Maybe others can comment as well, but I do think it's an important piece
of information, e.g. to determine recently popular tags.
Cheers,
Peter
François Dongier wrote:
Peter, maybe you could explain why you guys found it useful to date
tagging events in the first place. I suppose the point of it might be
that it could provide some context? If so, the date is only one aspect
of the context and probably not the richest one.
On Fri, Jun 12, 2009 at 1:50 PM, Danny Ayers <[email protected]
<mailto:[email protected]>> wrote:
2009/6/12 Toby Inkster <[email protected] <mailto:[email protected]>>:
> Lest I be accused of nonconstructive criticism, a route to
improving the
> vocab would be to properly align CommonTag with existing
ontologies by
> dropping ctag:taggedDate altogether.
>
> Of all the terms defined by CommonTag, ctag:taggedDate is
probably the
> one with least value to most publishers, so this change would
not only
> help align CommonTag with other ontologies, but also serve to
simplify
> and streamline the spec.
>
> The description of tagging *events* could then be considered an
> "advanced" use case, not directly supported by CommonTag. But
given that
> CommonTag would then be compatible with Richard Newman's
ontology, and
> MOAT, SCOT, etc, advanced users could go outside CommonTag to
add this
> extra meaning to their tags.
Makes sense to me.
While an RDFS/OWL inference based mapping between Richard's ontology
and Common Tag may not be be possible right now, SPARQL CONSTRUCT
could be an alternative.
Note also Richard's ontology allows:
<uri> tags:taggedWithTag <taguri> .
SPARQL (SELECT or CONSTRUCT) across those alongside Common Tag
taggings would be easy using OPTIONALs
Just as a little in-practice datapoint, not long ago I set up a little
proof-of-concept service [1] for pulling out del.icio.us
<http://del.icio.us> taggings into
Richard's Tag Ontology. del.icio.us <http://del.icio.us>'s RSS 1.0
feed gets the date
modelling wrong, funnily enough, so I was using XSLT on their API
(code at [2]). Although some of the string manipulation bits were
painful, the bit I decided to leave out because it was hard work was
reconciling the lists of values that could be the subject of
associatedTag.
Overall I was left with the impression that Richard's ont could use
simplifying, if it was possible to do this without breaking the
potential for maximally capturing data about the tagging event. I'm
optimistic the Common Tag mini-consortium can sort this one out :)
Cheers,
Danny.
[1] http://hyperdata.org/taglia/
[2] http://n2.talis.com/svn/playground/danja/taglia/
--
http://danny.ayers.name
