Thanks for the debrief here, Martin! Cheers, Chris
On Oct 8, 2012, at 9:15 AM, Martin Desruisseaux wrote: > Le 07/10/12 00:20, Adam Estrada a écrit : >> Yes! Please let the folks there know about SIS! I am interested to hear what >> feedback you receive. > > Sure, I will do. I have a "GeoAPI tutorial" cession tomorrow; I will mention > SIS there. > > In the main time, there is a quick report about the today meeting. We had > only one technical session (there will be more tomorrow). This session was > about geometries in Simple Features. Geometries are defined by the ISO 19107 > specification, and represented as Java interfaces in the org.opengis.geometry > package and sub-packages [1]. However this specification is reputed complex, > and the Java interfaces are still in the "pending" part of GeoAPI despite 2 > or 3 implementation attempts. The "Simple Feature" specification was used to > define a simpler geometric model, but it had more limitations (mostly 2D > Cartesian). There is some interesting ideas floating around for "completing > Simple Feature" or "simplifying ISO 19107" (depending on the point of view). > It may be too early for telling much more, but it could impact the way > geometries would be implemented in Apache SIS. It may be a reason for not > rushing too much on geometry and focus on other parts (e.g. coverage) in the > main time. > > One observation about the Coordinate Reference Model defined in the ISO 19111 > specification (and consequently the model expressed by GeoAPI interfaces): > this is a model describing quite extensively the reference system metadata, > but gives few information about how to calculate geometric properties > (distances, angles, etc.). Actually there is a package for the coordinate > systems [2], but each each leaf interfaces (CartesianCS, EllipsoidalCS, etc.) > is basically empty. For the EllipsoidalCS for instance, we can't do anything > without the axis length, which are stored elsewhere (indirectly in the > GeodeticDatum). For gravity-related geoid, there is no information at all. > Maybe some adaptations would be considered necessary. In such cases, we will > need to revisit the Java interfaces and see how we could eventually > incorporate the changes. > > Martin > > > [1] > http://www.geoapi.org/geoapi-pending/apidocs/org/opengis/geometry/coordinate/package-summary.html > [2] > http://www.geoapi.org/3.0/javadoc/org/opengis/referencing/cs/package-summary.html >
