Is a Timeline object needed for Calendars

[chris-little]

Calendars are constructed from multiple Timescales. Each Timescale is simple: nameId, epoch, clock, ticks, UoM.

A calendar has a "timeline" which is not a Timescale, but does have, generally, a monotonically increasing set of time instants. constructed from the multiple overlapping Timescales.

Does the "Timeline" need to be a separate object - which would then need the algorithm(s) and consituent Timescales? Or can we just have the "Timeline" as a Calendar attribute? In a sense, it IS the Calendar.

Or is the "Timeline" just a Notation?

[chris-little]

@cmheazel While clarifying the scope of Timescale, the multiplicity from Calendar should not be 1..* but 2..*.

If there is only one timescale, it is a Temporal CRS, not a calendar, as the normal rules of arithmetic apply!

[chris-little]

@cmheazel Two PRs now created: 35 and 38, for the two different approaches to include the concept of a timeline. I think both have all the extra typos that you trapped and incorporated into your PR#33

[chris-little]

Or is the "Timeline" just a Notation?
Of course not - a timeline could have several different notations, just like calendars.

[chris-little]

@cmheazel As the text of the Abstract Conceptual model uses the word timeline separately from timescale, there is some merit in having it as a separate object/class.
If it is an object/class, does is have various attributes - yes - similar to calendar, and it could have methods too. E.g. switch notation, change Units of Measure, extend or restrict scope, etc.
So I am in favour of the proposed PR#38 rather than #35.

[chris-little]

PR merged at Temporal DWG on 2023-08-31