Clipper respecting Coordinate Systems

Should be just redirected to Rejected if Clipper and Clippe have different coordinate systems perhaps?

+1 for me. It's very unusual that FME picks the right way out-of-the-box to reproject anything I'm working with.

Hi @sigtill, @david_r, Isn't it true that coordinate systems are meant to represent object locations on earth or even space? I would expect that a point defined in UTM83-10, which has its validy around North America will not overlap a polygon in EPSG:28992, which has its definition area in the Netherlands:

'

'

However, it is possible that a point defined in a world wide valid coordinate system like ll-wgs84 is overlapping the polygon in EPSG:28992.

'

My wish would be that Clipper (and other transformers) would respect this spatial coincidence.

In principle I agree completely, it's just that (in my part of the world) not all coordinate systems map 1:1 onto one another, there may be local deformations that must be considered.

Concrete example from the CsmapReprojector, where specifying the transformation rule is absolutely necessary for some combination of coordinate systems:

'

Then you have specific transformers for regional or national transformations, e.g. the ReframeReprojector, GtransReprojector or the GridInQuestReprojector. But are they always the right choice? And then what about the EsriReprojector, why might you want to use that rather than the FME libraries (maybe it's more important that your results line up with other data reprojected using the ESRI libraries, maybe not), etc.

Another point is performance, which is best: reprojecting once at the beginning of the workspace, or reprojecting inside every transformer along the way? Will the novice user of FME take this into account, or simply complain that FME is "too slow"?

My point is that reprojection is too complicated to leave to FME unattended :-)

Re-reading your question I'm starting to wonder if I misinterpreted your reply, do let me know if that's the case!

Hi @david_r I think I begin to see the light... thanks. It's just that the DataInspector is dealing it this way yet Clipper does not.