think more about constant shear on the sphere

[beckermr]

From #17, I wrote

I am still not 100% sure exactly what we want to mean by constant shear on the sphere. I can think of two meanings.

A sheared round object will have the same angle with respect to the local ra-dec lines at the object's center.

If you parallel transported a sheared round object to the center of the patch along a geodesic, they would all have the same shape.

These two things are not the same IIUIC. I have currently implemented the first one, since using codes we have now, this will result in a catalog of shapes that should be all the same (i.e. our current codes report shears wrt the local ra-dec basis.)

However, from a physical point of view, the second is absolutely more correct (and would better match assumptions about shearing the scene in metacal). The issue with it is that if we use it, then our codes will not report a constant shear and we will have to remove the effect of the coordinate system before reporting shear measurement biases. This is not hard (you parallel transport your shapes from their measured locations back to the center of the patch), but I am punting on this now.

We should think on this more.