[08/AUG/2007] Nicolas Le Novère (lenov@ebi.ac.uk)

	Move from isolatin-1 to utf-8

	version 4.2h

[06/SEP/2004] Nicolas Le Novère (lenov@ebi.ac.uk)

	Fix mistakes in the calorimetric table xia98a.nn.

	version 4.2g
	
[20/MAY/2004] Nicolas Le Novère (lenov@ebi.ac.uk)

	Fix a mistake in the calorimetric table xia98a.nn. Thanks to Wong
	Chee Hong (wongch@bii.a-star.edu.sg)

	version 4.2f
	
[10/APR/2003] Nicolas Le Novère (lenov@pasteur.fr)

	Minor modification of the verbose output. 
	
	version 4.2e
	
[11/FEB/2003] Nicolas Le Novère (lenov@pasteur.fr)

	Modification of the documentation, regarding the correction of
	strand concentration.  

	version 4.2d
	
[03/OCT/2002] Nicolas Le Novère (lenov@pasteur.fr)

	Bug fix: An amazingly obvious mistake was introduced in the
	approximate mode. The fractional GC was used instead of the
	percent GC (Personnal view: The approximate mode is a crap
	anyway). 

	version 4.2c
	
[14/MAY/2002] Nicolas Le Novère (lenov@pasteur.fr)

	Bug fix: a break was forgotten in the big switch decoding the
	options. Also the output of long messages did not comply with ISO
	restrictions on string length. I split the strings. Thanks to
	Mauro Petrillo.

	version 4.2b
	
[18/FEB/2002] Nicolas Le Novère (lenov@pasteur.fr)

	The threshold between exact and approximative computation is now
	controlled by an option.

	version 4.2
	
[14/FEB/2002] Nicolas Le Novère (lenov@pasteur.fr)

	Correction of a typo in all97a.nn An entropic factor was changed
	from -22.5 to -22.7 cal.M-1.K-1

	version 4.1h
	
[19/NOV/2001] Nicolas Le Novère (lenov@pasteur.fr)

        Corrections of the script profil.pl

        version 4.1g
	
[18/MAY/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Typographical mistake in decode.c:read_mm()
	If the environment variable was undefined, or there was no
	readable file at the specified place, the default defined at
	compilation was not read. 

	Several minor editing of the error messages.

	Modification of the Tk interface.
	
	version 4.1f
	
[03/MAY/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Big mistake in documentation (relative to the correction of
	nucleic acid concentration). Took advantage to improve the
	documentation (including benchmark results).

	Minor correction of usage function.

	version 4.1e
	
[29/APR/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	No need to enter a useless concentration of nucleic acid when
	using the approximative mode. Small (but important) edit of the
	script multi.pl. Addition of a benchmark in the directory TEST,
	which compare various methods, NN sets, and salt corrections. 

	version 4.1d
	
[20/APR/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Correction of several critical bugs in the reading of paramaeters
	for mismatches and dangling ends. The format of the file containing the 
	parameters for mismatches is modified (the undetermined Crick's pair 
	were previously commented out. They are now included with parameters 
	equal to 99999).

	version 4.1c         

[10/APR/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Correction of the code to prepare the implementation of dnarna and
	rnarna mismatches and dangling ends. Correction of the makefile
	for unices

	version 4.1b
	
[29/MAR/2001] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Implementation of the dangling ends analysis.

	Addition of a several security corrections (strcpy -> strncpy,
	check of malloc success ...).

	Improvement of the makefile for Unices. Simplification of the
	declaration of PATH for the nnfiles.
	
[07/AUG/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

        version 4f. Mistakes in the NN file san96a.nn found by Leo
	Schalkwyk
	(l.schalkwyk@iop.kcl.ac.uk)

[05/JUN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	version 4e. Correction of the output to stderr et stdout (number
	of spaces leading the lines). New perlTk interface.

[26/APR/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	version 4d. Typos in the code and the perl/Tk interface (unit of
	entropy was given as J.Mol-1 instead J.Mol-1.K-1)

[24/APR/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	version 4c. Addition of a script (profil.pl) allowing an iterative
	computation along a sequence, by sliding a window of a specified
	width. Minor modifications of the documentation.

[09/FEB/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	version 4b. Some fixes of the program output and of the
	documentation.  Handle of the references in the parameter files
	modified.

[08/FEB/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	MELTING4 release version. Corrections on the advice of Nicolas
	Joly.  Improved documentation.

[05/FEB/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Changed the behaviour of options -v and -q. Improved
	documentation.

[29/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Fixed bugs in the interactive mode. Improved online help (option
	-h).  Fixed a lot of bugs or unpleasant features in the option
	parsing.  Changed the behaviour of options -A and -H. The former
	has now always precedence over the latter regarding the et of NN
	parameters.

[27/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	Second beta version. Correction of some NN files. Modifications of
	documentation.

[26/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	First beta version. The program can now take in account mismatched
	hybridisations. Correction of a bug in the screening of
	Crick'pair.

[24/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	alpha5. The program can now read a file containing the
	nearest-neighbor parameters for some mismatched Crick's pairs.

[18/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk) 

	on suggestion of Nicolas Joly (njoly@pasteur.fr) alpha3 and
	alpha4. Code correction only.
	
[17/JAN/2000] Nicolas Le Novère (nl223@cus.cam.ac.uk)

	alpha2. Modified the system to localize files containing set of nn
	parameters. Now a directory is defined at compilation by
	preprocessor directive. However, an environment variable can be
	defined, which targets a directory where to find a nn file. If it
	was not found there, it will be searched in the default
	directory. Therefore, a user can define its own set of parameters,
	while still using those installed with melting.
