On Mon, 23 Oct 2006, Doug Tody wrote:
>
> I can see that one might want to plot a spectrum in magnitude units in
> a standard system. A good question to consider however, is whether this
> feature belongs in the user interface, where the spectrum is plotted in
> whatever units the user prefers, or in the protocol. This is similar to
> displaying RA/DEC in sexagesimal units, while we use decimal degrees in
> the protocol. This sort of conversion is easily handled in the user
> interface.
>
Unfortunately it is not easy, that is the problem. At least, I am assured by optical astronomers that for highly accurate conversion you need a raft of metadata and even to know the nature of the object. That is also the case for converting X-ray counts to physical units.
Part of the discussion which we have been having in VOTech has considered a two-level conversion. That is, if a standard magnitude system is specified, or a simple conversion factor for X-ray counts, you can convert to physical units to within a few percent accuracy. NB you also need to know the effective wavelength for the optical filter to get the spectral axis in physical units, or similar for the X-ray band.
For higher accuracy, we have discussed the possibility of metadata including a unique reference which would allow a VO query to the parent instrument archive to associate zeropoints etc. with that particular observation as required. This may happen in the near future for a few archives but will not be exhaustive (if anyone can find a web reference for e.g. plate N mag as in USNO-B, I will buy them a very large beer....) so we need to preserve the possibility of passing data in these bizarre units and hope taht tools will allow an easy manual option for specifying conversion factors.
cheers
a