proposed changes to use range types for nzlum

[alpha-beta-soup]

For fields that already accept multiple values (commod, etc.) I've added a stipulation that they should be sorted. This is to avoid situations where a dataset has features with values of both {cattle,sheep} and {sheep,cattle} which should be equivalent.

---

For fields that take single values recording facts about sources, source_date, source_scale, I'd like to propose instead using ranges and interval notation. Single values can still be specified but we can also express a range. This is to acknoledge that sometimes classification decisions are based on many contributing inputs, from a range of publication/edit dates, and geographic scales.

To normalise across geographic data with different methods of expressing precision or accuracy, I've also proposed normalising around expressing accuracy or precision in integer CRS units as an interval (e.g. in metres, [1,10)), and not a single indicative map ratio (e.g. "1:50000")

[alpha-beta-soup]

@ChocopieKewpie what do you make of this idea?

[ChocopieKewpie]

Really like these suggestions as they offer solely practical benefits.

[alpha-beta-soup]

I'd like to run these past more people, but in the absence of a group to which I can propose these changes, I'll merge for the sake of expediency.