W3C home > Mailing lists > Public > public-rdf-shapes@w3.org > July 2014

Re: Should we say "data model"?

From: Holger Knublauch <holger@topquadrant.com>
Date: Wed, 30 Jul 2014 09:14:05 +1000
Message-ID: <53D82ABD.5080006@topquadrant.com>
To: "public-rdf-shapes@w3.org" <public-rdf-shapes@w3.org>
I don't really want to do the work of the requirements task force here 
and now. But in general, I think we all understand that OWL cannot 
represent all possible use cases, especially those that require 
"variables" that refer to the same entity in multiple places.

Holger


On 7/30/14, 8:53 AM, Kendall Clark wrote:
> Can you give some examples of these "more complicated constraints" 
> that OWL cannot represent?
>
> On Tuesday, July 29, 2014, Holger Knublauch <holger@topquadrant.com 
> <mailto:holger@topquadrant.com>> wrote:
>
>     Let's wait and see what requirements this WG comes up with. Of
>     course there is a large overlap between the various Shapes
>     proposals and OWL in that they all support cardinality and range
>     restrictions. If it was just about those types then yes reusing
>     the terms (or local names) from OWL can work. However, as soon as
>     you add a few things that OWL cannot syntactically represent, then
>     you could just as well start from scratch and define a converter
>     script (or on-the-fly-conversion). It will likely be cleaner and
>     more honest to avoid confusion and have less ballast. And
>     independent of a high-level vocabulary for end users, there is
>     still a need to express more complicated constraints, and SPARQL
>     is the best available language to represent those. All we need to
>     agree on is a way to link SPARQL with RDF data models and SPIN is
>     one proposal to do that.
>
>     And yes, TopBraid also used OWL with closed-world interpretation
>     from day one, and it includes a SPIN library to interpret the OWL
>     vocabulary for constraint checking - at least for the restriction
>     subset of OWL and domains and ranges.
>
>     Holger
>
>
>     On 7/30/14, 4:00 AM, Kendall Clark wrote:
>>     Yes, it might (assuming you were addressing me) although it may
>>     be too late now. We re-used the OWL namespace mostly out of
>>     habit, not from some particular goal or aim (well, to not break
>>     existing tools like Protege which knew how to manipulate that
>>     *syntax* -- that was the main goal)... But at this point I'm not
>>     sure a different namespace is going to change anyone's mind. :>
>>
>>     Cheers,
>>     Kendall
>>
>>
>>     On Tue, Jul 29, 2014 at 1:58 PM, Paul <paul@proxml.be
>>     <javascript:_e(%7B%7D,'cvml','paul@proxml.be');>> wrote:
>>
>>         Would using different namespaces help in acceptance?
>>
>>         Kind regards,
>>
>>         Paul
>>
>>
>>
>>
>>         On 29-jul.-2014, at 19:46, Kendall Clark
>>         <kendall@clarkparsia.com
>>         <javascript:_e(%7B%7D,'cvml','kendall@clarkparsia.com');>> wrote:
>>
>>>         Yeah, I appreciate that concern. Everyone keeps telling me
>>>         that this seems like a problem in principle; apparently
>>>         we're the only ones who built it *as a real thing* and *in
>>>         practice* it's not a problem at all. Our customers don't
>>>         find it in the least bit confusing. In fact, as we
>>>         originally said, most people who wanted OWL always wanted
>>>         closed world semantics anyway, so giving it to them is a big
>>>         win.
>>>
>>>         Oh well. :>
>>>
>>>         Cheers,
>>>         Kendall
>>>
>>>
>>>         On Tue, Jul 29, 2014 at 1:42 PM, Bernard Vatant
>>>         <bernard.vatant@mondeca.com
>>>         <javascript:_e(%7B%7D,'cvml','bernard.vatant@mondeca.com');>> wrote:
>>>
>>>             Hi Kendall
>>>
>>>             I did not want to point at any specific syntax, but
>>>             since you mention it ...
>>>             Reusing OWL syntax with a closed world interpretation is
>>>             of course a seductive path (which I've been following
>>>             myself, as said before) but I've always been a bit
>>>             uneasy about it. OWA is built in the OWL Recommendation.
>>>             I would rather have a neutral language, with
>>>             non-ambiguous open world interpretation in OWL, and
>>>             another one in any closed-world language (SPIN, SPARQL,
>>>             you name it).
>>>
>>>
>>>             2014-07-29 18:07 GMT+02:00 Kendall Clark
>>>             <kendall@clarkparsia.com
>>>             <javascript:_e(%7B%7D,'cvml','kendall@clarkparsia.com');>>:
>>>
>>>                 On Tue, Jul 29, 2014 at 11:57 AM, Bernard Vatant
>>>                 <bernard.vatant@mondeca.com
>>>                 <javascript:_e(%7B%7D,'cvml','bernard.vatant@mondeca.com');>>
>>>                 wrote:
>>>
>>>                      Does that mean that we are looking for
>>>                     something (language, format, whatever) that
>>>                     could be interpreted either with the open world
>>>                     assumption to support open world reasoning, and
>>>                     (exactly the same piece) interpreted in closed
>>>                     world applications as a constraint for
>>>                     interfaces or a validation rule?
>>>
>>>
>>>                 I can't speak for anyone else, of course, but this
>>>                 is precisely what Stardog ICV does using OWL syntax
>>>                 and is (to my knowledge) the only such system that
>>>                 does. But, alas, it does not appear that there is
>>>                 consensus in the likely Validation WG to put that on
>>>                 the recommendation track. A mistake, in my view, but
>>>                 there you go. :>
>>>
>>>                 Cheers,
>>>                 Kendall
>>>
>>>
>>>
>>>
>>>             -- 
>>>             *Bernard Vatant
>>>             *
>>>             Vocabularies & Data Engineering
>>>             Tel : + 33 (0)9 71 48 84 59
>>>             Skype : bernard.vatant
>>>             http://google.com/+BernardVatant
>>>             --------------------------------------------------------
>>>             *Mondeca*****
>>>             35 boulevard de Strasbourg 75010 Paris*
>>>             *
>>>             www.mondeca.com <http://www.mondeca.com/>
>>>             Follow us on Twitter : @mondecanews
>>>             <http://twitter.com/#%21/mondecanews>
>>>             ----------------------------------------------------------
>>>
>>>
>>
>
Received on Tuesday, 29 July 2014 23:14:39 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:02:39 UTC