Calibration status: conclusions from
Calibration Meeting (Jeff Mangum)
Jeff gave us a summary of the calibration meeting the day before. A
Calibration Group has been formed. Among the news are:
If a coherent photonics apex radiator is implemented, its possible
uses are for amplitude bandpass, amplitude gain ratio,
amplitude polarization, but not the phase. Such a radiator
is considered desirable, but is not in the project scope.
Amplitude calibration: the choice apex dual load system
versus semi-transparent vane is still pending. A dual load
apex system will require switching at 20Hz.
In September 2001, ASAC suggested a core program for pipeline &
offline data processing. The goal was to define a narrow path
and expand from there, while getting a feedback from users on
the interfaces. In March 2002 the same concern was raised
again. Brian proposes that in the software release plan one
could announce particular events where testing will be possible
for the end user. One could consider making demos at ASAC
meetings.
For Proposal Review, ASAC suggested the name Proposal Review
Committee (rather than Time Allocation Committee). They stated
that the time parity among partners need not be balanced on
short time scales (month) but on longer time scales. As to
whether `time is time' or not (some weighting by weather
conditions) there is no definite answer, or whether the
structure of PRC's should be partner-based or subject-based.
We deferred issues about the relation of the archive with
Virtual Observatories to the Saturday session.
Next ASAC face-to-face meeting is expected in the week
containing Sep 11th, in Soccorro.
AIPS++/Plateau de
Bure test status.
I gave a short account of the AIPS++/IRAM test
status. End of Phase I is expected next month when A. Kemball
and K. Golap visit Grenoble; start of Phase II is expected
then. SSR is in favor of a Phase III as soon as possible (see
discussion of AIPS++ benchmarking).
Friday 26, Morning
Answers to comments
on Pipeline Requirements & discussion
F. Gueth gave a presentation of the
main issues raised by the reviewers. From the discussion that
followed he will (together with P. Schilke) prepare answers
to those comments and the relevant changes to the Pipeline
chapter.
I noted some important points:
Steve Scott asked whether the sub-integration variations
in system temperatures should be taken into account for data
averaging on these short time scales.
For quick look purposes it is reasonable not to do time
interpolation of passbands.
There is an unsettled issue on how frequent will systematic
pipeline reprocessing be needed after re-measurements of
antenna positions.
We will distinguish between full break points, where science
data reduction is performed, and the observer is required to
acknowledge before any further scheduling of the project is
possible, and simple monitor points, where the user simply
wants to have the science data reduction be done, without
halting any further scheduling for the project.
The above will imply that new option is available in the
Observing Tool.
We drop the idea of reversing steps in pipeline data reduction.
The last quick look image in a session will be available (on
the web) for 1 week after the end of session. Quick look
images are not archived.
What actually happens when the pipeline fails is left as an
operational issue.
In the afternoon Francois gave an presentation/demo of his
work on the simulator and the front-end observing tool he wrote
(e.g. for the correlator). Simulation of ALMA size data sets with
millimeter-wave atmospheric effects is achieved, producing AIPS++
measurement sets as output.
Simulator Requirements (Robert Lucas)
We discussed priorities for the various items in the draft list of
simulator requirements. The minimum priority is 1. I will write a
new version of this text including priorities; this will be
included in memo 11.
AIPS++ audit process and benchmarking plan (Steve Myers)
ALMA-SW-0018 was completed and reviewed. There remain some TBD
items, of which the benchmarks are the most important.
Steve will send us a summary of his propositions for the audit and
benchmarks.
The auditing will be performed on the SSR side by Steve
Myers, Francois Viallefond and Koh-Ichiro Morita, on the next
available AIPS++ version (June). The marking will follow the
proposition of Steve, though it is clear that the team will not be
able to test everything. It is important to see which features are
there and whether they are known to work and documented. A similar
marking will be done in the AIPS++ side but the SSR/ALMA
opinion should prevail.
The goal is to have first results for September (ASAC
Meeting). One should redo this every year or so the get the
information on the slope of ALMA-related AIPS++ development
process.
Define benchmarks will take longer. The same team will define
benchmarks that will enable to quantify AIPS++ performance on
a few fiducial data sets. They should simulate ALMA data sets
of different sizes, in standard modes, but should include some
extreme data sets. It is desirable to stay in contact with
other teams doing similar tests for other project (EVLA, SMA,
BIMA ...)
As it is extremely important to identify problems we may
encounter in the future using AIPS++, and as the performance
issue is regarded as a show-stopper, the SSR strongly supports
performing Phase III of the AIPS++/Bure test as soon as
possible. This should take as input simulated ALMA
interferometric data produced by F. Viallefond's simulator, in
a form that can be reduced using the tools developed
by the AIPS++ team for the Plateau de Bure data.
Saturday 27, Afternoon
Answers to Comments on Archiving Requirements & discussion
(K. Tatematsu). See Ken's slides.
Some of the main points (from my notes):
It is agreed not to separate technical and science archive
requirements.
We should archive all monitor data
In 7.1-R1 add 'regardless of where the information
physically resides'
Page 40, 7.2-R3: remove last sentence, as it contradicts
another requirement.
We should make sure that we have required from the pipeline
to put quality information in the archive.
Summary and plans (Robert Lucas)
Steve Scott mentioned some small consistency problems he
had identified in memo 11. R.Lucas will make the necessary
changes. We will not change the data rate in memo 11 but will
submit a change request, following the data rate note of
2001-12-06 by
S.Scott, S.Myers and M.Momose.
SSR members should express their preferences for the
subsystem thei would like to follow in Phase 2 as Subsystem
Scientists.
We expect to meet again face-to-face at the time of the
Software PDR as most of us will attend as Subsystem Scientists.
We may need a phone meeting at the end of May. I propose May
29th (15:00 UT).
Action Items
S. Myers: prepare a short description of AIPS++ auditing plan according to his propositions and following discussion (May 15)
S. Myers, F. Viallefond, K.-I. Morita: start the auditing process
(May 15)
F. Viallefond Prepare a simple ALMA-size data set for Phase III of
ALMA/AIPS++ test (June 1st)
F. Gueth, P. Schilke: edit replies to reviewers and new version
of pipeline chapter (June 1st)
K. Tatematsu, K. Nakanishi: edit replies to reviewers and new
version archiving chapter (June 1st)
R. Lucas: insert in memo 11 chapter on simulator with
priorities (May 15)
R. Lucas: correct in memo 11 points noted by Steve Scott (May
15)
All: give their own preferences for subsystem scientist choice (May 15)
$Id: 2002-04-25.html,v 1.11 2002/05/02 14:00:40 lucas Exp lucas $
