Minutes for UA MOPS meeting, April 6, 2006
In attendance:
Tim Axelrod, Kobus Barnard, Alon Efrat, Scott Morris,
Main points:
1. This was the week that a telecom with CMU/U.Pitt, but it did not work out.
We will shift the every second week telecom by a week, and try again next
week.
2. We discussed summer availability. Much of the UA team will be in town for
much of the summer. Students Scott Morris and Praveen Rao will put at least
as many hours towards LSST as they are doing now
3. Scott presented some success using OrbFit to characterize object prediction
errors. With differential corrections, we are getting much more reasonable
projection distributions. Scott has added a lot of checks to the OrbFit code
making it so that array bound errors force a return to the C wrapper,
instead of a program crash. He is now able to run large scale experiments.
4. Tim suggested that in the long run, the major source of error might be
systemic errors outside of what we have analyzed so far. He also mentioned
that he recently found a paper quite relevant to the issues that he will
circulate. Kobus pointed out that as we trust some of our fits, we will be
able to recalibrate our error estimates, and possibly characterize some of
these other errors. In other words, the system could learn as it goes about
how well it is predicting position, and make adjustments.
5. Tim mentioned that LSST and PANSTARS are close to agreements on the
pragmatics of working together. We discussed how this may impact our
activities. In particular, we may benefit from some MOPS pipeline
infrastructure. We are hopeful that we will be able to collaborate on orbit
fitting issues.
6. Alon initiated discussion on astronomical queries, making the point that
even
if we are not responsible for them in the long run, knowledge regarding
potential use might influence our design. We identified three general kinds
of queries: those which refer to orbital parameters, those which refer to
expected detections (and corresponding object points) in a section of the
sky at a specific time, and those wanting a list of detection events that
are likely to be associated with a given object.
--
Kobus Barnard
[EMAIL PROTECTED]
http://kobus.ca
730 Gould Simpson, U. Arizona (520-621-4237)
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data
