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

RE: OWL-Time - ISSUE-65: General purpose temporal predicates

From: <Simon.Cox@csiro.au>
Date: Mon, 9 Jan 2017 04:51:43 +0000
To: <janowicz@ucsb.edu>, <jlieberman@tumblingwalls.com>
CC: <public-sdw-wg@w3.org>
Message-ID: <2695b79ac5924cf3a6fddeff3f5b8d70@exch1-mel.nexus.csiro.au>
Josh, others –

Clearly my response last year was not explicit enough, as the discussion appears to be muddling a few issues. Let me have another go:

1.     A class TemporalEntity was defined in the 2006 version of OWL-Time. There are a lot of properties with TemporalEntity as the domain, including hasDuration, hasBeginning, etc, and its existing sub-classes (Interval, Instant, and further specializations) have even more properties – see Figure 1 here http://w3c.github.io/sdw/time/#topology .

2.     However, the principal classes in OWL-Time (shown in this figure) _only_ deal with the position and extent of temporal entities, and the relations between them.

I had been working on the assumption that these temporal entities were like Geometries (stgeo:GM_Object) rather than Features (stgeo:SpatialThing). (For ISO Harmonized Model aficionados, note that ISO 19108 TM_Object is comparable to ISO 19107 GM_Object, and I had been assuming that <http://www.w3.org/2006/time#TemporalEntity> matched TM_Object.)

An argument for why temporal things should *not* be subclassed from temporal entities is that a single temporal entity (interval, instant, etc) can be associated with an unlimited set of temporal things (events, happenings, activities), so it is better for a temporal thing to have a temporal property whose value would be one of these temporal entities.

This is similar to the relationship between a spatial thing and a geometry.


Looking at Josh’s stgeo1 model, things are connected to geometries using the generic properties hasSpatialModel and hasGeometry etc. where

  rdf:type owl:ObjectProperty ;
  rdfs:domain stgeo1:SpatialThing ;
  rdfs:range stgeo1:STModel ;

  rdf:type owl:ObjectProperty ;
  rdfs:domain stgeo1:SpatialThing ;
  rdfs:range stgeo1:Geometry ;
  rdfs:subPropertyOf stgeo1:hasSpatialModel ;

Doing the same kind of thing for temporal attributes, you’d get

  rdf:type owl:ObjectProperty ;
  rdfs:domain stgeo1:TemporalThing ;
 rdfs:range <http://www.w3.org/2006/time#TemporalEntity> ;

This is clearly inconsistent with the axiom:

            stgeo1:TemporalThing owl:equivalentClass <http://www.w3.org/2006/time#TemporalEntity> .

Rather, I’d say

            stgeo1:TemporalThing owl:equivalentClass stgeo1:SpatialThing .

(or maybe a common superclass STThing?) and maybe also

            <http://www.w3.org/2006/time#TemporalEntity> rdfs:subClassOf stgeo1:STModel .

(that would be if we are going for a unified space-time model ...)


From: Krzysztof Janowicz [mailto:janowicz@ucsb.edu]
Sent: Friday, 30 December, 2016 03:05
To: Joshua Lieberman <jlieberman@tumblingwalls.com>; Cox, Simon (L&W, Clayton) <Simon.Cox@csiro.au>
Cc: public-sdw-wg@w3.org
Subject: Re: OWL-Time - ISSUE-65: General purpose temporal predicates

Great discussion.

So I suppose the question to settle is just what is a TemporalEntity as distinct from a Duration or a TimePosition?

As a starter, a TemporalEntity may have an unknown duration or time position.  It can also have multiple durations which is something that I often see in the digital humanities domain where it is used to model uncertainty,  e.g., to distinguish an earliest start and end from a most likely start and end.


On 12/28/2016 02:19 PM, Joshua Lieberman wrote:
Whether  TemporalEntity has subclasses is not necessarily a problem. GF_FeatureType also has subclasses. But the spatial extent of a GF_Feature is unquantifiable without a model property such as a Geometry. It seems the temporal extent of a TemporalEntity also is unquantifiable without a property such as a Duration or TimePosition. So I suppose the question to settle is just what is a TemporalEntity as distinct from a Duration or a TimePosition?


On Dec 28, 2016, at 11:01 PM, simon.cox@csiro.au<mailto:simon.cox@csiro.au> wrote:

The class :TemporalEntity is already in the ontology. It has subclasses :Instant and :Interval. Which looks comparable to Geometry as superclass of Point, Line, Polygon etc at first glance at least. Which would make it _not_ a feature. Which is why I raised the issue.

