Recurring on Temporal Seq Member in OWL Time

Question

A paper about an extension of the OWL Time-Temporal Aggregates Ontology-details how to represent recurring event using OWL Time.

I am unsure about the representation of a particular recurring event. This is because it is unclear about whether a TemporalSeqMember can be itself a TemporalSeq. Secondly, I do not think that I correctly understood the essence of TemporalSeqMember

I have two questions:

1. What is the usefulness of TemporalSeqMember
2. Is the following modelling the right way to model represent a recurring event which happens yearly, in month April & May in the first 1,2 hour, i.e. at 00:00-00:59 and 1:00-1:59 respectively.

Churchwell

Answer 1

A CIDOC model of recurring events would require both a TemporalSeq (a recurring time span) and a disjoint persistent concept to define the observation of the second occurrence as related to the first. It seems that OWL is defining event frequency as a single recurring pattern of time as a single event aggregate, that contains sub events.

The endurant concept of the measured periodicity of an event cannot be modeled with TemporalSeq aggregation and seem to extend outside of TemporalThing completely. Your reference paper uses holiday as an example and I actually have a problem with modelling "Holiday" as a recurring event. "A holiday" is a recurring event in a temporal sequence, but it is also a concept. Every time "a holiday" occurs it is a new event, that event is part of a aggregate temporal sequence of events. The recurrence pattern of the instances of "holiday" is a persistent concept called "Holiday" that cannot be reasoned as a temporal sequence or event.

Nostalgia

Answer 2

From the paper that you linked to, I think one of the important parts is (emphasis added):

In order to encode the temporal aggregates ontology in OWL, we first defined temporal sequence. It has only one optional property hasMemeber [sic] which maps from a temporal sequence to any temporal thing.

So, it can have any temporal thing as a value. The question is whether TemporalSeq is actually a TemporalThing or not. The snippet shown in the paper doesn't show it as a subclass of TemporalSeq, and the ontology link appear dead. However, I found what I think is another copy at http://ontology.ihmc.us/temporalAggregates.owl. It doesn't appear to make TemporalSeq a subclass of TemporalThing (renamed to TemporalEntity?), either, so I don't think that TemporalSeq is supposed to be a TemporalThing/TemporalEntity.

Devi

thanks for your answer, according to you, if you ever modelled recurring events using the temporal aggregates ontology, is the way I modelled it in the question the proper way ? – Churchwell

I haven't used that ontology, so I'm not any kind of authority on its usage, but yours looks like a coherent and plausible use, at least at a first glance. – Devi