Date: Thu, 9 Oct 1997 11:19:05 -0700 (PDT)
From: jwf@ipac.caltech.edu
To: 2mass@ipac.caltech.edu
Cc: chas@ipac.caltech.edu, sstrom@donald.phast.umass.edu,
    stiening@ipac.caltech.edu
Subject: 2MASS WG Mtg #132 Minutes

           IPAC 2MASS Working Group Meeting #132 Minutes
                            10/07/97

Attendees: R. Beck, T. Chester, R. Cutri, T. Evans, J. Fowler,
           L. Fullmer, D. Kirkpatrick, G. Kopan, J. Mazzarella,
           H. McCallon, B. Wheaton, J. White


AGENDA

1.) 14-Bit Tests
2.) Telescope and Camera Status
3.) Observatory Duty
4.) Automated Observatory-Log-Dependent Processing
5.) Optical-Catalog Associations



DISCUSSION

1.) 14-Bit Tests

    A bug was found in the code that simulated 14-bit raw frame
data. A loop limit had a typo that cause the column loop to run
from 11 to 256 rather than 1 to 256. The effects were very subtle
but were detected by G. Kopan in a difference image made from
corresponding coadded images. The bug has been fixed, and the
test will be rerun. One other change is that the two least
significant bits will be forced to 10 instead of 00.


2.) Telescope and Camera Status

    R. Cutri reported that the telescope control errors reported
in last week's minutes have been fixed, and a new pointing model
has been installed. Current performance involves RMS pointing
control errors of about 6 arcsec.
    Consideration of sources for new camera electronics now
includes JPL as a possible manufacturer. The electronics for the
southern camera have been having problems and need to be repaired
or replaced.


3.) Observatory Duty

    A problem in staffing the northern observatory in the period
from 28 October to around 4 November has been solved. B. Nelson,
H. McCallon, and G. Kopan will donate their efforts to fill this
gap. Observers are still needed for the Dec. 25-31 run.


4.) Automated Observatory-Log-Dependent Processing

    The "Observatory Log Review" for 971004n includes a
special-processing request not to use evening flat sequences
unless no flat sequences were taken in the morning. This request
was honored by R. Beck by editing the "tls" file so that the
evening flat sequences would not be detected by the DARKS
program. But this was only the run for quality checking, i.e,
only calibration scans were processed for the purpose of
reporting possible problems to the observatory. G. Kopan pointed
out that there was no automated way to guarantee that the same
avoidance maneuver would be taken when the entire night's data
come up for processing later. Some discussion of techniques to
automate such things took place, but no satisfactory resolution
was achieved, and R. Cutri accepted an action item to pursue this
issue further.


5.) Optical-Catalog Associations

    Optical catalog associations are being made in the POSMAN
module POSPTS, as discussed in last week's minutes, but
discussion is still going on concerning what modifications need
to be made to satisfy all the users of this information. In
addition to reporting the radial separation between the 2MASS
point source and the catalog object, along with the corresponding
direction, some desire has been expressed for a likelihood ratio
parameter that indicates the plausibility that the associated
objects are in fact the same object. The proposed addition of the
chi-square match parameter supplies some of this, but it was
pointed out that this gives no information about local source
density, hence no hint of the possibility of a random coincidence
due to high source density. The likelihood ratio number was
proposed as a way to show the reduced confidence in a match in a
high density region compared to a match with the same chi-square
parameter in a low density region. It was not clear whether its
failure to distinguish between a tight match in a dense region
and a so-so match in a sparse region ought to be of concern.
Perhaps several new parameters need to be added, e.g., the
chi-square parameter and some indication of local density. The
galactic latitude is immediately obtainable from the celestial
coordinates, of course.
    In any case, it is desired not to keep changing the software
interfaces, since at least four subsystems are impacted each time
this is done, so it is important to arrive at a viable decision
soon. To this end, a splinter group was identified consisting of
T. Chester, R. Cutri, T. Evans, J. Fowler, C. Lonsdale, and B.
Wheaton to look further into the matter. A final decision within
two weeks is the goal.