Should we define a quantity Arbitrary or is Dimensionless enough?

[keilw]

#173 and #59 point to missing UCUM units like ARBITRARY_UNIT. The UCUM catalog defines these as quantity type "arbitrary", should we add a quantity type Artitrary or is Dimensionless enough?
There are other types like "arbitrary biologic activity" so the question is, if those could all use "arbitrary" or should have their own types often used for just one or two units? Or do we stick to Dimensionless for all of them?

[alexanderkiel]

I have laboratory data that uses the unit [arb'U]/mL for values of with LOINC 16135-6 "Beta 2 glycoprotein 1 IgG Ab [Units/volume] in Serum". Would it be possible to map that to ARBITRARY_UNIT? As I understand a simple entry in UCUMFormat_CS.properties would be sufficient. Can I somehow add such units on myself without editing this project files directly?