> describing an observation, rather than
modelling the actual measurements?
Martin
As we've currently approached it, the model can describe the
actual measurements to some level. Right now there's an ObsData
part of it (which is, you'll be *delighted* to know, thought
of as a Q instance) to hold the actual data etc. I expect that
in typical use, at least initially, the real data bytes will
be in a FITS file and we'll just use ObsData to point to that,
and there will be some evolution as more and more of the metadata
is used in XML - but for now the model is trying to at least model
a place for the data itself. Our focus, though, should be on
the metadata.