ChangeLog

Changelog.

Fri Jan 29 21:52)19 PST 2010
- if libmeas is not installed, do not compile the measurement interface.

Mon Nov 23 20:46:53 PST 2009
- changed from xmgrace to qtiplot (available from yum)

Mon Oct 12 11:54:27 PDT 2009
- Instrument support dropped for Varian and Bruker ER200.
- Updated instrument support for Bruker ESP-300, which includes
  old Bruker (Aspect based) ENDOR/TRIPLE.
- major code restructuring
- migrated to libmeas

--- version 0.8 ---

Fri May 18 20:40:08 PDT 2007
- Fixed compilation problems on Linux. Changed xmgr -> xmgrace.

Mon Jun 14 16:20:17 EEST 2004
- disabled use of ARPACK - it is more or less useless here
- disabled use of MINPACK - it is more or less useless here
- Update Makefile compilation options
- minor changes to kinetics measurement interface
- It is now possible to specify location of XForms in the main Makefile
- The code compiles now on MacOS X as well (using Apple X11)
- Two separate Makefiles: Makefile.linux and Makefile.macosx
  (use ln -s Makefile.xxx Makefile to activate)
- increased MAXPEAKS
- primitive parallel support using auto parallelization with OpenMP

Fri Feb 27 19:42:13 EET 1998
- Updated Simulate -> Print parameters
- Fixed Norris exchange simulation
- Fixed Rate constant dialog (no longer core dumps)
- Change normalization constant in 1st order simulation
  (now normalized according to the number of transitions instead
   of peaks)
- Numerical stabilization hacks for Heinzer simulation - it still
  doesn't work properly.
- Document updates
- Example updates

Sat Feb 28 09:58:36 EET 1998
- Keyboard shortcuts for: cursors, x & y zoom/unzoom:
  Insert = cursor speed, Home = Zoom-x, End = Unzoom-x,
  Page Up = Zoom-y, Page Down = Unzoom-y, left&right arrows
  = left cursor, up&down arrows = right cursor. F1 - F12 select
  the corresponding spectrum pages (ie. 0 thru 11).
- Added iconify (+ icon - atomic nitrogen spectrum in Argon matrix ;-)
- Added support more nuclei - changed text size in popup menus.

Sun Mar  1 12:23:26 EET 1998
- Fixed DFL/IS variable settings when using Norris exchange model.

Wed Mar  4 10:37:46 EET 1998
- NEQUIV is now 90 in xemr/tune.h

Thu Mar  5 10:41:00 EET 1998
- For FFT LS the intensities are now multiplied by 1E8 in order to 
  avoid huge amplitudes.

Fri Mar  6 13:52:19 EET 1998
- Switched to RedHat 5.0 (+glibc) and egcs. Workaround for egcs bug (see
  xemr/main.c and search for a word bug)
- Asymmetric lw example
- Quick powder LS (actually the right way to do it when only one spectrum
  present)

--- 0.4.1 released ---

Sat Mar  7 19:23:23 EET 1998
- Added identifier for variables
- Monte Carlo optimization now also reminds about exceeded limits

Sun Mar  8 10:03:52 EET 1998
- Switched to the latest BLAS and LAPACK releases (version 9).

Sun Mar  8 18:58:38 EET 1998
- Initial version of ENDOR spectrum generation. Argh... this is
  dead wrong... rethink! [this is fixed now]

Fri Mar 20 18:03:43 EET 1998
- Added field calibration (y = kx+b -type). The default values are for our 
  esp-380...
- fixed conflict in lib/file.c of keyword stream (changed to strm).
  This caused xxx2epr converters to fail.

Sat Mar 21 19:08:40 EET 1998
- Fixed transition moment operator for nuclei. Nuclear spin start
  at nel not neccesarily at 1.
- Fixed ENDOR routine. Added ENDOR single & powder simulations
  to libepr and xemr interface.

Tue Mar 24 21:14:40 EET 1998
- Complains if target and exp spectrum pages are the same in fitting
- axial symmetry can be now performed by restricting sp_theta and
  sp_phi (earlier this gave incorrect results)

Fri Mar 27 10:41:37 EET 1998
- Quick powder removed (is now default)
- Added L = 3 angular momentum data
- Documentation
- if the experimental spectrum length was not 2^n and FFT lineshape
  method was used then the fitting performed the RMS calculation
  incorrectly.

