Simon,
NcML (i.e. the netCDF Mark-up Language) is already able to formalize and
encode multidimensional gridded-based datasets. We developed the ncML-GML in
order to mediate ncML and GML and map these more general structures into a set
of "simple" coverages, according to ISO 19123 model.
Cheers,
---Stefano
Ron - I agree with you on Coverages, conceptually. There can be no
argument with this.
I spent some time the week before last with a team who are interested in
the variation of Gibbs free energy of a chemical species as a function
of temperature and pressure.
This is conventionally represented as a grid or map of values in P-T
space, which is structurally identical to a gridded coverage.
My point is that generalizing in this way has significant implications
to our claims that GML is an implementation of ISO 19123.
19123 is specifically limited to spatio-temporal domains, conventionally
defined.
My point really is that we need to consider whether we really want the
struggle of making "coverage" the generalized class,
or perhaps invent "functor" or "map" or "function" for this purpose, and
leave coverage alone, as originally defined.
Simon
