Changes between Initial Version and Version 3 of Ticket #492


Ignore:
Timestamp:
Mar 7, 2016 8:23:53 PM (8 years ago)
Author:
butler
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #492

    • Property Status changed from new to assigned
    • Property Work Package changed from SasView QA and testing to SasModels Redesign
    • Property Summary changed from check orientation angles to choice of parameterization of orientation angles
    • Property Priority changed from blocker to critical
    • Property Owner set to butler
    • Property Type changed from defect to enhancement
  • Ticket #492 – Description

    initial v3  
    1 Paul Kienzle has identified what he believes is a serious bug. The documentation suggests that Phi is the angle the particle's primary axis makes with the beam direction and theta the angle the particle's primary axis makes in the plane perpendicular to the beam.   
     1The current orientation angles, while they seem to be correct, are not natural for the user. 
     2For the cylinder, it would be preferable if theta were the angle of the cylinder axis with respect to the beam, with 0 corresponding to a cylinder along the beam, and a circular scattering pattern, and 90 corresponding to a cylinder perpendicular to the beam. This corresponds to latitude. The angle phi ranging from -180 to 180 would rotate the cylinder in a circle of constant angle theta, corresponding to longitude. 
     3This would allow users to set phi by inspection since it would be mostly independent of theta. It would also allow for faster phi polydispersity since a fixed 2D image could be used without having to recompute it for each phi, much like we will (eventually) be doing for 2D resolution 
    24 
    3 However he believes these have been reversed in the code and is causing an unphysical symmetry in the scattering pattern visible under certain conditions.  Others have oft complained of being confused when playing with these parameters which could easily be due to this as well. 
    4  
    5 We must verify this and fix it properly before releasing 4.0 which we expect, among other things to make 2D analysis more accessible.  
     5This would require redoing the model maths and needs broader agreement than just "fixing"