"DosyProc" Version 2.1
	for XWIN-NMR on Windows NT
	Niko Loening, October 17, 2000

	Copyright (C) 2000 Niko Loening
	All Rights Reserved

Index

	1. About DosyProc
	2. Notes
	3. Requirements
	4. Installation
	5. Usage
	6. Hints
	7. Version Changes
	8. Known Problems

-------------------------------------------------------------------------------
1. About DosyProc

DosyProc creates a DOSY spectrum from a series of diffusion weighted spectra.
It uses linear regression or non-linear least squares (mono or bi-exponential)
to determine the diffusion constant for each point and then creates Gaussian
peaks in the diffusion dimension based on this information.  The program
features an easy to use interface and on-line help throughout.

-------------------------------------------------------------------------------
2. Notes

-> DosyProc is freely available for academic use.

-> There is no warranty for damages caused by using this application.

-> Without permission from the author (Niko Loening), you may
   not distribute the program.

-> For requests, questions, or bug reports, contact the author by e-mail
   at "Niko_Loening@alumni.hmc.edu".

-> You may find the latest version of DosyProc at the following site:
              http://www-keeler.ch.cam.ac.uk/downloads

-> If you are using Windows 2000, the program will not automatically show
   up as the topmost window when started.  To change this behavior, use the
   Windows registry editor (regedit.exe) to change the \HKEY_CURRENT_USER\
   Control Panel\Desktop\ForegroundLockTimeout registry key from a decimal
   value of 200000 to 0.

-------------------------------------------------------------------------------
3. Requirements

This program was created for XWIN-NMR 2.5 and 2.6 running under Windows NT 4.0.

-------------------------------------------------------------------------------
4. Installation

1) Download:

	dosyproc.exe
	dosyproc.hlp
	dosyproc.cnt

   and place all three files in the XWIN-NMR\prog\au\bin directory.

2) That's it.

-------------------------------------------------------------------------------
5. Usage

1) On the XWIN-NMR commandline (where you type commands such as "ft" and
   "ased") type "dosyproc".  XWIN-NMR is case sensitive, so make sure you use
   lower case letters.

2) If you are not in a 2D data set, the program will generate an error.  If
   you are in a 2D data set, an application window will appear.

3) Flip through the pages and adjust the settings as needed.  There is
   on-line help to explain what the different variables are for.

4) Click on "Apply" to begin processing your data.  The XWIN-NMR status line
   will keep you updated about how the calculation is proceeding.  When the
   calculation is finished, the new DOSY spectrum will be displayed.

5) Click on "Cancel" to exit the program, or adjust the settings and click
   "Apply" to rerun the calculation.

-------------------------------------------------------------------------------
6. Hints

-> If the program doesn't start, make sure that you have version 6 of the
   Microsoft C libraries installed on your system.  These are automatically
   added to your system when you install Microsoft Visual C++ Version 6.0.

-> Everytime you click "Apply" the current settings are saved to a text file
   in the processed data directory for future reference.

-> If you get complete garbage, make sure the boundaries are okay.  Sometimes,
   if there is a lot of noise in the spectrum or the threshold is set to low,
   the automatic boundaries routine gets confused.  In this case, you should
   to try setting the boundaries to "Manual" and setting the range to 0 to 50.

-> The equation used for fitting the data is only exact for rectangular
   gradients.  For other gradient shapes, the shape factor gives only an
   approximate answer rather than the exact result.  However, as long as the
   gradient pulse length is much less than the diffusion delay, the results
   will be very nearly correct.

-> The results should be fairly similar for linear-regression and mono-
   exponential fitting.  Use the bi-exponential fitting at your own risk, it
   tends to add more errors to the spectrum than it removes.

-------------------------------------------------------------------------------
7. Version Changes

Version 1.0 was the original release version.

Version 1.2 featured a complete rewrite of the gradient dialog page.  This
allows for the gradient length or shape to be incremented instead of just the
gradient strength.

Version 1.3 added different scaling options on the display page.

Version 2.0 added non-linear least squares fitting, a rewrite of the data
processing methods, and several small bug fixes.

Version 2.1 added the "Bi-Exponential" page so that the user can
influence the starting parameters and rejection filters for the
bi-exponential fit.

-------------------------------------------------------------------------------
8. Known problems

There are no known problems at this time.  However, if anybody knows a better
algorithm for performing non-linear least squares than the one given in 
"Numerical Recipes in C", please let me know.