Dimensionless units

[Ed Shaya]

Anita Richards wrote:

>This problem is caused by the hidden assumptions made by astronomers which
>would be OK i.e. quantifiable if they are always the same..  In a lot of
>astronomy 'Jy' really means either Jy/beam (or Jy/aperture or etc.), or
>Jy/source.  In that case _if_ the beam or source size is known, then Jy/Sr
>can be scaled.  A very clever VO tool might be able to deduce the beam or
>source size from other metdata (e.g. we might record the effective
>apertures of major satellite instruments).  However this will not work if
>you are trying to compare observations which are dominated by an all-sky
>background (e.g. CMB) with point sources.  I cannot think of any sensible
>way other than making the user enter a value to produce dimensional
>compatibility.  So we do need to have something like 'angle' as a
>dimension in order to recognise that Jy/Sr != Jy .
>  
>

Such hidden assumptions are devastating for automated processes.  An 
automated process would have to assume that an object with N Jy means 
either that the object is smaller than the beam or that all of the flux 
of the object has been accounted for by adding up all of the beams that 
cover the object.  If it is just a surface brightness of an extended 
object then the metadata must say Jy/beam if that is what is meant.  In 
a survey map the units are Jy/beam and I would want a human familiar 
with the project to enter that to be sure it is done correctly, not a tool.

Ed