not dangerous.
>> We had a discussion about this point while writing this version of the spec.
>> I am ok with having 'related' replacing 'about'. The terms are both generic
>> but 'relates' does not imply the directionality.
>> Given the example you provided I don't see alternatives.
> I have always liked the "is somewhat about" definition. It has an
> implied directionality, which for most cases makes it easier to
> determine what is body and what is target.
> This is something I always found odd in the AO specification, where
> one had ao:annotatesResource, ao:body and ao:hasTopic - but it was
> there in particular to keep this distinction - classification would
> use hasTopic instead.
> My considerations back then in
> http://www.wf4ever-project.org/wiki/display/docs/2011-09-26+Annotation+model+considerations#2011-09-26Annotationmodelconsiderations-AnnotationOntology%28AO%29
>>  From this example above (using aot:Qualifier) one could strictly argue that for our annotation bodies, AO should be applied 'opposite' to how we used OAC, as the annotation bodies have the aggregated resources as their topics. We feel that this is somewhat counter-intuitive, as our motivation was to find a mechanism for attaching rich descriptions to aggregated resources. However, AO encourages specialisation through subclassing ao:Annotation, for instance an aot:Note relates an ann:body as a free-text note describing (a sub-selection of) the annotated document.
> With "relates to" it gets quite blurry. I think the classification
> example is the odd one out - and we have argued earlier to use
> something like oa:semanticTag instead of oa:hasBody for that purpose.
> So if Antoine case is "someone tags a web page with its subject", that is
> not classification, perhaps it is identification.
> I would, if I follow the current draft strictly, do this as:
> :ann1 a oa:Annotation ;
>    oa:hasBody<http://en.wikipedia.org/wiki/List_of_Presidents_of_the_United_States>
> ;
>    oa:hasTarget<http://dbpedia.org/resource/Bill_Clinton>  .
> And even for pictures:
> :ann2 a oa:Annotation ;
>    oa:hasBody<http://en.wikipedia.org/wiki/File:Bush_and_Clinton.jpg>  ;
>    oa:hasTarget<http://dbpedia.org/resource/Bill_Clinton>  .
> We can't say it the other way, because
> <http://dbpedia.org/resource/Bill_Clinton>  is not (in this case)
> "somewhat about"<Bush_and_Clinton.jpg>.
> I know however this reads counter-intuitive - we feel that the
> annotator should be "annotating the jpeg" above - not "annotating the
> former president".   We might above also do an annotation with both
> resources as oa:hasTarget and no body - but that does not say much,
> not without an appropriate motivation.
> With Antoine's "is related to" definition then this annotation could just
> as well been written both ways - so I'm not sure how this would help
> clarify the directionality, just open it for more confusion.
> It might also help if the appropriate motivations can help to relate
> the body and target, for instance oa:Classiciation does not state
> clearly where we can find the classification, and oa:Tagging how to
> find the tag. The use-case here is not classification, as
> <http://dbpedia.org/resource/Bill_Clinton>  is not a classification
> type, but rather an Identification.
Received on Tuesday, 8 January 2013 23:01:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2013 23:01:39 GMT