
Very impressive work, by a team led by Mark Linehan of IBM Research, on Date Time was presented by NIST’s Ed Barkmeyer at OMG this week. This is a “work in progress” for likely completion later this year but clearly could play an important role in temporal operations in event processing.
To quote from the latest draft, the objectives of this work are:
1. Provide a Standard Business Vocabulary for Date and Time Concepts. Provide a vocabulary of date and time concepts that business users can share and exploit in their business domain vocabularies and rules…
2. Support Machine Reasoning about Time. Provide a formal ontology that enables machine interpretation and reasoning…
3. Enable implementation…
It covers a time infrastructure (intervals, Allen Relations, durations and SBVR “states of affairs” such as events and situations), organizing time in calendars, “indexical time concepts” such as the meaning of “now” etc in a business context, and so forth. Indeed it seems the only thing NOT covered are the ad hoc adjustments of “leap seconds” to a “year” that are made every now again. And there are versions for UML, SBVR and CLIF, and plans an ODM / OWL version.
3 Comments
If it was meta-circular for time they’d need to be able to represent concepts such as “…likely completion later this year”.
Matthew – for sure there is a “fuzzy reasoning” aspect to most business references to time. Having a good time-and-date logic on which to base fuzzy reasoning is however a good thing
Cheers
FTR, “later this year” (“at a later time in this year”) is supported by the (draft) DateTime ontology. It is a combination of ‘indexical’ concepts — times related to ‘now’. And ‘completion at a later time this year’ is a special case of ‘event occurs in time period’. “Likely completion”, however, deals with concerns that have nothing to do with time! There was a time when we wondered …