For historical reasons, the telescope coordinates have only been recorded to integer precision in both RA and DEC in our observing headers. The problem traces back to the olden days when FORTH ruled the mountain. Despite the upgrades to the telescope control systems at the 4-m and the 2.1-m during the past decade, information passed to the instrument computers (CCD, IR) have remained in integer precision. While 1" precision might have been acceptable to most in declination, 1s of time may be as much as 15", making it difficult to run astrometry code on the frames or mosaic frames together.
Recent upgrades at the 4-m and the 2.1-m now result in the RA being recorded to 0.01 time seconds, and the declination to 0.1". The accuracy of these values depends, of course, on how recently and where the telescope coordinates were reset ("zeroed"), but the relative precision within a region should be comparable to the recorded value. Note that retaining the equinox to 0.01 years is an appropriate match to this precision; i.e., knowledge of the equinox to 0.01 years is roughly equivalent to a 0.2" precision.
These upgrades should soon be available within the Wildfire environment used with IR instrumentation. However, the upgrading of the 0.9-m may have to wait until its replacement with the 2.4-m control system!
We are grateful to Buell Jannuzi for urging someone to take on this project, and to Bob Marshall and Rob Seaman for carrying out the actual upgrades.
Phil Massey, Dick Joyce