Date: Wed, 8 Jul 1998 14:56:07 -0700 (PDT)
Subject: IPAC 2MASS WG Mtg #156 Minutes
To: 2mass
Cc: chas, stiening, bgreen
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-MD5: i8WNumKPsCjgQSYkMVj4sA==
IPAC 2MASS Working Group Meeting #156 Minutes
7/7/98
Attendees: R. Cutri, S. Van Dyk, T. Evans, L. Fullmer, D. Kirkpatrick,
G. Kopan, H. McCallon, R. Beck, B. Nelson, J. Mazzarella,
D. Engler, S. Wheelock
AGENDA
1.) RTBs running with 2MAPPS 2.1
2.) QA
3.) IRSA, CatScan, and The 2MASS Sampler
4.) New 2MASS Webmaster
DISCUSSION
1.) RTBs running with 2MAPPS 2.1
R. Beck is currently running the RTB nights 980114n and 980328s using
2MAPPS v. 2.1, which includes a new version of MAPCOR. These runs should be
done in a couple of days.
H. McCallon pointed out that he has found something disturbing about the
positions on southern nights, e.g., 980403s. There appears to be a bias in the
RA residuals from scan to scan in the overlap region. A comparison with the ACT
does not reveal this problem. R. Cutri provided a possible diagnosis: there may
be camera distortion in the south, and suggested that the distortion map be
stepped up to high priority. R. Cutri has also found a problem with the
photometric agreement in the overlap regions.
S. Wheelock and R. Cutri discussed the calibration zeropoint. Currently,
the zero point for each calibration scan is determined based on a standard star
(i.e., catalog mag - instrumental mag). When there is just a single standard,
this number can introduce some noise into the photometric correction for each
scan. To reduce the effect, R. Cutri suggests using the average of the set of
six zeropointsas the correction for each of the calibration scans. S. Wheelock
pointed out that now there are more than one standard per field, i.e., secondary
standards, in many of the calibration fields. Thus, the zeropoint used becomes
the average of all the magnitude differences for all the standards on a
calibration scan. This averaging of the "delta mags" should reduce the "noise".
R. Cutri concurred that our effort should be to get secondary standards in all
calibration fields.
In addition, R. Cutri suggested that we may need to output another quality
table from CALMON that gives the important dispersions and has a way to flag
those that may suggest a problem. This would allow CALMON to be run with
thresholds set high, allowing more data to pass through, but would have some
way of catching any potential problems.
The PSFs for the south need to be delivered by W. Wheaton. Artifact flags
need to be tuned as well by T. Evans. T. Evans will be away, starting on July
25. She hopes to have all the necessary tuning done by then, for the August 1
implementation of 2MAPPS 2.1. R. Cutri wants to be able to throw the switch on
August 3.
2.) QA
The Nightly Science QA page is currently working as designed, thanks to the
hard work of L. Fullmer. The Scan-by-Scan Science QA page, however, is in need
of revision. (The page exists; it's become a bit of a monster, though, and is
currently cumbersome to use.) After review of some 37 nights of full production
processing, we have learned that several new quality indicators are needed.
These include (1) an indicator that the RA has slipped during a scan and (2)
improved indicators for photometric and astrometric overlap analysis.
R. Cutri stressed that it is essential to move on getting all QA pages up to
speed. After the meeting, R. Hurt has expressed a willingness to help with
these tasks. He joins the 2MASS team on July 27th.
3.) IRSA, CatScan, and The 2MASS Sampler
a) R. Cutri met with J. Mazzarella and J. Good last week to discuss the
requirements of the IRSA system to suppport the 2MASS sampler data release.
J. Mazz will send the schedule and task list to meet these requirements to
R. Cutri on July 8.
b) D. Kirkpatrick had a problem where the file output by a long CatScan search
was removed from the server before he had time to download it. J. Mazz reminded
all that, by default, data files resulting from CatScan and image archive
services reside in temporary workspaces on the Web server. These files
currently get purged when they get *2 days old*, so J. Mazz stressed that users
should copy files to user disk ASAP. D. Kirkpatrick's problem brings up the
need to possibly increase the lifetime of temporary files in IRSA. However,
IRSA users have the option of obtaining a long-term/permanent workspace in which
data files will be purged after a longer time period. Please select
"Personal Workspace" (bottom of the left index frame at http://crystal:8001/)
for information on this feature.
Some details such as disk space quotas and lifetimes of data files in the
temporary and personal workspaces need revision.
c) The hw/sw configuration for the new operational Web 2MASS/IRSA server has
been discussed with S. Lo. The new 'babel' will require a connection to the EMC2
to enable access to the compressed 2MASS images. S. Lo will announce ASAP
when the machine will be available to us.
d) R. Beck reported that after the latest DBMAN load, indexing of pt_src took
*6 hours*. R. Cutri assigned the following questions/tasks for the DB group to
address:
i) If the money can be found, document what is the optimal hardware config.
and upgrade path to improve the indexing time to keep up with 2MASS
and eventually support ~600 million rows.
ii) What is our estimate of how indexing time will improve with additional
CPU and I/O channels on the EMC2? How does it impact the expected
n*log(n) scaling of a single CPU, etc.
(Some of the speed problem could be solved by simplifying the strings searched
for in queries. For instance, K purges and persistence probabilities are
currently character strings, which slow down indexing. T. Evans said that
maybe changing these to integers could speed things up.)
iii) Please schedule an Informix consultant to visit to help optimize the
current system and to plan upgrades. Have a consultant visit before
T. Evan's vacation or ASAP after (25 July - 1 Aug.).
e) Addition: BUG updates
i) A bug in the Survey Visualization/image server tool was encountered
while J. Mazz was coaching P. Seitzer (U. Michigan) over the phone to
get him started with 2MASS data access. Many of us have encountered this
problem, and J. Good has been asked to fix it ASAP.
ii) The "within" bug in isisql (used by CatScan), which produces errors with
small search radii, has been isolated. A fix is in progress, and it will
be tested and configured with CatScan by the end of this week.
R. Cutri asked T. Evans about scans that don't make it into the database.
These scans cannot be allowed to "pile up" without attention. DBMAN indicates
whether or not a scan has gone into the database. This information must make
its way to D. Kirkpatrick. R. Cutri asked T. Evans about the statistics of
scans that do not make it to the database. T. Evans reported that this is
about 1 scan per night.
R. Cutri emphasized that we should not wait to react to a problem and to
alert the rest of the group about the problem.
4) New 2MASS webmaster
S. Van Dyk is now the new 2MASS webmaster, freeing up J. Mazz of this
responsibility. All documents relevant to the 2MASS webpage should be referred
to S. Van Dyk.