NRAO Applications Users Group Meeting - DRAFT MINUTES
Date: 2007-02-14 (Wednesday)
Time: 1300 MST (1500 EST)
*******************************************************************
Minutes:
Send additions or corrections to smyers
1. NAUG News
o general news (Steve)
- NAUG testing in preparation for the ALMA test continues.
2. CASA Status Report
o It was pointed out that there was no way to get to the CASA
Wiki from the CASA Home pages, nor to the CASA home page from
official NRAO web pages.
ACTION: this should be remedied
Standard items:
o CASA status (Joe)
- see main event
o release status (Joe)
- stable status
Joe said that there would be a stable by 4pm and then at CV the
next day.
UPDATE: There was a stable on Feb 15.
The next stable will be Feb 26 (stuff in on Friday Feb 23) and
the plan is that this is all the new stuff for the ALMA Test,
with only bug fixes after that.
- CASA distribution and platform support
There is a 64-bit rack-mount machine on its way to the AOC.
The 64-bit casa port is scheduled for March 2007.
- defect status
We are still doing bugs via email to Joe.
Will move to JIRA after the ALMA Test.
o use and testing status (Steve, Mark, Joe)
- user testing
- CASA user interface coordination (Mark Claussen)
See UIWG draft document
http://www.aoc.nrao.edu/~smyers/aips++/uiwg/uiwg_report.pdf
- task status (Mark, Steve, Joe)
o development targets (Joe)
- See project office page:
http://projectoffice.aips2.nrao.edu
and CASA Wiki
https://wikio.nrao.edu/bin/view/Software/CASA
3. ALMA
o ALMA Heuristics has carried out a test, which Ed participated
in.
Pipeline Test 4:
http://almasw.hq.eso.org/almasw/bin/view/PIPELINE/UserTest4
I will post Ed's findings soon...
Standard items:
o ALMA Test status (Joe, Steve)
- upcoming test ALMA2006.10-6 (+single dish analysis). See:
https://wikio.nrao.edu/bin/view/Software/AlmaTest5
http://projectoffice.aips2.nrao.edu/ALMA2007.03/ALMA2007.03.html
This is a face-to-face test at the AOC with external testers.
It is a critical test politically, though the package is still
at the "alpha" stage. A test in Q4 after the "beta" release is
a better test of the system.
Debra has made a list of items that need NAUG and scientist
input to prepare for the test. I have posted this at:
http://www.aoc.nrao.edu/~smyers/naug/notes/2007-02-14-dshepherd-almatest.txt
o NAASC issues (John, Crystal)
- CV now has access to the daily series of builds. For "daily",
choose the most recent. It was requested that the most recent
"stable" build be clearly indicated (e.g. via a symlink or
other mechanism).
o ALMA development targets (Joe, Steve)
- mosaicing and simulator developments (Steve,Kumar)
During the week of Feb 5-9, a conclave on the CASA ALMA
Simulator was held in Socorro. Crystal, Rob Reid, and
Steve looked over the current task and tools and drafted
a proposal for a simulator
task architecture
and
parameter set.
The goal is to have something workable that Joe and Kumar
can bring to CV when they visit in late March.
One goal is to have something that astronomers attending the
Disk meeting (http://www.cv.nrao.edu/naasc/disk07/) can use.
Also see discussion below under EVLA.
- see schedule at
http://projectoffice.aips2.nrao.edu
and on the CASA Wiki
https://wikio.nrao.edu/bin/view/Software/CASA
4. EVLA
o Some concern was expressed by Bryan and Michael about the
development targets leading up to the Beta Release. In
particular, it was thought that any effort put in to the
simulator was inappropriate. It was also suggested that
we look into use of the SKA simulator or some other simulator.
Crystal pointed out that she has received a number of requests
for a casa simulator, in particular one that could be used by
attendees of her Disks meeting in June, and that this was a
requirement for the offline package to provide a simulator.
It was stressed by Steve that the planned work would minimally
impact the other development for the beta and that this would
be evaluated when Joe and Kumar go to CV in late March.
o At some point (soon) it will be necessary to know what
capabilities are required to be in the Beta release. What
modes of VLA data must be fully supported and what modes of
calibration and imaging must be available.
ACTION: Steve will make a first crack at this, though the
EVLA reps will eventually need to provide more input.
Standard items:
o EVLA test results (Joe, Michael)
Michael has produced a draft report from the EVLA test.
See:
http://www.aoc.nrao.edu/~smyers/naug/reports/evlatest_06feb07_summary.txt
- See:
http://projectoffice.aips2.nrao.edu/EVLA2006.12/EVLA2006.12.html
o EVLA development plan & requirements
- see schedule at
http://projectoffice.aips2.nrao.edu
and on the CASA Wiki
https://wikio.nrao.edu/bin/view/Software/CASA
5. Main Event - Presentations and Discussion Items
***************************************************************
- Joe McMullin :
This was another installment in his weekly follow-up, with reports
on further progress. Focus is on preparations for the critical
upcoming ALMA SD and regression Test.
