[Xrays@aps.anl.gov] Input on needs for scientific software

[Daniel Haskel]

1. What are the limitations of current tools for
   x-ray data reduction, analysis, modeling, and simulation?

Perhaps a distinction between needs during the experiment and after it is in
place. I believe there is a need for a flexible package for data 
manipulation for the beamlines to be used to evaluate the quality of
data and perhaps some preliminary analysis. Some of this is too
specific to implement in a broad sense, like XAFS and powder
diffraction, and there already
exists good packages for that (ATHENA, ARTHEMIS, IFEFFIT etc...). I see need
for a package that can read in multiple files, operate on columns
within a single file or between files, performing a large number of
mathematical operations (log, derivative, smoothing, integrals). This
should be accompanied by a nice GUI with good plotting capabilities,
allowing plotting of data columns from multiple files, results of
operations, zooming, cross-hairs, etc.

A number of packages do some of this, such as Newplot or the data
processing capability within XOP (XPLOT). Both of these handle SPEC files,
they can plot columns from more than one file, but they can't operate
between columns from different files. A very simple example is wanting
to calculate the rms variation in some data using 5 consecutive measurements.
Another is to take the difference between two data sets taken with
opposite applied field directions. Yet another is to fit a
polynomial background to a data set, subtract the background from the
data, and save/plot the background-corrected data. These are all "trivial"
things that we all do with our favorites IDL routines, C-PLOT, spreadsheets,
or whatever it may be, but I don't believe that we have here at the
APS a GUI/data manipulation software for the beamlines that is generic
enough that it can be effectively used by all. Newplot seems to be a
good beginning, but it is far from being very useful.

My personal opinion is that good beamline-based data-manipulation
software is very important to make the user's experience the best it
can be. This largely falls on the beamline scientists, resulting in
various sector-specific tools which make it harder for an experimenter
to move from sector to sector. A generic
powerful package for data manipulation will be a huge improvement to
all beamlines. 


2. What additional tools are needed?

See no. 1

3. How can the existing tools be improved?

See no. 1

4. What will most affect the scientific impact of your work?

Some users, especially those new to the synchrotron radiation arena,
need lots of help with data analysis. Since the beamline scientist
time is limited, this delays the turn around time between experiment
and publication a lot. It might be worth considering having a couple
of people in the theory group that do modeling/data analysis work in a
couple of major areas such as scattering and spectroscopy. These
people will be strong in modeling and could work with users that need
help in modeling their data. I'm not talking about exoteric
theory but simply help with analyzing data. Outreach activities such
as summer schools, etc, to train people in the use of existing/new
packages is also a great complementary way to pursue this problem.


--------------------------------------------------------------------

Dr. Daniel Haskel
Physicist, Magnetic Materials group                 

Advanced Photon Source, Bldg.431/E008          
Argonne National Laboratory
9700 S. Cass Ave.
Argonne, IL 60439, USA

Tel: (630) 252-7758
Fax; (630) 252-7392

http://www.aps.anl.gov/xfd/people/haskel/dani.html