Generalize the properties :numerator, :denominator, :multiplier, and :multiplicand

[DanCarey404]

The domains and ranges applied to the properties :numerator, :denominator, :multiplier, and :multiplicand (i.e., :RatioUnit, :ProductUnit, and :UnitOfMeasure) leave them more constrained than they perhaps ought to be.
Also, :mutliplier and :multiplicand should be inverses.

[uscholdm]

:mutliplier and :multiplicand should be inverses.
I don't think so. Commutativity is present, but that about argument order, not direction of the link. This is an easy mistake to make.

[uscholdm]

Can you give example of how the domains/ranges are too constrained? What do we potentially need that they disallow?

[philblackwood]

When the order is important, it would be easier to remember factor1 and factor2.

[rjyounes]

From issue review meeting 2021-07-22:

When would the distinction between multiplier and multiplicand be significant?

Why are they limited to ProductUnits?

Similarly, why are numerator and denominator limited to units of measure?

Deferring discussion for feedback from @uscholdm.

[rjyounes]

multiplier and multiplicand are not inverses, as @uscholdm points out. Both predicates relate one of the multipliers to the product, not to one another. It's analogous to the relationship between fromAgent and toAgent.

[uscholdm]

When would the distinction between multiplier and multiplicand be significant?

I don't know if it would ever matter, I presume this is just borrowed from mathematics which uses these (albeit rather obscure) terms. Factor1 and Factor2 would might be easier as Phil suggests. If we wanted to be analy consistently, we would use divndent and divisor rather than numerator and denominator, but lets forget that bad idea.

Why are they limited to ProductUnits?
Because they only make sense when you multiply two units which is what a ProductUnit is.

MORE IMPORTANTLY: Having spent untold hours painstakingly using multiplicand/multiplier/numerator/denominator to create many dozens of units, so far it has never paid any dividends. (so to speak).

[uscholdm]

multiplier and multiplicand are not inverses, as @uscholdm points out. Both predicates relate one of the multipliers to the product, not to one another. It's analogous to the relationship between fromAgent and toAgent.

Excellent analogy

[rjyounes]

We have a client who is using numerators and denominators for ordinary fractions, and has to define them because the gist predicates are overly narrow.

[uscholdm]

We have a client who is using numerators and denominators for ordinary fractions, and has to define them because the gist predicates are overly narrow.

I see. So relaxing the domain/range makes sense. You could add domain(range)Incldes annotations.

We made formulas for one of our clients (@sa-bpelakh do you recall whether we used numerator/denominator for that?)

[rjyounes]

Boris: numerator and denominator are needed for complex units of measure. owl:rational can be used in other cases.

DECISION: Remove domain and range from all four predicates.