Re: Comments to “SE Free text tagging a Image”

On Thu, Jan 31, 2013 at 12:13 PM, Robert Sanderson <azaroth42@gmail.com>wrote:

> >> About the section:
> >>
> >> We have a section “2.1.3 Tags and Semantic Tags” and there we introduce
> >> the class oa:Tag. However, we do not use it in this example; I think it
> is
> >> missing.
> >>
> >> From the description in the section, it seems that semantic tags are
> those
> >> tags corresponding to URIs? I do not think that is always the case.
> >> “http://dbpedia.org/resource/Paris” does correspond to a semantic
> entity but
> >> “http://wikipedia.org/Paris/”  does not. I think that for semantic
> tags only
> >> URIs corresponding to semantic entities should be use, any other
> thoughts
> >> about it?
>
> Yes, Semantic Tags should be non information resources to ensure that
> there isn't a collision between someone using the same URI, but
> intending it to be treated as a document rather than a concept.
>
> > In that particular example I used a free text tag and so I am assuming
> the
> > motivation ao:tagging is enough.
>
> I would include the oa:Tag to make it clear that the text is a tag,
> and not a comment.
>
>
> > However, in general the point you raise is
> > a general issue. The current spec allows classifying a URI as semantic
> tag
> > by using oa:Tag. However, the spec says  also:
>
> > "It is NOT RECOMMENDED to use the URI of a document as a semantic tag,
> as it
> > might also be used as a regular Body in other Annotations which would
> > inherit the oa:Tag class assignment. Instead a new URI should be created
> and
> > linked to the document using an ontology appropriate to the situation."
> >
> > Which I believe it is addressing what you are asking.
> > The problem arises when somebody uses document URIs as tags (not that
> > uncommon).
>
> I'm a little more optimistic. But without inventing lots of new
> predicates (like we had with hasSemanticTag previously) and then also
> for the multiplicity constructs, we can only do our best to recommend
> good behavior.  Hopefully also there won't be many collisions, as it
> seems unlikely that the document being used to tag something would
> also be used as the body of an annotation where it was meant as a
> comment or description.
>
>
> >> One question not necessarily related to tags. Could I use annotations to
> >> say that a 3D version of that image can be retrieved from “PDB link” in
> >> format “XXX”? If yes, how could that being expressed in OA? In general,
> the
> >> question is how to express links to other resources with annotations?
> >
> > Rob, was that the usage of the oa:linking motivation we removed?
>
> We could certainly reintroduce it.    I'm in favor of having more
> motivations, rather than fewer, to prevent future collisions where
> multiple communities all mint something rather basic.  If I recall
> correctly, it was seen as too similar to oa:annotating, which we now
> don't have any more.
>
> How about:
> oa:linking  The motivation that represents an untyped link to a
> resource related to the target.
>
>
I am good with reintroducing it. Meanwhile I also would like to see a
concrete example in the Cookbook about it because I think the problem is
also to characterize the link.
I would try to start by representing what Layla outlined: a 3D version of
that image can be retrieved from “PDB link” in format “XXX”

Paolo