See:
https://wikio.nrao.edu/bin/view/Software/2007-Feb-14
=============================================================
New features and Issues Resolved:
As requested in the last NAUG meeting, the .last file
now contains a single-line call to the task function with all
params (commented out with #).
--nolog option for casapy (does not bring up GUI, sends
messages to shell window). Steve requested disabling of
message output, as you can tail the casapy.log file.
Will try this out.
plotxy has been undergoing extensive development by Shannon
and Urvashi (Shannon is at AOC this week).
"locate" is said to be available in the msplot (mp) tool, which is
what the plotxy task calls. This is the mp.locatedata
function, which reports data from a region selected using
the mp.markregion function. Not yet available from a task.
Note: I don't think this is exactly what was wanted, which is
a way to get this from the plotxy gui (with a button?).
Interested NAUG parties should clarify...
Joe thinks the cause of the seg-faults has been located and
fixed. Please let him know if you get any more!
Regression scripts:
These are coming along, and will be used for the test.
You can find these online for each dataset linked from the
test page:
http://projectoffice.aips2.nrao.edu/ALMA2007.03/ALMA2007.03.html
Some issues are still being resolved, such as the mosaic
scaling and some calibration inconsistencies. There are some
missing tasks such as a MEM imaging task.
There are also issues with filling PdB data. The old aips++
filler (based on the ALMATI-FITS format) is all that there
is since CASA will not support the obsolete ALMATI format
in favor of the upcoming ASDM. So the regression scripts will
start from the already filled ms (from previous tests).
The original scripts are being converted from aips++ glish
(using the old tools) to casapy tasks.
Towards the Beta Release (Steve)
In the NAUG, and also in the previous week in separate meetings,
there were a number of discussions related to preparing for
the big Beta release. It is clear the Beta is make-or-break
(as opposed to the survive-or-break ALMA-T5) and the key issue
that keeps popping up is (as Nicole put it) "the gap between
the requirements and expectations". Whatever is in the Beta
must work robustly and in a sensible manner, and there must be
enough ways to fully reduce common data (such at VLA) that
some core cross-section of users will be satisfied. We
(the NAUG and project liasons) will be working over the next
few months to be sure we have a good understanding of what
these issues are. For example, some common expectations
for things that were missing are:
- line flagging based on a continuum image
- interactive clean box setting
- full-function flagging that works well
This will be the topic of an upcoming NAUG, so stay tuned.
Note also that the Beta cannot contain buggy or non-functional
tasks or tools, so some stuff that we have in casa now may have
to be backed out for the Beta (such as incomplete simulator
stuff) unless they are fully functional. I know that
nominally "Beta" releases can be sloppier but I'm afraid that
we don't have that leeway...
As a result of some of these discussions, I have decided to
put together a series of "tiger team" meetings where I sit
down with a few people outside our usual NAUG group and try
to push through some of their real data. That way I can get
fast feedback on what works and what doesn't to input to Joe.
It also places less of a burden on these "outside" testers
as it will typically be a 1-day event. As the first Tiger
Team, Ylva Pihlstrom and Claire Chandler have agreed to give
it a go. We are tentatively scheduled for late March
(probably while Joe and Kumar are in CV). If this works
well I will schedule more in the run up to the Beta.
=============================================================
Other info:
Test page:
http://projectoffice.aips2.nrao.edu/ALMA2007.03/ALMA2007.03.html
Test prep:
https://wikio.nrao.edu/bin/view/Software/AlmaTest5
Testing spreadsheet:
http://spreadsheets.google.com/pub?key=p9Q-V26qhzpmEtAF5AD4gOg
***************************************************************
Future Topics
o Selection and (ms) flagging implementation plan update (All)
o Beta Release form and functionality checklist (Steve)
o suggestions welcome
6. AIPS++ Developments - see latest targets and info at Project Office:
http://projectoffice.aips2.nrao.edu/
7. Upcoming meetings and deadlines:
o 2007 Feb 12 ALMA pre-testing starts
o 2007 Mar 5-9 ALMA external test (Socorro)
o 2007 June 18-21 NRAO 50th Anniversary Symposium (CV)
o 2007 June 22-24 Meeting "Transformational Science with ALMA:
Through Disks to Stars and Planets" (CV)
The agendas for past NAUG meetings are archived at:
http://www.aoc.nrao.edu/~smyers/naug/agenda/
The minutes for past NAUG meetings are archived at:
http://www.aoc.nrao.edu/~smyers/naug/minutes/