|
| |
Feedback from Data Analysis Workshop
Subject: [Fwd: Feedback from Data Analysis Workshop]
Date: Tue, 16 Nov 1999 16:18:51 -0500
Gordon Hurford wrote:
After what I think was a quite successful Data Analysis Meeting
on
Friday, here is an attempt at summarizing the feedback we
received.
Since we each hear comments (and especially criticisms) through
different
filters, please review your own notes and recollections and
send around
any corrections or additions. In the list that follows, I
have combined
user comments with issues that came into better focus as a
result of
various discussions. Some items are clearly must-do's while
others should
be discussed. In addition, I haven't tried to indicate which
is which,
but rather just to list the items so we can deal with them
at our next
meeting.
The following are organized by topic:
Data Issues:
Need transmission feasibility tests for HESSI data files SSL-GSFC an
SSL-ETH.
Need to clarify handling of delayed t/m in the case, for example of
downloads to other ground stations.
Need for closer coordination with developments at ETH
Suggestion to distribute CD of quick look catalog in addition to web
access.
Follow-up on possible HESSI data center at NRO.
Quick Look data:
Is 3-20 kev the best energy range for identifying flares? Should
there be additional energy ranges ?
Images in 8 energy bands seems optimistic, in light of our original
goal of about 10 images per day.
User packets are not yet eliminated in generating level-0 data.
Need to optimize/test variance measure - which grids, statistical
limitations, integration time....
Need to clarify how we handle spin period changes in QL rate data.
Possibiliy of adding a velocity vector to QL data.
Need to review default parameter/algorithm choice for coarse flare
location.
Add particle detector output to list of flags or rates
HSH strongly urged the inclusion of parameter values / software
version to QL data.
Add ground station ID to QL data.
Choice of coordinate systems for position specification in QL
data (and final images ?
Imaging:
Need for speed !!!!!
Need a more efficient MEM algorithm with ~2000 rather than 20000
modulation patterns
Progressive display to allow user to cut off CLEAN iterations
intractively.
Multiscale CLEANing ?
Photometry requirements need to be looked at carefully
Possibility of implementing forward-fitting for point sources
first, and then using intensity ratios for different subcollimators to
deduce diameter of symmetric (or asymmetric) gaussians.
Challenge posed by AB for imaging microflares, etc.
Apply the analytic interpolation used in Spectroscopic calibration
matrices to Grid response matrices.
Need to clarify terminology re grid chararcterization matrices
(1 or 2 grids, attenuators, blankets?)
BRD wants color movie (32 or 64 frames) within 24 hours of big
flare.
Perceived mismatch of "4 hours to process 10^3 photons".
Principle component analysis to determine eigenvectors (independent
of data) ahead of time.
We should firm up our plans for spectroscopic imaging - need a
'backend' tool for this.
We need to come up with a parametric model of livetime correction
so we can test/parameterize it using the flight hardware.
Who will do the direct FT ?
Need to break out a separate simulation module with better interface.
Miscellaneous:
Need to document the decimation scheme and get it into User Manual.
Suggestion of a "Virtual Data Analysis Workshop" with objectives similar
to our spring workshop, except done over the net.
Need to set up a Majordomo list.
We should plan on getting our contributions pulled together into a coherent
set on a web site, possiblly at Montana ??
Plans for the Spring Workshop
The numerical results of the survey are
attached. Main results are:
94% indicated that would definitely come or were interested. (The level
of enthusiasm should be somewhat discounted, of course.)
Most potential attendees (58%) would bring there own laptop with 32%
requesting Unix workstations. There was little interest in being
provided with a PC.
The location preferences (52%:25%:18%) for GSFC, UCB and Zurich clearly
favored GSFC, but on the other hand, this was a reasonable match to the
geography of the participants in the survey.
there was little preference between 2 or 3 days duration.
Most (79%) potential participants would have some or a lot of experience
with SSW.
In the comments, several people mentioned academic conflicts in May.
RESULTS FROM THE 10/22/99 SPRING WORKSHOP
SURVEY
LEVEL OF INTEREST:
20
Definitely will come
13 Interested, but will decide later.
2 Probably will not come
~10
No response = (estimated attendance) - (swteam) - (responses received)
(For the following, "Interested"
responses are weighted a 1/2; "Not
coming" are ignored.)
Computer Preference:
15.5
58% Bringing a laptop
8.5
32% Unix workstation
2.2
PC
0.5
Watch and listen
Location: (Multiple
preferences were weighted)
6.7
25% UCB
13.9
52% GSFC
4.9
18% Zurich
1
No preference
Responses:
11% West coast
43% East coast
18% Europe
28% Japan and elsewhere
Duration:
8
30% 2 days
10.5
40% 3 days
8
30% No preference
SSW Experience:
7
26% Extensive
14
53% Some
5.5
21% What is SSW?
COMMENTS:
4 mentioned neutral sites with Japan mentioned twice
4 cautioned about academic conflicts in May
Reminder about 17-21 June SPD
Suggestion to add science-related content
Suggestion to preload software and data to eliminate need for network connection
Suggestion to work in pairs or 2 rounds.
Reminder to finalize schedule ASAP.
|