Sun Mar 29 10:06:29 EEST 1998
- Updated convolutions (Lorentz, Gauss, other spectrum)
- Internal parameters are also saved to .sps file.
- Fixed sp_hamilton_divs bug which sometimes left the high field peaks out

Tue Mar 31 09:50:12 EEST 1998
- Variables that are not in range now report the indices correctly (+1)

Wed Apr  1 12:44:53 EEST 1998
- Heinzer exchange part rewritten. Switched to analytical solution.
  (now Heinzer and Norris give the exact same result ;-)

Fri Apr  3 17:09:23 EEST 1998
- Fixed MC limit error messaging
- Fixed intensities in Norris and Heinzer routines

Sat Apr  4 19:03:24 EEST 1998
- Numerical transition moment calculation now uses much less memory
  + added inlining to the intensity calculation routine
- Fixed '.' problem in file conversion.
- Updated esp2epr filter.
- Updated spectrum structure and file functions.
- Updated Edit->Parameters dialog (completed info for ENDOR and TRIPLE)
- Added export ASCII function
- Documentation updates (examples)

Sun Apr  5 16:20:12 EEST 1998
- Hydroquinone exchange simulation example

Mon Apr 13 17:37:13 EEST 1998
- Updates to hydroquinone example
- Fixed rpm spec file
- Examples get installed now too

Mon May 11 08:34:30 EEST 1998
- Fixed a typo in file conversion

Wed May 13 17:56:31 EEST 1998
- Added 1st order ENDOR simulation (isotropic only) which is useful
  (in combination with fitting) of getting HFCs with better accuracy
  from ENDOR spectra

Fri May 15 12:00:36 EEST 1998
- Fixed shrink (x-axis is now also updated)
- symmetry smooth cancel no longer generates core ...

Sat May 16 16:02:23 EEST 1998
- M-L non-linear least squares algorithm from the MINPACK library
  can now be used in minimization

Tue May 19 18:02:47 EEST 1998
- Fixed the varian EPR measurement interface
- Added copy current page function to Edit menu
- Changed the kinetic run behaviour

Sat May 23 13:07:21 EEST 1998
- Added support for manual scan mode for measurement
- Interface for Bruker ER-200D EPR, ENDOR and TRIPLE started

Wed Jun 10 14:18:56 EEST 1998
- 1st order ENDOR lw & amp no longer shows params for nuclei > ngr
- 1st order ENDOR lw & amp high and low fields were mixed.

Wed Jun 24 19:12:02 EEST 1998
- Print simulation params now also prints the exp. spectrum name

Mon Jul  6 21:36:52 EEST 1998
- Fixed G (or mT) -> MHz conversion in xemr/dialog.c
  it now accounts for g value.
- It now again compiles on sgi
- Cleaned up measurement routines

Wed Jul 22 18:57:39 EEST 1998
- EPR (rewritten), ENDOR, and TRIPLE measruement routines
  (See lib/systems/README for specification)
- Low level spectrometer routines rewritten
- Initial implementation for Bruker ER200D EPR/ENDOR/TRIPLE spectrometer

Mon Jul 27 21:29:13 EEST 1998
- Aspect slow serial bus hacking. The temperature controller works now.
  Still no docs on wavetek :-(
- Added temperature controller routines.
- Modified the main makefile for selection of spectrometer
- Added directory search for spectra

Tue Jul 28 21:57:58 EEST 1998
- Added clear simulation parameters function
- Added xemr.csh front end script for handling different spectrometers

Wed Jul 29 12:30:50 EEST 1998
- Added number of scans display

Fri Aug  7 09:46:12 EEST 1998
- ER200D EPR part works now - first EPR spectrum measured!
  (ER 033M lock sucks! It drops IEC625 command every now and then)
- Added set center field and scan width by cursors
- Params dialog now has center field and sweep width settings rather
  than begin and end.
- Modified variable dialog (mclimit in %, value is shown and is editable,
  active buttons for variables)

Wed Aug 26 08:37:55 EEST 1998
- Fixed 1st order ENDOR paramter printing
- ER200D debugging
- Fixed multiple electron coupling to nuclei (numerical SH)

Thu Aug 27 16:49:05 EEST 1998
- Fixed bugs in numerical SH transition moments
- baseline/spline: no longer crashes if no marks are selected.

Sun Aug 30 20:57:01 EEST 1998
- The "Knight example" works now...

Sun Sep 13 22:19:12 EEST 1998
- A bug in ENDOR measurement: no RF power is coming out (switches or
  attenuators) - no fix yet
- upgraded to egcs-1.1. Minor fixes to utils/*.c for egcs...

