From: "Dennis, Brian R." <Brian.R.Dennis@nasa.gov>
Subject: Thoughts on the Goddard HEDAW
Date: Sat, 10 Feb 2001 10:28:29 -0500

While they are fresh in my mind, here are my thoughts on the Goddard HEDAW. 
I would encourage everybody else who was there to send me your comments. Then we
can discuss them on the Tuesday telephone call and decide what changes 
should be made for the Berkeley HEDAW.

Generally, I thought the HEDAW went remarkably well and I congratulate all 
of the organizers and presenters for doing such a wonderful job. In 
particular, I thought Chris Johns-Krull's write up proved extremely valuable. Many people 
used this as their basis for learning how to use the system. The graphical user
interface was also a very popular feature. Some people also were really 
getting into the objects and how to extract more information from them than is 
available through the GUI. I got a lot of positive feedback from the attendees.

Surprisingly, there seemed to be relatively few installation problems. 
Most people with laptops came pretty well prepared with IDL and SSW already
installed. However, I did think that the first morning's presentations went 
on for too long before people had a chance to actually try to use the software 
and do some simple tasks. The talks didn't finish until after 11:30 so people 
only had a half hour before lunch and the next set of presentations on imaging 
in the afternoon. There was plenty of time later in the workshop so we shouldn't 
try to cram so much into the first morning.

The pacing for the rest of the workshop was more relaxed, I thought, with 
plenty of time after the presentations for people to try out the different aspects 
that they had just heard about. I got some positive feedback about my 
semi-jocular "assignments" after each session so perhaps we should make them a little 
more formal and think about what would be appropriate - perhaps have some worked
examples or something. We did have quite a few people stay over for the 
third day even though nothing formal was planned. Perhaps that would be a good 
time to go into more details about objects and the inner workings of the software 
for the real aficionados. Anyway, here's my suggestions for changes:

1. Let's have paper copies of Jim's presentation available so that he 
doesn't  have to go through every single item on his viewgraphs but can just 
refer  people to the general type of item on the lists and have them read it
later. That way, Jim could shorten his presentation and help to 
alleviate  the time crunch on the first morning.
 
2. Another way to alleviate this time crunch would be to put the second 
half  of Andre's talk to a later time. I thought that the detailed 
discussion  about objects at that time went over most people's heads. They hadn't 
had a  chance to understand the basic aim and scope of the software by 
trying to do simple tasks themselves so they couldn't grasp at that stage the 
need  for or the advantage of objects over any other scheme. I would 
suggest  having Andre review the installation and set up of the software on 
PC's and Unix systems and give a brief overview of the software but leave the
detailed discussion of objects until later, perhaps as late as the 
second afternoon, with the really specialized stuff on the third day.
 
3. Kim's presentation of the GUI is a good lead in to a first simple
assignment to obtain a light curve for a selected flare, say the 
brightest one in the simulated data. That makes sure that everybody's setup is
correct and starts people off slowly in using the software. One thing 
that should have more emphasis at some stage is the capability to switch 
back from the GUI to the command line interface. It needs to be made 
clearer how to do this (in the write up) and its value in allowing you to use the 
GUI for what its good at without losing the flexibility and full 
capability of IDL. Perhaps this could be mentioned by Kim at this time with more 
details later when Andre talks about objects.

4. We need a Chris Johns-Krull write up for spectroscopy. People need a
printed and an electronic version of the script of necessary 
commands to do the basic spectroscopy. It's not a good idea to try typing in the
commands in real time, especially for the introduction when people 
are trying to learn how to do the basic operations. That would be more
appropriate for advanced students on the third day.
 
5. Ron Murphy's presentation of the gamma-ray line code was useful, I 
thought.  We need somebody to present it in Berkeley. Perhaps Richard can do that.

6. We had barely enough space in the main meeting room for everybody to 
sit down for the presentations. Make sure that there are going to be 
enough seats for everybody at SSL. There was enough space and network 
connections available for laptops. The nine UNIX terminals seemed like more than 
enough as many people without laptops preferred to use the five PCs we had 
set up next to the main meeting room.
 
7. Let's have a printed list of assignments for people to try after each
series of presentations. Here are my suggestions:
After the first mornings presentations, generate a log-linear light curve using the
GUI for the largest flare in the simulated data set. Provide the flare number, peak rate, 
and FWHM duration of the flare in the 20 - 100 keV energy band.
After the imaging presentations Reconstruct images at the peak of the default
flare using back projection, clean, MEM-Sato, and MEM-VIS. Compare and contrast 
the results After the spectroscopy presentations
Construct a spectrum at the peak of the default flare and fit it to a thermal
plus power-law function. Report the temperature, emission measure, and power-law
spectral index.   After the ancillary data presentations Overlay a HESSI image at the peak
of the default flare on to an H-alpha image nearest in time. Access the HEDC web site and
do something unique to prove you could use it.
 
8. We should set up a bulletin board and ask (demand?) that everybody 
using HESSI data indicates what research activity they will be conducting - 
what flares they will be using, what topics they will be addressing, what
ancillary data they will be using, who they will be collaborating 
with, etc. Perhaps we need to generate a form asking for this information 
in some standard format. It might make it easier for people to provide the
information that way than if we just asked them to send it free-form.
 
9. We should develop a list of frequently asked questions (FAQs) with 
answers. People that were at the workshop should send me questions that they 
were asked with the corresponding answers and I will assemble an on-line 
list of FAQs.

Here's the first FAQ with answers that we generated at the workshop:

Question: What should I do if IDL crashes when running a HESSI routine?


Answer:

1. Note the statement that appears in the IDL text window when the
crash occurs. See if it helps you understand what went wrong.
 
2. Type retall and see if you can rerun the program and repeat the problem.
 
3. Exit and then restart IDL.

4. Exit IDL, update to the latest version of SSW. Make sure you are
using the appropriate released or developmental version. Restart IDL.

5. List the local variables by typing the following immediately
after the crash occurs:
help, /name='*'
See if you can determine if any of these variables are causing the problem.

6. Capture the IDL path by typing the following:
print,!path
See if this could be at fault.

7. If all else fails, email all the above information with a
description of what you were trying to do to the appropriate
HESSI team member with a cc to Brian Dennis at
Brian.R.Dennis@nasa.gov.

List of HESSI Team Members with their areas of expertise:

--
Brian R. Dennis

Responsible government official:    Brian Dennis   Brian.R.Dennis@nasa.gov

              Web design:    Merrick Berg     Merrick.C.Berg.1@gsfc.nasa.gov

This site last updated:  February 20th 2001

This site last updated November 10, 2008.