Subject: brief diskussion of JEM-X OSA software
 8bit

***** The primary uses of JEM-X and the OSA software

1) Detection of new sources and of transient eruptions of known sources.
   (source detection in individual science windows)
   j_ima_iros
   (j_ima_cross ? + j_src_locator)
   (source detection in mosaics of multiple science windows)
   j_ima_iros + j_ima_mosaic + j_src_locator

2) Determination of source spectra.
   (extraction of spectral parameters from individual science windows)
   j_src_spectra
   j_ima_iros
   (j_src_properties ?)
   (extraction of spectral parameters from mosaics of multiple scw's)
   j_ima_iros + j_mosaic + mosaic_spec

3) Determination of light curves
   (light curves at short time scales << science window duration)
   j_src_lc
   (j_src_properties ?)
   (light curves at long time scales - multiple science windows)
   j_ima_iros + j_mosaic + mosaic_spec

4) Period folding based on PIF selected event lists
   j_src_extract_lc ?
   (j_src_properties ?)


***** Development status of software packages:

 Source detection:

Original package (j_ima_basic_recon) now replaced by j_ima_iros.
j_ima_iros is developed by NL and NJW. It is based on an Iterative-
Removal-Of-Sources (IROS) algorithm, and on a Pixel-Illumination-Model
(PIF). The pixel illumination model is very detailed and is separated
out in a subroutine package so it is available for the j_src_properties
software package to be discussed below.

Development status: j_ima_iros (in OSA 7.0) is still containing some
bugs which degrades the source positions and fluxes. We believe we have
solved these problems by now and a revised version has been delivered
to ISDC but not yet released in a new software release.

Further, j_ima_iros (in OSA 7.0 and a possible OSA 7.1) does not
correct for the significant variations in electronic efficiency
(pulseheight dependent) which are known to exist in JEM-X. A new
version of j_ima_iros which will include corrections for electronic
efficiency is almost ready (have you ever seen that statement before ?)
 - and may become available for OSA 8.

Problems: The detailed modelling involved in j_ima_iros could mean
that the results produced are prone to errors and noise types which
could be avoided with a simpler analysis. The main issue is the
handling of the collimator shadowing, which can only be taken into
account with a finite precision. An alternative analysis imaging
method which does not attempt to do any detailed modelling of the
collimator has been developed by Carl as a stand-alone IDL package.
This software package is now being ported to the OSA environmment by
NJW as j_ima_cross.

The disadvantage of not modelling the collimator is the the signal-
to-noise ratio for off-axis sources is lower than it could be. At
this moment it is not clear which imaging method will work best.


 Source spectra:

The original package, j_src_spectra, is still the OSA-7.0 default
tool for spectral extraction. (mosaic_spec is an alternative).
However, j_src_spectra (and j_src_lc) was developed by Stefan Larsson
which has not working on this project since several years.

We therefore intend to develop a new spectral extraction package,
j_src_properties,  based on the PIF-subroutine package developed for
j_ima_iros. Of course, the same questions about detailed modelling
which were discussed above comes in here, but for light curves on
short timescales there is no way around using the pixel illumination
function to select source relevant events from the event list.