Coordinate system convention in Positions class about holopy HOT 6 CLOSED

Yes, I wrote that to use conventional spherical coordinates. At the time I didn't realize that there was a different convention in scattering. There are, I think, three options for how to deal with this

1. Leave it like it is. This avoids a break with backwards compatability, keeps things sensible except for scattering
2. Change to use scattering conventions. This would be a break with backwards compatabiliy, albiet on something that very few people use explicitly at the moment as far as I know.
3. Add another argument to the Positions (and similar classes) constructor, allowing you to specify whether you want this object to be in conventional spherical coordinates or in scattering coordinates.

I think I lean towards option three, but would welcome comments from @vnmanoharan @jeromefung @RebeccaWPerry @anna-wang

from holopy.

I think my earlier post was incorrect. I think the convention used in the code is consistent with scattering coordinates (theta = 0 corresponds to forward scattering). However, this means that the documentation is incorrect or incomplete, for example here:
http://manoharan.seas.harvard.edu/holopy/users/concepts.html

Perhaps the documentation needs to be updated? Also, the documentation doesn't seem to include any information on the conventions used for specifying polarization.

from holopy.

Ah, so you are running into the perennial coordinate problem in holopy, we
choose to define a z axis such that it measures a positive particle
distance from the camera or focal plane of a lens. This unfortunately means
that in a standard orientation, light is then propagating towards negative
z.

If you can see a way to write that more clearly, we would welcome a pull
request. Feel free to ask me if you need help understanding things enough
to write it.

On Mon, Aug 11, 2014 at 1:39 PM, nschade [email protected] wrote:

I think my earlier post was incorrect. I think the convention used in the
code is consistent with scattering coordinates (theta = 0 corresponds to
forward scattering). However, this means that the documentation is
incorrect, for example here:
http://manoharan.seas.harvard.edu/holopy/users/concepts.html

Perhaps the documentation needs to be updated? Also, the documentation
doesn't seem to include any information on the conventions used for
specifying polarization.

—
Reply to this email directly or view it on GitHub
#10 (comment).

from holopy.

Spoke with @vnmanoharan about this offline. He pointed out that it might be worth adopting the convention from fourier optics of using ξ and η to denote imaging plane coordinates and x, y, z for particle centered coordinates.

We can't ever completely switch to particle centered coordinates because there are many cases where we don't know the particle position when we are setting up the problem. However, x, y, z, η, ξ, could allow us to keep a right handed coordinate system while switching to having positive z along the direction of light propagation which makes more sense for scattering calculations.

from holopy.

Talking with Victoria, it looks like the way to make this clear is to put in a 3d picture in the docs showing the detector, the particle, and the light coming in.

from holopy.

We should also make link to the coordinates page from somewhere in the scattering tutorial because people thinking about scattering are the most likely to need to know about the differing z conventions.

from holopy.