Clarification of trackerHitPlane members

[BrieucF]

edm4hep::TrackerHitPlane is heavily inspired by the lcio equivalent but the member description seems to somehow got lost.

For instance, the u and v members were described in lcio with "Direction of first/second measurement - given as (theta, phi). Defines spanning vector of measurement plane.".
Now we have "measurement direction vector, u lies in the x-y plane" and "measurement direction vector, v is along z", which is much less clear to me. Was there any intention to change the content w.r.t. the lcio version? If not I would propose to use the description from lcio. And we should probably add units ([rad] for u and v and [mm] for du and dv).

I am also guessing that the position member should be in global coordinates, that could maybe be added as well.

Finally, can someone comment on the intended usage (or how it is used in ILCSoft)? I understood from a discussion at an edm4hep meeting that it was enabling the creation of 'measurements' for tracking without the need to access the geometry. But for instance, some tracking packages require planarMeasurement in local coordinate of the sensor plane, this can currently not be constructed without accessing the geometry as far as I can tell.

Thanks a lot for your input!

[tmadlener]

See also the discussion in #160

[andread3vita]

Hi all,
Since u and v are given in cartesian coordinates from geometry, do you think it is more useful to have them in polar coordinates or directly in Cartesian coordinates? I am not aware of all the tools that use u and v from trackerHitPlane, but for track fitting with genfit, it would be very useful to have them as 3d vectors along with the origin of the plane to which the hit belongs.

Ideally, with origin, u and v it is possible to calculate the local coordinates from the global without retrieving the geometry.

Thanks a lot for your input!