Observing descriptions & FITS header

One issue in defining the observations and providing proper input to the instrument set-up is the need to provide a complete and well define list of items of the optical elements (filters, slits, grism, etc). From the data base point of view each item should be unique and easily distinguishable (e.g., all the filters are different than all the grisms, each of which have specific numbers), while from the instrument set-up point of view the data-base should provide a way to include the appropriate parameters. Specifically, for each place (wheel) where items can be mounted it should be defined what can be mounted there (e.g., the cross-dispersion grisms can be mounted in the filter wheel), and in what way (horizontal, vertical, blue/red shifted). This defines what one should be able to specify for the set-up of an instrument from the data-base and, vice versa, what the data-base should be able to contain (as defined by the instrument set-up) as for where and how an item is installed. Complete lists have been made for all available optical elements for each instrument. Furthermore, it was defined what are the parameters that need to be specified for each type of element. A conceptual design is being made for the optical element entries in the data base.

Given the need to be able to automatically process images it should be possible to uniquely define the type of observation (imaging, spectroscopy, calibration, test, etc) from the FITS keywords. Our current set of observing description keywords do not follow any specific system and in general the post-processing and analysis tools we use rely mostly on specific instrument keywords (e.g, if there is a slit in the light path) to determine the type of observation. It was considered that the system of ESO using 3 different keywords to define observations giving the category (science, calib, test, etc), the type (object, std, sky, flat, lamp, etc) and (observing) technique (image, spectrum, echelle, polarimetry, etc) where each of them can have more than one entry is well defined and should cover most observations and provide a full description. By following ESO all our data will automatically be ``Virtual Observatory'' compatible as well and adapting to any reduction packages which can process ESO data should be relatively easy.

Like the above, their are many issues that concern FITS keywords in relation to observations and the resulting data. This is particularly relevant for things like automatic scripts as part of calibration plans, or data reduction and analysis through pipelines, but also for extended Virtual Observatory compatibility of our data and we have planned a specific working group meeting consisting of the astronomy staff and software group to define this in detail.

It was discovered that the FITS keywords for the pixel scale for some of the instruments did not correspond to the latest values. The differences are relatively small, but significant. We are in the process of updating the values to the correct ones.

Thomas Augusteijn 2010-02-09