From: David Z. <dz...@re...> - 2004-10-19 23:29:40
|
On Tue, 2004-10-19 at 16:24, Martin Desruisseaux wrote: > David Zwiers a =C3=A9crit : >=20 > > Heya Martin, > >=20 > > I agree with the existence of a naming issue. I prefer a suffix as we= ll, > > but I am concerned with the possible confusion of ...GEO and ..._GEO > > trying to figur eout off hand whether it is geoapi or geotools. The '= I' > > just isn't very appealing. >=20 > Maybe GT then, as you suggested. >=20 >=20 > > I guess I should make a suggestion ... my preference is to have the > > default as an inner class of the interface, but that isn't always > > possible (and would be a case of calling the kettle black with my sch= ema > > stuff). >=20 > Inner class is really not possible in the case of implementation of=20 > GeoAPI interfaces. >=20 >=20 > > ParameterValueGT (Geotools) > > ParameterValue (GeoAPI)=20 > >=20 > > I would rather have the interfaces maintain the 'original' name, and = the > > implementations add postfix initials to distinguish themselves. >=20 > I agree that interfaces should have the original name. The "GT" suffix=20 > for Geotools implementations seems a good choice. It keep rooms for=20 > other suffix like "JTS" for Geometry wrappers around JTS objects. >=20 > Wonder if we should put a "_" in front of the suffix, e.g.: >=20 > ParameterValue_GT >=20 > The reason is that some GeoAPI interfaces already have a suffix. For ex= ample >=20 > CartesianCS (GeoAPI) > CartesienCS_GT (Geotools) > GeographicCRS (GeoAPI) > GeographicCRS_GT (Geotools) I like it, sounds like a good plan to me. David >=20 > It would also separate visually the suffix (which is about=20 > implementation) from the name (which is about functionality). >=20 >=20 > Martin. >=20 |