Re: [Geoapi-devel] UML Diagram and Spec to Interface instructions

[Martin D. <mar...@ge...>]

Jody Garnett a écrit :
> Please note that Geometry (and Feature for that matter) usually behave
> as "Data Objects" - two instances of a Point with the same CRS and
> Coordinate should be considered equal(). So use of primitives is spot on.

I'm not sure to understand... Are you making a connection between 'equals'
contract and use of primitive types? To me, the return types ('double' vs
'Double') have no connection at all with the 'equals(Object)' contract...

> #1- if you are a data object (like Geometry, Feature, SpatialExtent).
> The entire data object is to be represented as null if the required
> values are not present.

I though that metadata were data object, and we do have non-null metadata with
required values not present. It was even the whole reason for GEO-103 and all
this "primitive type --> primitive wrapper" changes as I understand it.

> Based on your changes I thought you were confused between the difference
> of "Data Object" and the "Metadata" interfaces. I did not want to single
> metadata out for special abuse - lots of other specifications can be
> treated in the same way (Catalog, Feature and Style, ...).

It sound like that I'm confused. Metadata are not Data Object then? What is the
difference?

> Thanks for the review Martin, and I am glad this discussion is wrapping up.

Thanks for your patience and answers :)

	Martin