Please add more details, and describe the point and contents of the file rather than referring to Craig.

The file contains the predicted wavelength for each pixel in each FiberTrace. It is used to find the FiberId from the center position of the FiberTrace, and to get a preliminary line list for the wavelength calibration.

I can see generating a seed solution, with the slit and ccd at nominal position and focus.

But I think that any input list of line,lambda will need to be slightly stretched to best fit the measured peaks, and the 1-d trace of detected fiber centers will need to be correlated to the expected pattern for a given slit (also slightly shifted/scaled). [JEG has pointed out that the bounds of the fiber centers for the entire slit will be a good measure of focus.]

As a strawman how about two input files:

• per-detector nominal wavelength map image. spLambdaMap-$arm-$spectrograph?
• per-spectrograph slit map, enumerating the fibers along with their type, status (e.g. broken), known positions, etc.? spSlitMap-$spectrograph?

I'm confused, because Andreas seems to think this is a file that already exists and just needs appropriate handling, whereas Craig seems to be discussing the potential contents of two other (but related) files. Can somebody clarify?

I have been using the existing but now outdated file in question for 2 different things:
1. to assign the fiberID to each FiberTrace based on the center positions of the FiberTrace on the CCD image,
2. to create a preliminary line list for the wavelength calibration from the predicted wavelengths for each row in the extracted FiberTrace.
Craig would (after discussion with Jim sensei) like to split this into 2 different files, one for the wavelengths and one for the fiber positions on the CCD. While this does make perfect sense, I think we should file a ticket to do this later and for now simply recreate the old file (if possible?) I was using as this is blocking me in PIPE2D-80.

Maybe Craig can comment on this further...

Obsoleted.