W3C home > Mailing lists > Public > public-sdw-wg@w3.org > March 2017

Re: Using rdf:Property class for properties whose URI contains string "time"

From: Maxime LEFRANCOIS <maxime.lefrancois@emse.fr>
Date: Wed, 29 Mar 2017 00:42:20 +0200
To: janowicz@ucsb.edu
Message-ID: <33f-58dae700-6f-4e673400@140150008>
Cc: "SDW WG Public List" <public-sdw-wg@w3.org>

If we can lower the number of properties, without preventing any use case and requirement to be fullfilled, then let's go! 

plus: it would be better for implementation evidence that: data that previously used the deprecated property now can be inferred to use the actual normative property.

best,
Maxime

Le Mercredi, Mars 29, 2017 00:23 CEST, Krzysztof Janowicz <janowicz@ucsb.edu> a écrit:
  Hi,

Yes, but as I tried to describe on the wiki page[1], this is for good reasons and we discussed them a few times several months ago. PhenomenonTime needs to be able to deal with more complex inputs.

The problem that I was trying to explain was that the currently proposed alignment axiom 'ssn:observationResultTime rdfs:subPropertyOf sosa:resultTime' is not in OWL2 DL as one is a DataTypeProperty and the other one is an ObjectTypeProperty.

The second case 'ssn:observationSamplingTime owl:equivalentProperty sosa:phenomenonTime. ' is simple because both are object type properties and equivalent anyway.

I liked Raul's proposal (if I understood it correctly) to deprecate observationResultTime and observationSamplingTime and then reuse the sosa properties resultTime and phenomenonTime in ssn without the need to do anything in addition.

Best,
Jano

[1] https://www.w3.org/2015/spatial/wiki/Time_in_SOSA_and_SSN
On 03/28/2017 03:14 PM, Maxime Lefrançois wrote:Dear all, 

If I took the minutes correctly today, some of the properties whose URI contains string "time" are object properties and other are datatype properties, so that's not really consistent. 

It has been proposed to declare them as instances of rdf:Property instead of having to choose between ObjectProperty and DatatypeProperty. 

This could be interesting, these are the side effects I can think of now:
- we would need to assert these properties are instances of AnnotationProperty, else the ontology would not be OWL DL;
- no ontology that extends SSN can assert it's also a ObjectProperty or a DatatypeProperty;
- one cannot make this property be involved in a OWL logical axiom in any possible way, apart from rdfs:domain, rdfs:range, and rdfs:subPropertyOf;
- still, people can create non-OWL rules ()e.g., SPARQL Construct or SPIN rules) that can generate new knowledge out of some pattern that involves this property.

Best, 
Maxime 
 -- 
Krzysztof Janowicz

Geography Department, University of California, Santa Barbara 
4830 Ellison Hall, Santa Barbara, CA 93106-4060 

Email: jano@geog.ucsb.edu
Webpage: http://geog.ucsb.edu/~jano/
Semantic Web Journal: http://www.semantic-web-journal.net



 
Received on Tuesday, 28 March 2017 22:42:53 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 28 March 2017 22:42:53 UTC