Add MonetaryPerDuration and CurrencyPerDurationUnit classes. Fixes #846

[rjyounes]

Fixes #846.

[dylan-sa]

One thought on the CurrencyPerDurationUnit formal def, but otherwise looking good to me.

[dylan-sa]

Can ratio units have numerators or denominators w/ more than one unit of measure? If not, is this a case where an allvaluesFrom + cardinality = 1 construction would better capture the meaning? (Maybe the cardinality restrictions would be better placed in the RatioUnit definition; and if so, maybe this should be raised as a separate issue.)

allValuesFrom + cardinality = 1 version:

gist:CurrencyPerDurationUnit
	a owl:Class ;
	owl:equivalentClass [
		a owl:Class ;
		owl:intersectionOf (
			gist:RatioUnit
			[
				a owl:Restriction ;
				owl:onProperty gist:hasDenominator ;
				owl:allValuesFrom gist:DurationUnit ;
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasNumerator ;
				owl:allValuesFrom gist:CurrencyUnit ;
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasNumerator ;
				owl:cardinality 1 ;
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasDenominator ;
				owl:cardinality 1 ;
			]
		) ;
	] ;

[uscholdm]

@dylan-sa @rjyounes

Can ratio units have numerators or denominators w/ more than one unit of measure?

Not sure what you mean. For example, if the denominator is a duration unit of one hour, then as stated you are asking if the unit of one hour itself has more than one unit of measure. But a unit of measure doesn't generally have even one unit of measure, it IS a unit of measure. So maybe you are asking whether a ratio unit can have more than one denominator (or numerator). I think the answer should be no.

[dylan-sa]

So maybe you are asking whether a ratio unit can have more than one denominator (or numerator). I think the answer should be no.

That is the core of it. If it doesn't make sense for a ratio unit to have more than one numerator or denominator, we could codify that in the formal definition of gist:RatioUnit. This is what I have in mind:

# Amend definition of gist:RatioUnit - switch someValuesFrom to allValuesFrom and add cardinality constraints

gist:RatioUnit
	a owl:Class ;
	owl:equivalentClass [
		a owl:Class ;
		owl:intersectionOf (
			gist:UnitOfMeasure
			[
				a owl:Restriction ;
				owl:onProperty gist:hasDenominator ;
				owl:allValuesFrom gist:UnitOfMeasure ; # Change from someValuesFrom to allValuesFrom
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasNumerator ;
				owl:allValuesFrom gist:UnitOfMeasure ; # Change from someValuesFrom to allValuesFrom
			]
                        [
				a owl:Restriction ;
				owl:onProperty gist:hasDenominator ;
				owl:cardinality 1 ; 
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasNumerator ;
				owl:cardinality 1 ;
			]
		) ;
	] ;

# Amend proposed definition of gist:CurrencyPerDurationUnit

gist:CurrencyPerDurationUnit
	a owl:Class ;
	owl:equivalentClass [
		a owl:Class ;
		owl:intersectionOf (
			gist:RatioUnit
			[
				a owl:Restriction ;
				owl:onProperty gist:hasDenominator ;
				owl:allValuesFrom gist:DurationUnit ; # Change from someValuesFrom to allValuesFrom
			]
			[
				a owl:Restriction ;
				owl:onProperty gist:hasNumerator ;
				owl:allValuesFrom gist:CurrencyUnit ; # Change from someValuesFrom to allValuesFrom
			]
		) ;
	] ;

[uscholdm]

That is the core of it. If it doesn't make sense for a ratio unit to have more than one numerator or denominator, we could codify that in the formal definition of gist:RatioUnit.

Yes, that would make sense. Note that it's possible to have two equivalent units with different numerators and denominators. For example, there might be a currency unit for 100USD and a DurationUnit of 100 Days. The Ratio unit 100USD/100Days is equivalent to the ratio unit USD/day, but each has different numerator and denominator. But even here, a single unit still has only one numerator and one denominator.

[rjyounes]

@uscholdm @dylan-sa This looks correct, but this is not the place to redefine RatioUnit. Are these definitions correct in that they follow current practice in defining magnitude and unit classes? This is what needs to be reviewed. It will cause confusion if one definition diverges from the others: it will make people start wondering why this one is different.

Please feel free to open another issue for the larger topic.

[rjyounes]

This was closed due to technical issues related to branch names. It's been recreated as #856.