On 3 Nov 2009, at 18:37, Nathan wrote:
Alexandre Passant wrote:
Hi Nathan,
On 3 Nov 2009, at 18:16, Nathan wrote:
Hi All,
Hoping for a little bit of guidance here on tagging & assigning
subjects to content etc - I can't quite grasp how to describe what
an item of content is about; particularly in the context of a
normal blog post and with relation to tags/subjects/moat/commontag/
scot etc.
In short I've build a little mashup of a few services and some
linked data which extracts terms & subjects from an item of
content; and now I'm unclear of which ontologies to use.
The info I can extract is "tag string" and mainly a dbpedia uri
for the tag (to give it real meaning I guess)
example..
string: Nuclear program of Iraq
URI: http://dbpedia.org/resource/Nuclear_program_of_Iran
also bearing in mind that I'll typically have 5-10 of these per
"post".
On the face of it I'd assume I should be using the following for
each "tag" and leaving the string literal value out of the triples
altogether
http://purl.org/dc/terms/subject
http://www.holygoat.co.uk/owl/redwood/0.1/tags/taggedWithTag
however, with MOAT/CommonTag/SCOT (and no doubt others) added in
to the equation I'm totally lost as which is the most fitting and
widely recognised for tagging content in this manner; is it worth
adding something to say that it was automatically tagged by a
machine? or including the string literal value of the tag(s)?
SCOT does not directly address the issue of 'tag meaning' but focus
on modeling tagclouds and making them interoperable.
MOAT and CommonTag serve the same general purpose (defining what
tag means, in terms of URIs) so you can use whatever you like -
however, CommonTag is indexed by SearchMonkey so that is a clearer
advantage for it and I'd then suggest to use that one if you
develop an app on the Web.
A few differences between them however so far (it may evolve in the
future, with ongoing work on CommonTag)
- CommonTag provides ways to make the difference between
ctag:AuthorTag, ctag:ReaderTag and ctag:AutoTag while MOAT just
make the difference between manual and auto-tag.
- MOAT models the "tagging action" (i.e. tri/quatri-partite model,
based on - and extending - the Tag Ontology) and 'global
meanings' (that can be used if you want to setup a tag server that
deliver URIs / meanings for each tags, e.g. in a company.)
Hope that helps,
cheers, it does.. but also leaves me thinking I need to be using:
dc:subject
tag:taggedWith
ctag:means
moat:tagMeaning
surely this is an issue if they're all essentially the same?
and leads me to a further question.. is there any way to express
that [dc:subject tag:taggedWith ctag:means moat:tagMeaning] are all
equal?
They are actually not the same.
The relationships ctag:means and moat:tagMeaning are used to define
links between a tag and its meaning, not for linking the tagged
resource to the meaning of the tag.
For that direct relationship , ctag:isAbout is the appropriate
relationship (I'm just realizing it's not in the doc but in the
ontology only [1]).
There is also moat:taggedWith that serve a similar purpose.
In addition, tag:taggedWith is there to link a resource to a tag, not
to the URI that serves as a meaning for this tag.
Finally, regarding dc:subject, a tag can be used not as a subject
(think of a webpage tagged "cool" or "todo", they are probably not
used as subject) so the semantics of dc:subject is probably not what
you want here.
However, this property can be enough if you know that the tags used
are here as subject / topics.
Best,
Alex.
thanks again,
nathan
--
Dr. Alexandre Passant
Digital Enterprise Research Institute
National University of Ireland, Galway
:me owl:sameAs <http://apassant.net/alex> .
