                    List of corrections and modifications in FDMNES

---------------------------------------------------------------------------------------

September 1st 2016
Modifications in radius and SCF for when H atoms.
Some technical modifications.

July 15th 2016
Bug in the reading of occupancy rate in cif files fixed.

July 6th 2016
Bug, introduced in June versions in SCF excited option, fixed.

June 23rd 2016
Bug, introduced in June versions on the reading of cif files, fixed.

June 23rd 2016
Zero energy at K edge set corresponding to the exctited absorbing atom when not SCF calculation (gives a slight shift).
Spin-orbit E1E1 transition channel included. 

June 9th 2016
Introduction of Film_zero keyword

June 2nd 2016
Zero energy is nos set at the Fermi energy. All spectra are thus shifted versus
previous version. To get sma origin as before use keyword "Old_zero".
This new version contains also the extension to 2D film diffraction (see manual).

March 1st 2016
Technical modifications. Reading of Cif file without atom_site_occupancy now possible. 

February 9rd 2016
Parameter p_self_max, for SCF added.

February 9rd 2016
Core level is now versus Fermi level (--> better relative shift when more than 1 non equivalent absorber).
Slight change in the parameters for SCF.

February 3rd 2016
Core level energy better calculated (effect only with SCF).

February 1st 2016
Self-consistent option slightly improved.
Default options are now 100 iterations max and 0.1 eV step for convergence.

January 8th 2016
Automatic reading of the cif files.

January 6th 2016
Technical modifications.

December 15th 2015
E1M1 term now analyzed with spgerical tensor (one bug corrected also in its evaluation for mahgnetic system).
Polarization defined in the indata not anymore in reduced unit mesh unit, but in the normalized unit cell
(one still has to take into account the angles between axis but not the difference in modulus of the unit cell parameters). 

October 28th 2015
Bug fixed for simulation FDM + Spinorbite with some point group (2/m, 4/m...)
Screening keyword does not need anymore the associated use of Atom keyword.

October 6th 2015
Edge energies for Z > 93 now complete.

September 1st 2015
Bug fixed with "sphere_all" keyword.
Double counting of nbr of atom fixed for Bormann.

July 28th 2015
Bug fixed when several adimp or radius

July 26th 2015
Bug fixed in MPI.

July 24th 2015
NRIXS simulation are included.

July 20rd 2015
One (big) bug fixed in Epsii calculation at L23 edges. Introduced in previous version. 

July 3rd 2015
Speeding up of the FDM solution by Sergey Guda.
TDDFT slightly improved.
Numerical instability sometimes occuring with Spinorbite + Hubbard fixed.

June 5th 2015
Instabilities in TDDFT fixed.

April 27 2015
Two variables were not initilized what could make problems on linux systems.

April 23 2015
First bug of April 20 was not completely fixed. It is now the case (I hope).

April 21 2015
Minor bug fixed when using SCF keyword (sometimes the code stopped at the begining of the job).

April 20 2015
- 2 minor bugs corrected :
   1) When using extract keyword on files longueur than 1 000 000 lines (no more limit).
   2) When using weight to get an average under Polarise keyword, when their sum was zero.

April 15 2015
- Introduction of the MUMPS solver to solve the FDM matrix, in collaboration with Guda, Soldatov et al, Rostov group.
  It is still possible to work like before if compiling with mat_solve_gaussian.f routine in place of
  mat_solve_MUMPS.f
  Simulation are up to 40 times faster in FDM with MUMS. Results are the same.
- Occupancy rate is now taking into account for atoms at the same site.
  Scattering matrix is evaluated taking choosing the surrounding atoms corresponding with maximum rate
  when more than one are at the same position.
  Sum over xanes and scattering amplitudes corresponds to the occupancy rate.     
- Fit on occupancy rates is now possible

February 13th 2015
- Bug found (with rare effect) when using spinorbit without magnetism.
- In order to save memory space, the way of calculation of the variation with energy
 of the interstitial potential is changed.
 Give very tiny modification of the calculated spectra.
- Modification of the default core-hole width. Follow now precisely Krauze and Oliver.
The change is slight for K edge but bigger for L1, L2 and L3 edges.
- Fundamental constant values (hbar, bohr, Rydb) given with last precision. 

January 5th 2015
- Other magnetic ponctual groups were not found whan rotation was needed. Now it is ok.
- Bug recently incorporeted for "Crystal_t" keyword fixed.
In both cases the code stopped.
- Specific bug occuring with gfortran compiler fixed.
It needed the suppression of the "Length_word" keyword.

