We currently have the flux error fields in reference catalogs listed as "optional" fields. Making them required (and filled with NaN if not available in already existing refcats) would simplify code in meas_algorithms, and ensure that newly created refcats have more useful minimal defaults. All of our new sources for reference catalogs will provide flux errors, and fluxes by themselves are not meaningful without their errors.

The reference catalog schema is defined in the docstring for LoadReferenceObjectsTask in meas_algorithms, where the flux and fluxErr fields are defined as:

• *referenceFilterName*_flux: brightness in the specified reference catalog filter (Jy)
• *referenceFilterName*_fluxErr (optional): brightness standard deviation (Jy);

In addition, Russell Owen and Jim Bosch agreed in comments on RFC-271 that the camFlux fields could be removed from the schema, so I also propose to remove them as part of this change. The existing docstrings do not make it clear how they are different from the regular *_flux fields. We're not using any such fields in the existing refcats on lsst-dev. These fields are defined in LoadReferenceObjectsTask as:

• camFlux: brightness in default camera filter (Jy); omitted if defaultFilter not specified
• camFluxErr: brightness standard deviation for default camera filter;
  omitted if defaultFilter not specified or standard deviation not available that filter
• *cameraFilterName*_camFlux: brightness in specified camera filter (Jy)
• *cameraFilterName*_camFluxErr (optional): brightness standard deviation
  in specified camera filter (Jy); omitted if no data is available;
  possibly nan if data is available for some objects but not others