Re: reasoning with DCAT

Note that if properties are defined without domains and ranges, then it
should be possible to add those to the property when it is included in
an application profile. I agree that reasoning is not commonly used, but
some applications make use of domains and ranges in their validation
routines. (I won't go into whether that is a "proper" use of D&R or not;
just that one commonly sees this kind of use of the declarations.)

This therefore may become a requirement on profiles that has not been
made explicit.

kc

On 3/21/18 10:30 AM, Annette Greiner wrote:
> Yes, I think reviewing them is fine and wise. So far I haven't seen
> anything that makes me worry, just the statement about reasoning in
> general.
> -Annette
> 
> 
> On 3/20/18 9:55 PM, Simon.Cox@csiro.au wrote:
>>> really don't need to be bound to a particular class - dcat:theme and
>>> dcat:keyword.
>> What I mean is, I can't see anyone relying on an entailment like "this
>> resource has a dcat:keyword therefore it must be a dcat:Dataset", right?
>>
>> -----Original Message-----
>> From: Simon.Cox@csiro.au [mailto:Simon.Cox@csiro.au]
>> Sent: Wednesday, 21 March, 2018 15:23
>> To: amgreiner@lbl.gov; public-dxwg-wg@w3.org
>> Subject: [ExternalEmail] RE: reasoning with DCAT
>>
>> Hi Annette -
>>
>> Thanks for your comments. It's good that my provocation triggered
>> someone. The whole reasoning thing - which was one of the big selling
>> points of semweb early, has fallen into some disrepute. But it's good
>> to hear that there are still applications looking to leverage it.
>>
>> The point of the UC is to explain the motivation for doing a wholesale
>> review of global domain/range constraints. It is not to motivate a
>> wholesale elimination of global domain/range constraints. So we can
>> evaluate them one at a time, retain the ones that the group thinks
>> matters, and move the ones that we think are unhelpful into a separate
>> graph (where people who want DCAT 1.0 axiomatization can still access
>> them [1]).
>>
>> I recorded my 'votes' against a bunch of them today, and went with 4
>> 'no change' vs 2 'relax the domain constraint'. Both of the latter
>> relate to what are essentially _annotations_ which really don't need
>> to be bound to a particular class - dcat:theme and dcat:keyword.
>>
>> - Simon
>>
>> [1] e.g. https://github.com/w3c/dxwg/blob/gh-pages/dcat/rdf/dcat10.ttl
>>
>> -----Original Message-----
>> From: Annette Greiner [mailto:amgreiner@lbl.gov]
>> Sent: Wednesday, 21 March, 2018 09:19
>> To: public-dxwg-wg@w3.org
>> Subject: reasoning with DCAT
>>
>> Sorry, this doesn't really fit into one single github issue.
>>
>> One sentence in the use case about relaxing axiomatization concerns me.
>> I know we just voted on it, but I didn't get to read it carefully
>> until just now. We are saying "...we are not aware of any catalogs in
>> which reasoning is an important part of their operation."
>>
>> That statement worries me, since part of the potential value of a
>> vocabulary is to enable reasoning if someone wants to build an
>> application that does that. I may even have a counterexample use case.
>> I've been working on an application that works on log data from
>> supercomputing systems, where the various log files are described
>> using DCAT. We have logs for a wide variety of subsystems, like
>> networking, node operation, job scheduling, boot cycles, etc. We want
>> to be able to reason about relationships between these different log
>> sets (e.g., if a network link went down, we want to find log data
>> about the nodes on the ends of that link.) At this point, I'm not sure
>> that our use case requires us to keep stricter axiomatization than the
>> group is planning to preserve, but it's probably worth thinking through.
>>
>> -Annette
>>
>> -- 
>> Annette Greiner
>> NERSC Data and Analytics Services
>> Lawrence Berkeley National Laboratory
>>
>>
> 

-- 
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
m: 1-510-435-8234 (Signal)
skype: kcoylenet/+1-510-984-3600

Received on Saturday, 24 March 2018 23:48:20 UTC