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

Re: Semantic Tags (was several threads)

From: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
Date: Tue, 5 Feb 2013 16:11:12 +0000
Message-ID: <CAPRnXtn=NSeoSt0Rs-2fY-_7OpaO4P2ry+ncnDFhhmPHeG3Mxw@mail.gmail.com>
To: Raphaël Troncy <raphael.troncy@eurecom.fr>
Cc: Robert Sanderson <azaroth42@gmail.com>, Paolo Ciccarese <paolo.ciccarese@gmail.com>, Antoine Isaac <aisaac@few.vu.nl>, public-openannotation@w3.org
On Tue, Feb 5, 2013 at 3:10 PM, Raphaël Troncy
<raphael.troncy@eurecom.fr> wrote:
> Perhaps I will say something that you have already discussed at large, but
> the more I think, and the more I'm under the impression that you want to
> re-invent ctag:Tag with oa:SemanticTag. Am I wrong?

I don't think you are wrong, but I would call it more like "Perform
ctag functionality within the common frame work of Open Annotation

ctag, while a very beautiful and simple model (
http://www.commontag.org/Specification ), would in the light of OA
also have big overlaps with the more general OA model.

For instance - http://www.commontag.org/QuickStartGuide shows how to
tag sections, paragraphs etc - which we have more general solutions
for with selectors.

With OA we give a common approach to do annotation, be it tagging,
comments, semantic identification, highlighting an image, etc.  I
don't see why, when you look for annotations of a resource, you should
then not be able to see the semantic tags described in OA?

Using most ctag terms:

<target1> ctag:tagged <anno1> .

<anno1> a ctag:Tag, ctag:AuthorTag ;
    ctag:means <term1> ;
    ctab:label "term1" ;
    ctag:taggingDate "2013-01-28T12:00:00Z" .

Using OA, here doing a (valid) "dual text/semantic tagging" as in the
ctag example:

  <anno1> a oa:Annotation ;
    oa:motivatedBy oa:tagging, oa:identifying ;
Using ctag:

<target1> ctag:tagged <anno1> .

<anno1> a ctag:Tag, ctag:AuthorTag ;
    ctag:means <term1> ;
    ctab:label "term1" ;
    ctag:taggingDate "2013-01-28T12:00:00Z" .

    oa:hasBody <term1> ;
    oa:hasTarget <target1> ;
    oa:annotatedBy <agent1> ;
    oa:annotatedAt "2013-01-28T12:00:00Z" ;

  <term1> a oa:SemanticTag, cnt:ContentAsText ;
      cnt:chars "term1" .

So clearly the ctag version is easily expressed in OA, and a mapping
should be possible to form quite easily;

(forgive my not-quite-Manchester syntax below, I know it's
embarrassing considering my work place!)

ctag:Tag subclassof oa:Annotation
ctag:tag subclassof [ oa:motivatedBy oa:tagging ], [ oa:motivatedBy
oa:identifying ]
ctag:means SubObjectPropertyOf oa:hasBody
ctag:means rdfs:domain oa:SemanticTag
ctag:tagger SubObjectPropertyOf (inverse oa:hasTarget)
ctag:taggingDate SubObjectPropertyOf oa:annotatedAt
(ctag:label would need a property chain, although that's tricky with literal)

The only ones not covered are the specializations AuthorTag,
ReaderTag, AutoTag, which we would have to explain as a more elaborate
chain of the agent's type and relation to the document.

ctag:means is a bit stricter than our tagging. We have different
motivations you could use, oa:tagging is quite loose, oa:identifying
is probably closer to what ctag means (and so I've included it above).

However there is no direct reference to <agent1> in ctag, which is a
bit odd, we can however say something about the role of <agent1>,
using ctag:AuthorTag. I guess regular dcterms:creator,
prov:wasAttributedTo or pav:authoredBy could work on the ctag:Tag.

Note above that a ctag:Tag is not equivalent with a oa:SemanticTag -
as the ctag:Tag represents that particular tagging rather than the tag
itself. Our semantic tag is what is in the end of ctag:means, which is
untyped in ctag.

Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester
Received on Tuesday, 5 February 2013 16:11:59 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:22:03 UTC