OWL Time Ontology adoption

The tables below list a number of ontologies and datasets that utilize elements from OWL-Time. The list is not exhaustive.

Original OWL-Time features

The original version of OWL-Time was published as a W3C Note in 2006, and has been recognised as the primary basic time ontology by many parts of the community. However, universal adoption of OWL-Time, alongside core vocabularies and ontologies like FOAF, Dublin Core, OWL, PROV-O, SKOS and the W3C Organization Ontology, has been compromised by its uncertain status. Though there are many examples of adoption of OWL-Time, their temporal distribution is uneven, and may have diminished as the community became aware of the actual status of the 2006 note, though there was no challenge to the quality of the ontology. The current (2017) revision and standards-track activity is intended to be a long-overdue correction to this situation.

New features

The original version of OWL-Time limited the description of time position to use of the Gregorian calendar and 24 hour clock. This prevented adoption of OWL-Time in geology, archeology, cosmology, and in applications designed for other cultures. New features in this revision remove the restriction, by adding super-classes and siblings to the original features, while respecting and preserving all the original features. While the extensions were triggered by well documented needs, the specific implementation only has a few implementations. Nevertheless, the value of the new features appears to have been demonstrated.

Transition criteria

The W3C criteria for exiting 'Candidate Recommendation' status are based on

(a) number of implementations of each element of the ontology

(b) the expectation that these will include both producing and consuming applications.

For a low-level ontology like OWL-Time, the appearance of elements from the ontology are likely to be a minor feature of a specific application. Nevertheless, we can broadly distinguish between

1. ontologies which use OWL-Time classes or properties as the basis for more specific application classes or properties through sub-classing, or which use OWL-Time classes as the domain or range of property definitions, or in constraints in class definitions, and
2. data which mentions OWL-Time classes and properties in 'marked-up' datasets.

These two cases are related to the 'producing' and 'consuming' types mentioned by W3C. For that reason, in the tables below we indicate the general division between 'ontologies' and data by rendering the latter in bold.

Some OWL-Time implementations

Usage of OWL-Time features

Additional testimonials

f1: From Rob Warren warren@muninn-project.org on Rationale for keeping unused elements of Time OWL:

Currently, two items within the Time Ontology remain unused by the known implementations: hasXSDDuration and GeneralDateTimeDescription and are "at risk" features:

hasXSDDuration is defined as the "Extent of a temporal entity, expressed using xsd:duration". Since the early web had this resource to document durations, it stands that the feature is needed to reference legacy data.

Furthermore, the exemplar in section in "5.8 Legal interval" makes use of it and the interval ontology has a similar property "hasXsdDurationDescription" which is poorly defined but seems to be the related property. I think this is sufficient evidence that there is a use for this property for legacy linkages to XML.

GeneralDateTimeDescription is needed for a generalization for other calendars as well as a structural ontological "tie-together" for other time units such as Section 5.8's financial quarter "ex: hasFinancialQuarter". I know that ontologists are not too keen on leaf nodes within ontologies, but these are needed for integrations to other views of time across ontologies and datasets. I'm think here of other time-descriptions or units that aren't necessarily formalized, such as school periods, ships's bells or rugby time periods.

In short, my view is that both these properties have yet to be seen in action but are needed to support time structures that have not been previously well supported by computing systems. This may be the reason itself for the lack of use but if we expect ontologies to tackle some of the harder "human time" problems, the hooks need to be in place.

All my best,

rhw