---------------------------------------------------------------------------------------

December 22nd 2014
Bug fixed for the symmetry in magnetic system with symmetry 4'/mmm.

December 11th 2014
A bug fixed in case of antiferro-magnetic compounds (very special case).

November 6th 2014
Calculation of atomic scattering amplitudes (f', f") improved for the L23 edges.
Extrapolation included for their calculation, and concatanation of the corresponding files
before integration (well, it is rather technical... in other words it is easier for the user). 

October 29th 2014 
A bug fixed (I hope) when spin-orbit + Hubbard. Gave sometimes numerical instabilities.

October 24th 2014
A bug fixed in the search of equivalent atoms in magnetic systems for some symmetries.

September 19th 2014
A bug fixed in resonant scattering for antiferromagnetic systems.

September 5th 2014
A bug found in the magnetic non resonant scattering (for 5 years, gave zero...)

September 3rd 2014
A bug found in the linux compiler version (gave NaN...)

July 18th 2014
Bug (recent) found with "Hubbard + spinorbite".
Keyword "Density_comp" introduced.
Keyword "Doping" introduced.

April 30th 2014
Bug found on E_cut energy evaluation when excited valence occupation used (very rarely used).

April 16th 2014
Bug found in the automatic shift when summing non equivalent atoms. (old bug...)
Bug found on keyword "EFermi" on linux.

April 7th 2014
Keyword Folder_dat included to choose the directory for spacegroup.txt and xsect.dat files.
Routine are now called **.f90 and are in free format.

March 28th 2014
Stokes parameters included for resonant diffraction.
Right and left definitions reversed (!).
A bug corrected for forward scattering calculation, that is (h,k,l) = (0,0,0).
Keyword "Folder_dat" included.

-------------------------------------------------------------------------------------

December 12th 2013
A bug found in the "CXY" for data from Wien2k.

December 10th 2013
Bulk center case now included for potential coming from Wien2k.

November 15th 2013
Bug found in case of Perdew-Zang exchange-correlation potential.

October 29th 2013
Bug found occuring sometimes for self-consistent calculations in parallel mode.

October 1st 2013
Bug found for ponctual group with complex representation (as C3, C3h, C4, C4h) with FDM.
Bug present from May 2013.

September 2013
Formula corrected for the non resonant magnetic amplitude (error in the Blume formula)
Bug corrected in the birefringence correction.
Bug (introduced in May) for m3m symmetry when working with FDM (gave 0!)
Magnetic dipole transition in E1M1 and M1M1 corrected.
Reading of pdb file included (Crystal_p)
Reading of orientation matrix and DAFS reflexions from experimental files (Dafs_exp)
Keywords Z_absorber and Atom_conf included.
There is a new file to compile (optic.f)

May 27th 2013
Magnetic dipole transition in E1M1 and M1M1 corrected.

May 18th 2013
Speedness of the code improved (vs 2013 version).

May 3rd 2013
E3E3 term included.

April 26th 2013
One bug corrected.

April 17th 2013
Bug corrected when Hubbard + Spinorbite. Density of state calculation improved.
Minor bugs corrected.
Note that there is no more the hubbard.f file to compile. (Hubbard staff is included in 
the other routines.

-----------------------------------------------------------------------------------------

November 2012
One bug corrected in case of magnetic calculation using the One_run process.
Modification of the intersticial potential in order to make the average 
in area at distance less than 2.5 Angstroem from any atom (useful in some cases for molecules).

July 9th 2012
Optimization of the SCF (better convergence).
One bug corrected for magnetic RXD calculation (rare occurence).

June 7th 2012
SCF calculated by default with a fixed magnetic polarization.
Hubbard energy correction included in the total energy.

May 16th 2012
Bug corrected when Linux + TDDFT (gave an aleatory error message).

April 24th 2012
Bug corrected when Hubbard + FDM.

April 8th 2012
Bug corrected in option "One_run".

February 21th 2012
One output format correction.

February 15th 2012
A bug corrected for calculation TDDFT + self absorption.

February 2nd 2012
A bug corrected with spin orbit and imaginary energy.

January 31st 2012
Birefringence effect included
Self absorption for RXD improved
Interstitial potential calculated in a better way.
One_run option included.

-----------------------------------------------------------------------------------------

August 12th 2011
Some bugs fixed in the "scan" part for RXD.

