	
--

As usual, I have totally forgotten the previous discussion with respect
to rummaging around in the USNO-B database, but that is just normal for me.
I think that there are two ways to proceed, and both stem from the exercise
of attaching WCS headers to the individual PMM camera footprints that we
did for Google.  I should note that Andy Connolly swears that Google
will soon serve the 10 TPixels that we gave them, but I am not going
to hold my breath.

The PMM scanned each Schmidt plate as 588 individual footprints.  The
scale is about 0.9 arcsec/pixel (2X optics, 6.8 micron camera pixels,
67.5 arcsec/mm for Schmidt plates).  The readout format was 1394x1037
(not even an integer multiple of 4!) of which we claimed that the
useful area was from (67:1378 and 3:1035).  There are no dead pixels
in either camera, the bias frames are zero, and we have the flats.
Each image covers 17.8 by 14.0 millimeters but the step-and-stop
was done on a 16 by 12 millimeter grid giving about 2 millimeters
(about 2.2 arcminutes) overlap between footprints.

I wrote software that extracted each footprint from each plate,
made a short, paired list of (X,Y) and (RA,Dec), fed this to imwcs(),
and saved the image+WCS as separate FITS files.  We scanned 12070
Schmidt plates of which we saved the pixels for all but the 897 UJ
(3 minute exposure blue plates from POSS-II).  So far as I know,
We have 588 times 11,173 FITS files spinning on our disks somewhere.
These files have been given to Google and SDSC, and they are available
to anybody else that could seriously use them.  We do not have the
bandwidth to make soft copies, so we must negotiate for reasonable
ways to copy the 10 TPixels involved.  We have yet to play the same
game with the Yale Astrograph plates, but will eventually add these
3 TPixels of FITS images to the archive.

So Option 1 would be to just hand you guys the FITS WCS images and let you
play.  You can figure out whether IPP or SExtractor works on photographic
pixels, how to do photometry, how to handle the overlap zones, and
all of the other unpleasant things I had to do for PMM for which I
have been so roundly criticized.

Option 2 would require a small hack in the pixel processing pipeline
that would save the PMM detection lists and the WCS transformations
in some reasonable way.  This would save you the work and bandwidth
of dealing with the images, but would then rely on the PMM's image
processing software for computation of the detection parameters.

Depending on which, if either, seems to be a reasonable way to
proceed, I think that we have enough bandwidth (or could use DVDs)
so send a small sample of what the data would look like.  Did I send
you the flat ASCII file of the nominal pointings and epochs (12070 lines)?

Let me know.  I will be in Flagstaff for the next couple of weeks,
but then my travel schedule goes nuts.

-Dave

--

I had thought that I passed you this file before.  The files
I would be passing you would start with the 7 character root
in the third column, SSnnnnn where SS is a unique survey
identifier, and nnnnn is a unique plate number from that
observatory.  You can safely ignore the first 2 columns.
The plate is the third, followed by DDMonYYYY HH:MM
HHMMSS sDDMMSS Emulsion Filter Exposure H:MM.  It is
believed that the UT is for the start of the exposure,
but it might be for the middle.  Many things are lost in
the cobwebs of antiquity.  I can give you a table of
nominal wavelengths for various Emulsion+Filter combinations.
The final column is the hour angle of mid-exposure.

-Dave

example:

0         1         2         3         4         5         6         7         8         9         0         
01234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789
so0001 /nofs/leo1/sel/jb0/se100/so0001a.inf PO00570 23AUG1952 09:22 194300 +900000 103AO NONE    10 0:13E
so0002 /nofs/leo1/sel/jb0/se100/so0002a.inf PO00567 22AUG1952 08:21 223100 +900000 103AO NONE    10 1:14E
so0003 /nofs/leo1/sel/jb0/se100/so0003a.inf PO00568 22AUG1952 10:53 000624 +843140 103AO NONE    10 1:10W
so0004 /nofs/leo1/sel/jb0/se100/so0004a.inf PO01277 23NOV1954 06:45 025031 +842224 103AO NONE    12 0:29W
so0005 /nofs/leo1/sel/jb0/se100/so0005a.inf PO01328 29JAN1955 06:42 052420 +840404 103AO NONE    12 2:19W


emulsions:
098
098-0
103AD
103AE
103AO
IIIAF
IIIAJ
IVN

filters:
#12
AMB2
AMB3
AMB4
AMB5
AMB6
AMB7
AMB8
GG358
GG385
GG395
MULTI
NONE
OG590
RED
RED66
RED67
RED68
RED69
RED70
RED71
RED73
RG2444
RG600
RG610
RG630
RG715
RG9
RP2444
WR88A
YEL3
YEL8

combinations:
098 RED 
098-0 RED70 
098-0 RG630 
103AD MULTI 
103AD YEL3 
103AD YEL8 
103AE #12 
103AE AMB2 
103AE AMB3 
103AE AMB4 
103AE AMB5 
103AE AMB6 
103AE AMB7 
103AE AMB8 
103AE NONE 
103AE RED66 
103AE RED67 
103AE RED68 
103AE RED69 
103AE RED70 
103AE RED71 
103AE RED73 
103AE RG2444 
103AE RP2444 
103AO NONE 
IIIAF OG590 
IIIAF RG600 
IIIAF RG610 
IIIAF RG630 
IIIAJ GG358 
IIIAJ GG385 
IIIAJ GG395 
IVN RG715 
IVN RG9 
IVN WR88A 
