Enchufa2
commented
4 years ago I see this a little bit out of the scope, yes. I think this could be just another subclass playing along with {quantities}, something like c("censored", "quantities", "units", "errors"). Every method defined for "censored" would be executed in the first place, so that it can check the method, impose limits... and then forward the rest of the functionality to {quantities}.
vorpalvorpal
Thank you for this fantastic package. It's exactly what I dreamed of creating, but did not have the skill to create myself. I really feel like it should be a part of base R.
I'm not sure if this would fall into the scope of what you want for the package, or not, but a lot of measurement data has lower (and less frequently upper) detection/quantification/reporting limits. I work a lot with water sampling data, and this almost always has a lower reporting limit. Values less than this limit are reported as "below reporting limit", rather than as numerics. I realise this might be outside the scope of the package because the correct way to treat censored data is not as clear cut as units or errors.
On a related note, many measurements have a standardised method associated with them. It might be nice to treat this a bit like units, though less strictly. Possibly just giving a warning if an operation acts on two objects with different methods. Again, this may be outside of what you see as the legitimate scope of the package.