May 18th 2011
Correction bug on reading of keyword "Self-abs".

April 30th 2011
Minor bugs corrected. Default edge width modified for Mn, Sc, V, Ti L23.

March 18th 2011
Minor bugs corrected. Detectable only with specific compilers.

March 2011
Minor bugs corrected.
TDDFT part is by default with spinorbite.
In LAPW, atomic radius is by default the Norman one.

February 2011
Relative shift of initial states calculated, including magnetic shift.
By default, L23 and M45 are now calculated without core-hole.
By default, broadenings at L2 and L3 have different widths.
Default atomic radius repaced by the Norman radius when the default is too big or too small. 
Hubbard correction included.
TDDFT calculation included with local Kernel.

-----------------------------------------------------------------------------------------

August 4th 2010
Correction of a bug for the Wien2k interface when structure is not cubic or tetragonal.
For Wien2k interface, no more need in general to modify the number of radius point.
Calculation of energy of core state now relativistic.
Definition of starting energy for SCF slightly improved.

July 6th 2010
Some minor bugs corrected.

June 1st 2010
- One bug coorected and existing from version March 2009: the magnetic E1E2 was not calculated!
- E1M1 and M1M1 component now calculated.

May 11th 2010
- Some bugs corrected in Hubbard and symmetrization in multiple scattering mode.

March 22nd 2010
- Symmetrization in the multiple scattering mode in the spin-orbit case (thus calculations faster).
- Hubbard correction implemented
- TDDFT calculation implemented
- Some Variables not initialized, now initialized (gave error with specific compilers).

February 1st 2010
Two minor bugs corrected.

-----------------------------------------------------------------------------------------

May 2009
One bug corrected in case of punctual symmetry S6.

July 2009 2th
Some bugs corrected. Convolution slightly modified for XANES.

December 2009
Bug corrected in the magnetic case with spin-orbit for trigonal symmetry.
Keyword "scan_conv" created to write the output file name of the scan convoluted file.
Crossing spin term is now taking into account. This changes very slightly calculations with spin-orbit
at L23 and M45 edges. (No differences at K edges). 

December 31st 2009
Bug corrected in Radial Matrix when two edges for E0 and E2 term.

---------------------------------------------------------------------------------------

August 2008 - March 2009
- Self-consistent calculation included.
- Some modifications for the evaluation of the muffin-tin radius (in some case values were wrong).
- The sign of the E1E2 component is changed for DAFS. 
- There is no more the nesparam.inc file. There is is a new file : scf.f.

June and July 2008
- Some other bugs corrected

May 2008
- Some bugs corrected

April 2008
- sigma polarization defined reverse.
- azimuth defined anti-clock wise for the rotation of the crystal around the Q axis.
- Atomic potential calculated more precisely.


March 2008
- Symmetrization of the multiple scattering calculation (but with spin-orbit).
- Calculation of negative energy in multiple scattering.
- Photoemission calculation.
- Improvement of the MPI.

---------------------------------------------------------------------------------------

March-June 2007
- Many modification the program in order to improve the speed when working in FDM model.
  Real only with the existance of 3 fold axis and sometimes with 4 fold axis. 
- Non parallel spin axis are now possible
- With MPI, results are given continuously during the job and not just at the end.
- Different bugs were corrected.

14 February 2007
- In routine Symsite, declaration of base_orth as logical forgotten previously. Gave problem
in compilation with some compiler

6 February 2007
- Bug found and corrected for the automatic screening in magnetic configuration
when the non excited electronic configuration is such that there is less than
 half electron free for one of the spin for the orbital where comes the screening electron
(very rare situation).
- Establishment of the groups of equivalent atoms made less memory consuming. Previously
when the number of non equivalent atoms was to high one got stack overflow.

19 January 2007
- Bug found and corrected when symmetry plane along y=z (or x=z) and no symmetry
 plane for y=0 and z=0 (or x=0 and z=0). (Occured just in FDM).

15 January 2007
- Bug found and corrected when 2 fold axis along x (or y) and symmetry plane
 along y=z (or x=z). (Occured just in FDM).

10 January 2007
- Anormal memory increase under MPI on some compiler corrected (by R. Wilcke)
- Bad initialisation of naltm parameter corrected (previously could stop the program).

8 January 2007
- For 1 year, for the 4f elements, the screening was by default on the 5d, 
  it is now again by default on the 4f.
--> This can modify the results comparing with previous version
- Better output in bav file to get the atomic electronic configurations.

