It would be great if others on this list could chime in with agreement
or disagreement on Karen's position so that she can represent the
opinion of this group with confidence. My own view:
+1 completely agree re: "if you accept closed world OWL then you might
as well throw out RDF/RDFS/OWL standards altogether".
I completely agree that overloading RDF/OWL concepts risks both
confusion and the erosion of meaning. I do not buy the argument that
overloading existing terms is easier than learning a new set of terms.
On the contrary, learning new terms sounds easier to me than trying to
get one's head around subtle distinctions between variant meanings of
existing terms.
As Danbri likes to say, what people do in the privacy of their own
databases is their own affair. People can choose to interpret OWL for
their own purposes with CWA. It would IMO be very counterproductive
for W3C to specify alternative global semantics for OWL.
Tom
On Fri, Jan 23, 2015 at 11:58:09AM -0800, Karen Coyle wrote:
> On 1/23/15 11:08 AM, Bernard Vatant wrote:
> >Beyond the name, grounding a language which will be mostly use for
> >validation in closed worlds, upon RDFS which relies on the open world
> >assumption, might be at risk of muddling those waters a bit more.
>
> Bernard - exactly.
>
> This is indeed one of the underlying issues that the group has not
> yet resolved. Some members of the group are advocates of the
> closed-world interpretation of OWL (e.g. ICV/Stardog), so some
> closed-world re-use of RDF/S seems mild in comparison. The response
> from these people to those who prefer something entirely distinct
> from RDF and OWL is that it requires learning an entirely new set of
> terms. Those opposed to over-loading cite potential confusion and
> erosion of meaning.
>
> So we have the "over-loaders/less learning curve" vs. the
> "no-over-loaders/don't muddy the water" philosophies going
> face-to-face.
>
> If DCMI has a preferred position, we should put it forth, as members
> of the W3C working group. My own gut tells me that overloading is
> quite dangerous, but I don't know how others feel. This is an
> appropriate architectural discussion, I would think.
>
> kc
> p.s. To reveal my personal view without holding back, if you accept
> closed world OWL then you might as well throw out RDF/RDFS/OWL
> standards altogether and just let everyone make up rules for
> whatever suits them best. But I haven't said this in the W3C group.
--
Tom Baker <[log in to unmask]>
|