[QUANTITY] Data Model for Quantity v0.5 - inheritance vs aggregation
gtr at ast.cam.ac.uk
Mon May 10 09:12:21 PDT 2004
There seems to be a conflation of structural stuff about a quantity - its
value, its error bounds, its unit - and contextual stuff such as UCDs. I think
it does help to define the structural bits as a standard structure - i.e.
something for which IVOA controls the schema. If that looks like a class,
then that's fine: it's a piece of data model so it's a noun, not a verb.
I'm not sure whether having the meaning of the quantity inside the Quantity
structure really works. If you want to denote the meaning by aggregating
metadata - e.g. UCDs to the quantity, then you probably need to do this. The
alternative is always to use named types derived from Quantity in data models,
s.t. the name of the type implies the meaning. To put it another way, if we
want to go on having UCDs, then we need somewhere to keep them and inside a
Quantity seems like a good place.
More information about the dm