I found and fixed a bug in the KPNO Mosaic astrometry code (mosaic_wcs in the astrometry.net codebase). I re-ran the Willman data and I started the hand-inspection of the astrometry in cases marked by the code as possible problems. All inspected so far are fine, although robust object detection of relevant objects (not stars, since the USNO-B1.0 is dominated by galaxies!) remains an issue.
Chris Stubbs gave a very nice talk at MIT on the use of astrophysics in testing fundamental physics; in particular he discussed the issue of absolute radiometry in the supernovae projects, an issue I sometimes get upset about.