Add a check-list when comparing results between software

[hakonanes]

There are some reported issues where users don't get the pattern they expect because of some detail in how the detector-sample geometry is set up or how the orientations are defined (I include myself in "users"!). Here're some of them:

• Issue regarding to EBSDmasterPattern.get_patterns method #746
• Set correct sample tilt for TKD dataset read from Bruker HDF5 #669
• How to plot a phase map #666

It would be good to have a sort of check-list to go through when this happens. Something like double-checking:

• Detector-sample geometry:
  • Sample tilt
  • Detector tilt
  • PC convention
  • PC
• Orientation:
  • Sample reference frame (90 deg @ [001] between kikuchipy/EMsoft/EDAX -> Bruker/Oxford)
  • Angles in degrees or radians

This should be an example in the examples gallery.