2010-07-05:
 * MiRA 0.9.10 released.
 * Changed T_CHAR, T_SHORT, etc. to Y_CHAR, Y_SHORT, etc. to work with
   Yorick >= 2.2 and Yeti >= 6.3.1

2009-05-14:
 * MiRA 0.9.9 released.

2009-04-23:
 * MiRA 0.9.8 released.
 * Changes to use OptimPack-1.3 and to display Fdata and Fprior.
 * Complex visibilities can be fitted with (or without) Goodman
   approximation.

2008-12-12:
 * MiRA 0.9.7 released.
 * Fixed oifits.i for empty ARRNAME or INSNAME. 

2008-12-09:
 * MiRA 0.9.6 released.
 * Fixed a bug in oifits.i with oifits_new_* when master is provided
   (thanks to Thibaut Paumard).
 * Fixed a bug in mira_new_fft_xform with spatial frequencies exactly
   equal to zero (thanks to Stéphanie Renard).

2008-10-03:
 * MiRA 0.9.5 released.
 * Fix bug when no target is specified in mira_add_oidata.
 * In mira_add_oidata/mira_new, keyword cleanup_bad_data can be set
   to 0, 1, 2 to achieve different levels of filtering of invalid data.
 * Various fixes in oifits.i for creating/saving OI-FITS files (thanks
   to Sylvestre Lacour).

2008-09-26:
 * MiRA 0.9.4 released.
 * Added the possibility to select the target in mira_new.
 * Hack for FLAG column to deal with AMBER data.

2008-09-23:
 * MiRA 0.9.3 released.
 * Fixed a bug in __mira_build_coordinate_list__ which prevent to use
   data files with central frequency (0,0) measured.
 * New function mira_dirac.

2008-09-07:
 * Massive rewrite of oifits.i to optimize the code, make it easier to read
   and let the users create OI-FITS data on the fly and save it to a file.

2008-09-04:
 * MiRA 0.9.2 released.  This is the version used for the MiRA demonstration
   (see mira-demo.i) at SPIE 2008 Conference in Marseille (France).
 * Some change in plots.

2008-07-12:
 * MiRA 0.9.1 released.  This is the version used for the VLTI 2008 Summer
   School at Keszthely (Hungary).
 * Missing fft_utils.i is now part of the distribution.

2008-06-06:
 * MiRA 0.9 released.

2008-05-02:
 * MiRA 0.8 released.  

2008-01-31:
 * MiRA 0.7 released.
 * Orientation of u-v and image coordinates fixed to match astronmical
   conventions.
 * Use OptimPack1.
 * Noisy works again.
 * New Monte-Carlo optimizer.
 * Fix a bug in mira_add_oidata with loading of complex visibilities.

2007:
 * Penalty with respect to data is now computed for cartesian representation
   of a complex (not a polar).
 * Huge rewrite of code: all data of given type are collected in a single
   table (more compact representation and faster computation/plotting when a
   lot of interferometric data-blocks are fitted).
 * Wavelength selection is more flexible (you can specify the effective
   wavelength and bandwidth to consider).
 * Very fast approximation of the Fourier transform by means of FFT (or FFTW)
   and spectral interpolation.

