Date

15 Oct 2019

Attendees

Apologies:

John Gemski (insufficient notice given for this meeting)

Agenda

The agenda this week is to create a ‘Sprint’ for a number of thematically related content matters. These will include

Time / temporality concepts as needed for Ratings, Loans and others
Agree on groupings of Jiras for inter-related topics, including Address ones, Service etc. – we will aim to identify what can be achieved in the Q4 time frame and agree next actions to progress these. Decide which topics to include as sprints for Q4.
For address related topics, we will review the current Jiras, close any that are no longer applicable, consider the intended scope of FIBO for this content and therefore which of the remaining Address related Jiras are applicable.

Proceedings:

Sprints / Grouping

Grouped and created Sprints for:

Time and Temporality
Address related concepts
Goods and Services

However it turns out that Jira, unlike PM tools, does not let you specify activities to be carried out in parallel. The ‘Definitions’ Sprint needs to be finished (in Jira’s perspective) before the Time and Temporality one is started. In reality, we started on the Time and Temporality sprint today.

Time and Temporality

Initial discussion and goals setting.

Key points:

Time Interval

What matters is concepts not physics e.g. Time Interval is a concept expressed in data and management decision making (the criteria for use of an ontology), even though there is ‘no such thing’ for a purist perspective. Earlier FIBO FND conceptual work addressed this by saying that for any Time Interval, there is a related granularity, expressed in terms of some time unit. This remains in the ‘Informative’ conceptual material and is not planned for inclusion in the functional Release FIBO ontologies.

Conceptual v Physical

MB went over the history of the current FinancialDates ontology (shown on screen using CCM diagrams) and explained the distinction between a conceptual ontology and a data-focused ontology. CO will have both temporal subject matter (days, periods, instants etc.) and information constructs that have some reference relation to these (similar to, in Addresses, the distinction between address elements and geo / real estate concepts). In the case of time (not temporality), the Mark Linehan created ‘FinancialDates’ ontology should reflect only the information constructs (dates, date/time), so there is a clear boundary (we think) between conceptual and operational Time model constructs.

We will need to see how that stacks up against the requirements from Ratings – these will need to identify what if anything needs to be added, and we need to figure out how that add these within the FIBO rules whereby we reflect data elements (and thereby, information constructs), not conceptual matters.

Noted in passing that one of the concepts in the FinancialDates ontology looks a bit conceptual.

The Ratings requirements may include phase and status.

Temporality is a matter of tenses (past, present, projected futures). We modeled this fairly naively in conceptual FIBO, though subsequent readings from the academic ‘Applied Ontology’ community (Guarino et al) suggest we were not far off the mark. Applying the ‘Principles of Best Practice’ for FIBO OWL resulted in the removal of ‘Value’ from the ‘Parameter’ v ‘Value’ pair of high level concepts and the partial conflation of these concepts. That principal, while not yet documented, needs to be reflected in the additional work we do or bring forward for Ratings in terms of tenses, phases and the like.

We will continue with Time and Temporality next week, starting with the detailed Loans and Ratings requirements.

Address

JB: We should look at the LEI Ontology – this includes a significant number of address related concepts including e.g. headquarters. Also deals with mailing label (line 1 etc.) parts, and other parts of the address, structured according to specific semantics e.g. country, postal code, as well as optional ones like floors.

This will soon be on an OS GitHub.

Action: JB to send in the RDF and TTL stuff on that.

FIBO: ‘Address is an index to a location’ (a pointer to something). This may be physical or virtual.

Any information constructs that refer to some physical OR virtual thing. Places where information is delivered (PO boxes; virtual addresses). It is an end point for some communications.

This over-arching pattern applies to postal addresses, to computer network addresses, to geo-coordinates and to others e.g. the Three Little Words system. Note that for IP addresses, there is a possibility that a particular end user uses dynamically allocated IP addresses, in which case the end point of that IP address is not a specific machine for all time. IT remains the case that the address identifies somewhere that is the end point for some communication.

Actions: Eliminate address Jiras that are now moot.

Use outcomes of Temporality as a guide as to where to draw the line between conceptual things and information constructs.

With Time this will be simpler because it is one dimensional.

Decisions:

Action items

Tahoe Blue JB to send in the RDF and TTL stuff on GLIEF Ontology
Mike Bennett Eliminate Address Jiras that are now moot

Browser not supported