1. Notable changes from 12 Oct 1994 release to 26 Nov 1994 release

1.01 plot2 : option for contouring timer

	There is now an option in the levels module which allows
	a timer to be on or off during the contouring process.
	This is in case a sensible contour level is unknown.
	The default is that the timer is on.

1.02 process : commands given as comments in par file

	The input file, output file, script file and script
	commands are now written to the output par file.
	This is to provide a record of the processing done.

1.03 plot1 : loading object module parameters

	The code to load (saved) object module parameters
	(e.g. whether the x axes are visible, etc.) has
	been added (it was missing due to an oversight).

1.04 plot2, contours : fixed contouring bug

	There are two (similar) routines in the contouring
	code which deal with the case when there are two
	contour lines crossing a particular square.  One of
	these routines had a backwards if statement, thus
	instead of
		if (X) then A else B
	the code should have been
		if (X) then B else A

1.05 plot2 : fixed ppm bug

	There were various places where using a region defined
	by ppm instead of points created a problem, e.g. when
	presssing the whole button in the region module, when
	trying to go to a saved region, and when trying to use
	the tracking display to move the display region.  This
	arose because of the backwards orientation of ppm space.

1.06 plot1, plot2 : more paper sizes in output module

	The paper size options now are A4, A3, A (used to be
	called US Letter), B and 'other'.  The 'other' option
	comes with a text field, in which user defined paper
	sizes may be given.

1.07 plot2 : fixed extracting subset bug

	The code for extracting chosen subsets had a bug in it
	to do with the step size (it only manifested itself in
	certain circumstances).  Instead of
		N = M / S
	the code should have been
		N = 1 + (M-1) / S

1.08 plot1 : fixed lower and range bug

	The code for the lower and range commands in the script
	file had a minus sign where there should have been a
	plus sign, thus instead of
		D -= V
	the code should have been
		D += V

1.09 plot1 : fixed plot_y multiple drawing lines bug

	The code for plot_y was written in a lazy way, in that
	independent of the displayed range the entire array was
	plotted.  Normally this is fine because X does all the
	clipping, but when looking at small ranges for large
	arrays there was a wrap-around effect which resulted in
	multiple lines being drawn on the screen.  (This may be
	a 16-bit problem.)  The code now only draws those parts
	of the array in the chosen display range.

1.10 plot2 : added ruler options in 1D display

	There used to be no y ruler drawn on the 1D display and
	the x ruler was always drawn.  Both x and y rulers are
	now available, and they can be made visible or invisible
	by selecting the appropriate toggle in the slice module.

1.11 plot2 : automatic toggling in some cases

	In the output module, typing in the 'other' text field
	automatically makes the 'other' toggle get selected.
	Similarly in the extract module, typing in any of the
	first, last or step text fields automatically makes
	the chosen subset toggle get selected.

1.12 plot1 : automatic toggling in some cases

	In the output module, typing in the 'other' text field
	automatically makes the 'other' toggle get selected.
	Similarly in the region module, typing in either the
	x region or y region text fields automatically makes
	the chosen fit region toggle get selected.

1.13 plot1 : fixed texts module bug

	There was a bug in the code that created the texts
	module so that the program crashed if there was more
	than one text.  The code loops over each text field
	that needs to be created, and inside this loop only
	one text field should be created, so instead of having
		for each of N texts create N textfields
	the code should have been
		for each of N texts create 1 textfield

1.14 plot1 : scrollbars if too many sliders, buttons or texts

	The original code did not have scrollbars for the
	slider, button or text modules.  Thus there was an
	upper limit to the number of sliders, etc., that
	could be used (depending on the screen size).  This
	has been changed so that scrollbars appear once
	there are too many sliders, etc.

1.15 plot1, plot2 : no value change upon slider selection

	When a slider is clicked upon to bring it into active
	focus, the slider now does not change in value (second
	and subsequent clicks do change the value, of course).
	This separates out the two distinct actions (bringing
	a slider into focus and changing the slider value).

1.16 plot1 : baseline correction added

	The baseline correction routines have now been added.
	These are base_const, base_const2, base_poly,
	base_poly2, base_trig and base_trig2.

1.17 plot2 : fixed slice drawing bug

	If the display region did not intersect the region
	defined by a data set then the program crashed.  This
	is because the size to draw became negative, and this
	condition was not trapped.

1.18 azara : added html files

	A new directory, html, containing html versions of the
	help files, and some other files, has been created.

1.19 process : fixed multi-dimensional commands bug

	If the number of points in any dimension was not a
	multiple of the block size in that dimension for
	multi-dimensional commands (currently only maxent2
	and maxent3) then the processing would not work as
	it should (although the problem might not actually
	show a noticeable side effect).  This is because the
	(four) inequalities that checked for this condition
	all had '>' instead of '>='.

1.20 process : added Hilbert Fourier transform

	Added the Hilbert Fourier transform as a command.

1.21 process : added non-linearly sampling in maxent

	Ordinarily experiments are done in a manner that
	is linear in time, that is, the sampled points
	are given by t = T+1, T+2, T+3, T+4, ..., for
	some time T (and each time actually occurs twice
	if the data is complex in the given dimension).
	The maximum entropy code now allows the original
	data to be non-linearly sampled if desired.

1.22 process : fixed file initialisation bug

	The C files were not explicitly initialised, which
	caused a crash when the operating system did not
	initialise them to NULL.  This was only a problem
	because of the way the files were opened.  The code
	for the latter has now been changed so that it does
	not matter what the initial values of the files are.

Azara help: CHANGES-1.0-1 / W. Boucher / azara@bioc.cam.ac.uk