after the 0.9b: 0.9RC1 then 0.9.0 and 0.9.1 for bugfix of minor adds before 1.0
create SVN branches/tags/trunk for the first stable release

v0.9RC1	WEBSITE: make some tutorial videos and update documentation for recently added features
v0.9RC1	WEBSITE: advertise on cloudynights, stargazers and other famous UK/US forums
v0.9	CODE: parallelise slow algorithms on CPU cores
v1.0	CODE: evaluate GPU computation of HDR mapping or colormapping for display acceleration
v1.0	CODE: fix arrays of fixed size 256, should often be dynamic. Needs to be done for file names, in sequences, preprocessing and so on
v1.0	CODE: seq files are loaded when searching for sequences in the dir to identify of SER sequences. A type-only loading could be done to avoid allocating memory in readseqfile for nothing
v2.0	CODE: replace com.seq(sequence) by com.seq(sequence*)


TODO: command line and console
	: test all commands and make the command line robust
v0.9	: reactivate / verify / test the ddp function -> is this depracated with the new Midtone Transfer Funtion ?
v0.9	: update maximum of commands in the graphical interface
v1.0	: reenable the lrgb and rgb commands
v2.0	: make a console version of Siril, for preprocessing at least


TODO: import, conversion, sequences
v0.9	: add the use of the dark-flats
v0.9	: save regparam and imgparam.quality in sequence file
v0.9	: remake sequences to work with strings instead of numbers
v0.9	: fix spaces in filenames for the load command and sequences
v0.9	: film import: fix frame loading (delagated to FFMS2, but film conversion is possible)
v0.9	: allow any sequence to be imported and converted on-the-fly instead of requiring implicit conversion
?	: save hi and lo for each layer in the sequence file (it's done for all layers in the FITS)
v1.0	: the show_excluded variable is not used, the checkbox is useless!


TODO: preprocessing
v0.9	: manage sequences instead of master files for offset, flat and dark
v1.0	: apply a function or command to sequences (modifying files) like flipping images or croping them
v1.0	: for native sequences (SER and FFMPEG), there is no preprocessing and should never be. It requires an on-the-fly pre-processing. This technique could also be used for FITS sequences, if the user prefers to wait a bit more instead of creating new pp_ files.


TODO: analyse
v0.9	: provide a better indicator of planetary image quality and enable stacking filtering on it
		- what is the quality value (given by function quality()), should it be displayed?
		- one solution to quantify quality for stellar images is to use both the Strehl ratio on unsaturated stars (http://en.wikipedia.org/wiki/Strehl_ratio) and roundness given by FWHM.
		- for planetary images entropy can be used (since 0.9b)
		- sigma, the background noise standard deviation is a good indicator for deep sky images
v0.9	: add the "cut" function of Iris, allowing a 2D visualization of pixel values over a user-drawn line
v1.0	: give a reference magnitude for a star and allow star selection in the PSF to be done on magnitude


TODO: registration
v0.9	: registration data is stored per channel, it should be copied on all layers by default, with a clear indication (planetary images should not use the same alignment between all layers when an ADC (Atmospheric Dispersion Corrector) was not present in the setup.
v1.0	: sub-pixel alignment? is it possible without transforming images as wavelets and handling them as such for processing?
v1.0	: manual rotation method (possible for centered rotation on gray images)
v1.0	: Add Drizzle x2 and x3 (using OpenCV to resize image), also called : Variable Pixel Linear Reconstruction
		http://www.stark-labs.com/craig/resources/Articles-&-Reviews/Drizzle_API.pdf
v1.0	: add automatic rotation method
		algorithms used by deep sky stacker:
		http://adsabs.harvard.edu/cgi-bin/nph-bib_query?bibcode=1995PASP..107.1119V&db_key=AST&high=39463d35aa24090
		http://lear.inrialpes.fr/pubs/2004/MR04/
		See also: http://pixinsight.com/doc/tools/StarAlignment/StarAlignment.html


TODO: processing and display
v0.9	: remove blurred display when zooming
v0.9	: implement median stacking for SER and AVI sequences, probably with read_partial_frame
v0.9	: allow the default rendering mode to be different than linear (histeq in particular)
v0.9	: zoom with wheel or + / - keys centered on the mouse cursor position if on image
v0.9	: enable preview in the unsharp mask dialog.
v0.9	: fit the RGB picture to the RGB window. Independant behaviour (?)
v0.9	: reenable a proper hi/lo setting with button 2
v0.9	: add of spline method to extract background.


IDEAS ON IMAGE LIST/SEQUENCE (long term):
v2.0	CODE: use the image_layer_struct and allow for complex RGB compositing
each layer is related to a layer inside an image sequence
there may be different sequences for each layer
for each layer, should be displayed: layer name, originating sequence and layer, hi and lo scales, auto level (max range and dynamic range), a list of images with { image name, registration data, selected, FWHM, reference image }
for each displayed frame should be displayed: originating image, registration data, selection status in sequence, FWHM on registration selection, is reference image

