Telecon with Bill Reach, John Stauffer, G. Fazio about IOC GLIMPSE AOR.

GLIMPSErs: Ed, Barb, Remy, Brian, Christer, Marilyn

highlights:

1.  They are going to recalibrate the data and send us a new set.  It will
have better flats, hopefully proper pmasks (and dmasks we hope).
Proper distortion will have in time for OSV. The pre-launch
pmask (what we're using) is outdated and should not be believed.

2.  they were using pre-launch flats. They think that explains our issues
with the band 3&4 gradients which show up in the mosaics.  Band 3 has
vignetting on the left side of 10-15%. That is probably what we saw.  It
should be corrected with flats.

3.  Pointing:  They have 4 levels of pointing calibration.  when one fails
it goes to the next.  There's a keyword, USEDBPH (use basic pointing
correction).  Not sure yet what we are supposed to do but it sounds like
they are going to do pointing correction and perhaps we don't need to.
they said gyro calibrations are worse in pitch than yaw.

4.  Distortion.  Don't use the distortion keywords in that dataset.  they
are garbage.  

5.  Dmasks:  They were surprised that their dmasks don't seem to be
finding saturated stars.  Regarding all those flags in channel two (I
forget what they were), they might be real because Channels 1 & 2 are low
background so have this whatever it is.

6.  cosmic ray detection.  we said we couldn't find any flags.  They said,
we don't like our single-frame detector, and may not include it.  they
said, we should use the SSC mosaicer (which creates an r-mask) to detect
cosmic rays.  unfortunately this doesn't fit our single-frame pipeline at
all.  We have to investigate this and see if there is some way to get what
we want from their SSC mosaicer without rewriting our pipeline.

7.  Column pulldown.  The GOODS team has made a correction. it
seems like we can make a correction.  I'm thinking we'd subtract the
median from the affected column, add back in the median of adjacent ones.

8.  banding.  they are interested to know our findings on banding
correction.

9.  photometric calibration.  we told them our comparison results.  they
seemed to think that was okay (?).

10.  missing data in frames.  should only happen 0.1% of the time and
usually recoverable with a new transfer. 

11.  What they wanted from us.  Did we find many latents?  need to
investigate. Stray light?  This is high priority for them.  they are
making 3 boxes in channel 1; 2 in channel 2, none in 3&4 at this point.
they will be doing a stray light characterization observation in about
2 weeks.

12.  If we could tell them which of our AORs contain 0 mag stars,
 they could adjust the observation schedule to minimize latency.

13.  Issues for them.  channel 1 latents.  solution is annealing. channel
4 residuals seen a week later.  issues for 1 or 2 pointings. (?).
(what's the difference between residuals and latents?)

14.  Overall, IRAC is stable, they won't have to calibrate as much as they
were thinking.