Contents of this distribution:

Files:
  COPYING:          copyright terms (GPL license)
  InversionKit.jar: compiled program (can be run directly - see doc.pdf)
  Makefile:         a makefile which will replace InversionKit with a new
                    compiled version (this should be identical to previous
                    version if no changes are made to the source files)
  README:           this file
  doc.pdf:          a short manual which explains how to use InversionKit

Folders:
  src:              contains java source files 
  classes:          contains java class files (files obtained after compilation)
                    including class files extracted from jlapack

!------------------------------------------------------------------------------
! List of changes:
!------------------------------------------------------------------------------
2.2:  - can handle discontinuities in models
      - improved spline treatment based on Dierckx package
      - improved splitting inequalities tests based on latest version of 
        Reese (2015, A&A 578, 37): includes simplex test and Monte Carlo runs
      - includes other inversions: acoustic radius and age indicator (see
        Buldgen et al. 2015, A&A 574, 42), rotation gradiant and angular
        momentum gradient (see Reese 2015, A&A 578, 37)
      - supplementary interactive features for treating and modifying stellar
        models

2.1:  - can read other model formats
      - can calculate its own pulsation modes, or read them from a file
      - echelle diagram included
      - a page with adjustable physical constants has been added
      - improved management of text fields
      - receives directly the frequencies rather than frequency differences
      - includes various tests on rotational splittings, assuming a decreasing
        rotation profile, or one which obeys Rayleigh's stability criterion
        (see Reese, 2015, A&A 578, 37)

1.4:  - the structural kernels have been corrected
      - a different way of simulating noise has been implemented: the amplitude
        of the noise is now given by the product of the value on the noise slider
        and the individual error bars on the rotational splittings or frequency
        shifts.
      - colar has been added to various parts in order to improve legibility
      - a new tab dedicated to estimating the mean density using 3 different
        methods (SOLA, large frequency scaling, Kjeldsen's approach) has been added
      - rho_Gamma1 and Gamma1_rho (= c2_rho) kernels are now calculated, thereby
        allowing the user to do inversions on the structural pair (rho,Gamma1)
      - the calculation of the Eulerian perturbation to the gravity potential,
        which intervenes in the rho_c2 and rho_Gamma1 kernels, has been improved
        and now works at high l values
      - Gaussian-like functions which go to 0 at r=0 are now used as the target
        functions for the SOLA inversions (see Eq. 11 of Rabello-Soares et al.,
        1999)
      - the end points are excluded from the inversion grids (for both the
        sliding window and Chebyshev approaches)
      - horizontal error bars have been implemented for SOLA inversions
      - fortran binary AMDL files (which contains a stellar model) can now
        be read directly
      - fortran binary AMDE files (which contains eigenfunctions) can now
        be read directly
      - nfmode=1 AMDE and FAMDE files are now accepted
      - the format of the numbers in the Frequency Shifts and Rotational Splittings
        table has been improved
      - the documentation contains more details on the calculations which are
        carried out in InversionKit

1.3:  - can now parse doubles which use "D" or "d" instead of "E" (this
        slightly slows down reading files)
      - fixed a bug in navigation buttons for kernels and changed their layout
      - fixed a bug in the selection of models from the HELAS model tree
      - improved order of models in HELAS model tree
      - set DISPOSE_ON_CLOSE option in duplicate plot frames
      - gzipped format accepted for all files which are loaded
      - add a NaN check on the readDouble utility
      - better verification on validity of frequency shift data or rotational
        splitting data (checks for infinity and NaN added)
      - corrected an error in the calculation of EnlCoefficient when reading
        eigenfunctions from an FAMDE file

1.2:  - can now use full grid to calculate the different integrals (using a
        sliding window technique) which intervene in RLS and SOLA inversions.
        For the RLS inversions, this is supplemented by the use of b-splines
        which allow the use of a different grid for the output inverted
        functions.
      - corrected bug in which the inversion would sometimes update twice
        when changing the number of grid points in the inversion
      - averaging kernels are only calculated as needed so as to gain time when
        changing parameters such as Lambda, Delta of Theta

1.1:  - can read eigenfunction in FILOU format
      - can read .osc file from earlier versions of CESAM (i.e. before CESAM2k)
      - addition of separate field for IDL titles in plots so as to avoid
        difficulties with special characters
      - removed index array for rotation splittings and frequency shifts:
        - sort is very slightly slower
        - deleting rows is much faster
        - simpler code

1.0: First version to be distributed
!------------------------------------------------------------------------------
