[Geotools-devel] Specialized Coverage and Domain Objects

[Bryce L N. <bno...@fs...>]

Nothing is ever simple.

There are at least two areas in which ISO 19123 performs "field hiding" by
overriding an attribute in a child with a type totally unrelated to the
type possessed by the attribute in the parent.  Worse, the "unrelated
types" chosen by the authors of 19123 take their carefully constructed
ability to handle spatio-temporal coverages and convert it into something
that is spatial-only.

I've proposed fixes to 19123.  These fixes preserve the spatio-temporal
nature of the domain objects, while establishing a parent-child
relationship in the specializations (e.g. I did it with type-narrowing.)
Obviously we will need to address the type narrowing when these interfaces
are converted to GeoAPI, but at least the types are related now.

For those interested, please review and comment.  Silence means these
departures from 19123 will be implemented in GeoAPI and GeoTools.

Bryce
PS: These are part of the Reference guide on
http://docs.codehaus.org/display/GEOTOOLS/Coverage+Implementation+Central

They are incomplete: only sections 1.1 and 1.2 in each contain the fixes.
http://docs.codehaus.org/download/attachments/46835/Specialized+Domain+Objects.pdf?version=1
http://docs.codehaus.org/download/attachments/46835/Specialized+Value+Objects.pdf?version=1