I had a discussion with Robert earlier today and he indicated that as far as data handling and reduction is concerned, the numbering of the science fibers is not an issue, however we want to handle it. I therefore propose for simplicity (well, not quite simplicity, which would say we do a zero-based scheme) that we number them from 1 to 2394, so the science fiber sf ranges are

• 1-600 would be in order from the top to the bottom of the slit in spectrograph 1
• 601-1200 would be in order from the top to the bottom of the slit in spectrograph 2
• 1201-1797 would be in order from the top to the bottom of the slit in spectrograph 3
• 1798-2394 would be in order from the top to the bottom of the slit in spectrograph 4

Note that each spectrograph has also a fiber HOLE number, which runs from 1-651 from the top to the bottom of the slit. Most of these are occupied by the 600 or 597 science fibers, but some are occupied by engineering fibers and some are blank; there is a large block of blanks in the center to address the gaps beween the optical CCDs.

Yuki and I are preparing a document and tables with more than you wanted
to know about the fiber mapping, and we would like to include this ID, so
I will proceed with this unless I hear desperate screams. The 597 fibers
in sp 3 and 4 is a pain, but one can just make a table.

Just to confuse things, remember that there is a running COBRA number from 1 to 2394 which bears only a very complicated relationship with the science fiber number, which for data reduction and handling reasons clearly needs to be tied to the spectrographs.