Subject: JEM-X imaging software - how to start?


Dear Paul:

It was very good to meet you in Helsinki and have an opportunity
to discuss your future work in the JEM-X software team. You 
certainly bring to us a lot of very relevant expertise.

Trying to explain my imaging software to you was a rather unpleasant
experience, because I realize how much of a unstructured patchwork 
it has now become.

You asked me how I output my images and as I told you: I dont look
at images - but I suspect that doing that, looking at images, could 
turn out to be extremely useful for understanding better both the 
instrument and the software.

Below I suggest a way to get going. Please comment.

With best regards

Niels

PS: I will send you also a collection of mails and notes describing
midisky8.

*****

A suggested initial work plan:

1) You start by taking the Midisky software complex as it is and 
get it installed and running on your linux machine. The instructions 
for doing such an installation exist (I think you already got them 
from Silvia), and both Silvia and Jerome Chevenez here have succeeded 
with such an installation. 

The software actually already output some images, the format should 
possibly be changed, but that you can do yourself once you know the 
arrays in which I store the raw image and the residual images. 
I will be available for e-mail or phone discussions as need be.
It would be of immediate use for me (and for Victor too I suppose) 
if we can begin to produce some nice JEM-X images. Your knowledge 
of existing image display packages could be essential here.

I have thought more about how to assign weights to the individual
skymaps when we want to combine them into a mosaic, I think I now
see a way forward and I expect to distribute a short note next week.
We are planning here at DSRI to make a PR-event in connection with
INTEGRALs "birthday" in October, and for that we want to make a
"galactic plane seen by JEM-X" image. So I have my own interests
it getting this image processing going. But I dont want to 
reinvent the wheel when it comes to the actual display routines
if I can avoid it.

Once we see how the midisky images really look under different
conditions of source and background fluxes this will help us decide
whether you should take midisky as a starting point for improved
JEM-X imaging software at ISDC or you think you can put together
something better or equally good using pieces with which you are
familiar.

***

2) I start to clean up the midisky software, introducing a better 
structure. It is obvious (now!) that I should take the processing 
of a single shadowgram out into a subroutine rather than doing this 
all inside a very messy main program. 

But there are limits to how much time I can use on such a clean-up 
process, so I will concentrate on removing dead and superflouos code
and improve the overall structure. 

I suppose that you may at a later time anyway take the software apart 
and replace part of it with better things, but that job will be easier 
if the software already has a minimum of structure.

***

3) Niels Joergen has put together a package with all the routines he 
has prepared for implementing an earlier version of my program at ISDC.
He will send you an e-mail about where you can find it. I am sure that
much of what he has done can be directly taken over by you or at
least serve as inspiration if you take up the challenge of transforming
midisky for ISDC.