Instances of different sky coverage in target and best

Every attempt was made to make the best area cover the same area as the target area, and to make the chunk boundaries the same. For the most part the target and best start and endMu's for the same chunk of sky agree to within 1'', however there are a few exceptions which we list below. For these few exceptions, if a whole stripe is considered, then the areas agree to 1'' -- only the end/begin points of internal chunks within a stripe are slightly offset.

The startMu and endMu coordinates listed below are in arcseconds, divide by 3600.0 to get degrees. These are inclined great circle coordinates and so mu = RA (J2000) only on stripes 10 and 82 (the equator). The numbers can go beyond 360*3600 to avoid endMu < startMu).

• Stripe 9, startMu 614999: best ends at 724412, target ends at 723324
• Stripe 9, startMu 724412 (best) , startMu 723324 (target): both end at 858918

Note that together the two stripe 9 chunks both start at 614999 and end at 858918.
• Stripe 10, startMu 439001: best ends at 533284, target ends at 522543
• Stripe 10, startMu 533284 (best), startMu 522543 (target): best ends at 849998, target ends at 851094
• Stripe 10, startMu 849998 (best), startMu 851094 (target): both end at 899949
  Note that together the three stripe 10 chunks both start at 439001 and end at 899949.
• Stripe 82, startMu 1176576

  Target and Best skyVersions go from startMu = 1176576 to endMu = 1499432. These chunks match in length, however note that in the 'Best' version of the sky, we replace runs 1033/1056/94/125 with (improved quality) runs 3388, 3325 2662 and 3388. We needed to minimize the impact of the 'astrometrically offset by 74 arcsec' early run 1056 and thus avoid 'gaps' in the filled best skyVersion of the stripe 82 between scanline (camCols) at the beginning of the run. This run 1056 exists in the 'resolved' target skyVersion, and we can't remove it from the target imaging data (or we lose that information), however, the plates (plates 371-380) were targeted off of run 1033/1056/94/125 have a slightly non-uniform tiling pattern (i.e. small horizontal gaps) where no objects are present. Large scale structure folks may wish to consider this information in making their region masks.

Users interested in large scale structure map calculations should note that the 'lambda limits' are not used for the three southern stripes, 76, 82 and 86, instead the 'startMu' and 'endMu' limits are followed. For the 'target' skyVersion of data (from which spectroscopic targets were selected), 'lambda limits' override mu limits in the Northern survey. For the 'best' version of sky (for DR2/DR3), 'lambda limits' are not recognized for any stripes, mu limits (as listed in the tsChunk files) are always used when determining if an object is 'PRIMARY' (mode = 1) or 'SECONDARY' (mode = 2).