From: Krzysztof Janowicz <janowicz@ucsb.edu<mailto:janowicz@ucsb.edu>>
Sent: Wednesday, 28 December 2016 9:08:33 PM
To: Joshua Lieberman; Cox, Simon (L&W, Clayton)
Cc: public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>
Subject: Re: OWL-Time - ISSUE-65: General purpose temporal predicates

I like Josh's argumentation and agree on the separation of a temporal entity from the time model the same way we deal with spatial entities and geometries. This would also enable us to easily support multiple temporal granularities the same way we can have multiple geometries per feature.

On 12/28/2016 01:03 PM, Joshua Lieberman wrote:
Doesn’t that depend on what a Temporal Entity “is”. The ISO - OGC feature-geometry disjunction is based on Geometry being a model property for a Feature (Spatial Thing), not the Feature itself. It seems to me that this could have a parallel in OWL-Time, with Temporal Entity being the thing with an extent in time, and Duration or TimePosition being a model property for that Entity. If that parallel is valid, then it wouldn’t be inconsistent to say that a Feature is-also-a Temporal Entity if it has a Duration or TimePosition model property.


On Dec 28, 2016, at 8:16 PM, simon.cox@csiro.au<mailto:simon.cox@csiro.au> wrote:

The primary goal of OWL-Time is to implement Allen’s temporal relations in OWL, so all the OWL-Time predicates have Temporal classes as both domain and range. For example, ‘hasBeginning’ relates a temporal entity to a temporal instant. This means that attaching timing information to any event or activity using one of these predicates implies that it _is_ a “Temporal Entity”. This would be inconsistent with the approach used in the OGC/ISO Feature Model for associating geometry with a feature, in which feature types are _nt_ subclassed from geometries, but have associations with geometries. At least that would be the argument if time is treated the same as geometry.

As there appears to be interest in standard predicates to associate timing information to events or activities, we have a problem. One solution (ISSUE-64) would be to relax the global domain constraints on the existing predicates. Alternatively, we can create some general purpose object properties, such as the following:

  rdfs:comment "Beginning of an event or activity."@en<mailto:%22Beginning%20of%20an%20event%20or%20activity.%22@en> ;
  rdfs:range :Instant ;
  rdfs:comment "Duration of an event or activity, expressed as a scaled value"@en ;
  rdfs:range :Duration ;
  rdfs:comment "Duration of an event or activity, expressed using a structured description"@en ;
  rdfs:range :GeneralDurationDescription ;
  rdfs:comment "End of an event or activity."@en<mailto:%22End%20of%20an%20event%20or%20activity.%22@en> ;
  rdfs:range :Instant ;
  rdfs:comment "Supports the assignment of a temporal entity (instant or interval) with an event or activity"@en ;
 rdfs:range :TemporalEntity ;

The slightly awkward names are because hasBeginning, hasDuration etc are already in use.
Not at all wedded to activity*. Could be event* or something else if anyone has any smart ideas.

I’ve added these to the branch here:

OTOH, some upper-level ontologies make a fundamental distinction between time-bounded entities (occurrent or perdurant) and non-time-bounded entities (continuant or endurant). If we accept this viewpoint, then we might just use the original OWL-Time predicates and accept the entailment.  I guess it depends which fundamental commitment we are willing to make.


Simon J D Cox
Research Scientist
Environmental Informatics
CSIRO Land and Water<http://www.csiro.au/Research/LWF>

E simon.cox@csiro.au<mailto:simon.cox@csiro.au> T +61 3 9545 2365 M +61 403 302 672
   Mail: Private Bag 10, Clayton South, Vic 3169
   Visit: Central Reception, Research Way, Clayton, Vic 3168
   Deliver: Gate 3, Normanby Road, Clayton, Vic 3168

The information contained in this email may be confidential or privileged. Any unauthorised use or disclosure is prohibited. If you have received this email in error, please delete it immediately and notify the sender by return email. Thank you. To the extent permitted by law, CSIRO does not represent, warrant and/or guarantee that the integrity of this communication has been maintained or that the communication is free of errors, virus, interception or interference.

Please consider the environment before printing this email.


Krzysztof Janowicz

Geography Department, University of California, Santa Barbara

4830 Ellison Hall, Santa Barbara, CA 93106-4060

Email: jano@geog.ucsb.edu<mailto:jano@geog.ucsb.edu>

Webpage: http://geog.ucsb.edu/~jano/<http://geog.ucsb.edu/%7Ejano/>

Semantic Web Journal: http://www.semantic-web-journal.net<http://www.semantic-web-journal.net/>


Krzysztof Janowicz

Geography Department, University of California, Santa Barbara

4830 Ellison Hall, Santa Barbara, CA 93106-4060

Email: jano@geog.ucsb.edu<mailto:jano@geog.ucsb.edu>

Webpage: http://geog.ucsb.edu/~jano/

Semantic Web Journal: http://www.semantic-web-journal.net

Received on Monday, 9 January 2017 04:52:28 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:31:28 UTC