The chaos in here is because there were several experiments on the various
steps that were necessary to re-reduce USNO-B.

* bycd2table.py -- hack to produce input for plates.rd
* dump2masscross.py -- dumps 2MASS positions with USNO ipixes
* dumpppmxcross.py -- dumps PPMX positions with USNO ipixes
* mergeDumps.sh -- puts together the various dumps (PPMX, 2MASS, USNOB) to
  a common list of observed positions.
* obspos.pyx -- a Cython thing that generates the dump from USNO.
* setup.py -- pythons obspos


Validation Scripts

* checkPlateAssignment -- a validation script for plates.rd data
* findMeanErrors.py -- some validation script


Obsolete stuff

* dumpobspos.py -- this would dump the obspos table that is no longer built.
* makecross.py -- this is now basically in the RD
* makeobsposusnob.py -- superceded by obspos.pyx
* merge.py -- python version of mergeDumps.sh