During engineering tests, a series of both spectral and image calibration frames should be acquired at many different telescope positions. This series should be done again after any corrections or re-engineering are complete, to serve as a reference for the future. From then on, we might expect that the observer takes one or two calibration frames either at the start of each run or at the start of each night. All of these calibration frames should be saved as part of a continuous history of the instrument's performance.
If the instrument is stable enough and/or the flexure compensation system is effective enough, it may be possible to produce and archive a standard set of calibration data which can be used by many successive observers. The information management system should accommodate this possibility.
de@ucolick.org