Tue Sep 22 18:15:52 EEST 1998
- Fixed a bug in ER200D endor interface. The documentation on attenuator
  settings is incorrect in the manual. Now ENDOR and TRIPLE general
  work ok. Still problems with TRIPLE special.

Wed Sep 23 20:40:39 EEST 1998
- Inital implementation of ESP-300 measurement interface. NOT TESTED YET
  and contains bugs for sure.

Thu Sep 24 20:09:20 EEST 1998
- The hyperfine editor should not show NaNs anymore
- Cancel delayed works now
- Transition moment diagrams

Sat Sep 26 15:40:07 EEST 1998
- added ENDOR induced EPR mode
- removed the manuals stuff from measure.c. If magnet or nmr control is
  not available then the program aborts.
- Break / Pause key turns the mw and nmr fields off. (ie. ctrl-e in aspect)
- spectrometer calibration setup + auto tune mode for max. modulation
  and signal phase.
- fixed modulation, signal phase, and receiver gain settings in esp300
- fixed (nmr) modulation in er200d

Tue Sep 29 16:14:46 EEST 1998
- The choice numbers were incorrect for powder spectrum generation
  (at least for 1st order epr) - fixed in xemr/main.c
- The transition moment diagram stuff had mixed up the numerical SH
  -- works now again...
- numerical SH for EPR: the transition moment matrix is symmetric,
  only half of it is now computed.

Thu Oct  1 14:03:13 EEST 1998
- ER200D ENDOR modulation scale fixed (again).
- ENDOR induced EPR mode works now.
- TRIPLE special mode works now.
- added igloo type mesh to powder integrator (now both regular and igloo
  can be used)
- The user interface now displays cpu time consumed in particular simulation

Fri Oct  2 16:07:48 EEST 1998
- Fixed ER200D EPR field settings (wrong lock offsets)
- A test implementation of even megahertz problem on endor/triple (ER200D)
  (the frequency appears to jump around somewhat when fine freq setting
   is zero)

Sat Oct  3 15:45:33 EEST 1998
- Simlation / fitting now also reports cpu times
- Fixed PO23+ example (FFT linshape bug removed)

Sun Oct  4 13:45:14 EEST 1998
- Changed to FFT routines (libfftw-2.0.1). Unfortunately this changed
  the intensity scaling, so amplitude settings need to be updated.
  This change also resolves the patent issues that were involved with
  the previous FFT routines. Also the power of two requirement is gone.
- Spectrum differentation now works in Fourier space when generating
  derivative spectra
- Switched to linux-gpib-2.04 which supports variety of gpib cards.
  The driver appears to much better than the one I used previously!

Mon Oct  5 10:21:39 EEST 1998
- SGI support dropped. Too many things are broken, different, etc.
  I hope SGIs run Linux soon so that things start working again ;-)

Thu Oct  8 22:12:55 EEST 1998
- Added fictious spin representation to numerical spin hamiltonian
- Fixed FFT lineshape generation bug (with certain width/resolution
  combinations strange spectra were obtained)

Fri Oct  9 14:27:36 EEST 1998
- Fixed kinetic operation mode of EPR and ENDOR measurement
  (xbeg = xend, samples amount of steps)
- Fixed DE(J) setting in S.D.S for fictious spin case

Sat Oct 10 18:36:06 EEST 1998
- Documentation updates
- Due to public demands the font size is now bigger...
- Template spectrum concept (template spectra for EPR, ENDOR, TRIPLE, etc)
- Breit-Rabi graphs now respect fictious spin selection
- Higher optimization flags for gcc

Fri Oct 23 08:47:12 EEST 1998
- The new compilation options caused the conversion utilities to fail
  (fixed)
- More problems fixed in esp2epr conversion utility
- Varian measurement interface fixes (it works again...)

Thu Oct 29 08:42:33 EET 1998
- Added a new linshape model. Gaussian distributed Lorentzians.
  Both can have their own linewidths (ie. llw = 0 -> pure gaussian).

Wed Nov  4 10:20:40 EET 1998
- fixed epr kinetic mode
- TRIPLE pumping freq. problems fixed for ER200D

Thu Nov 12 10:01:28 EET 1998
- fixed x and xy conversions

Thu Dec 03 17:50:22 EET 1998
- Fixed a typo in parameter print out (the exp. file spectrum name is now
  printed correctly)
- Added microwave frequency to parameter print out

Tue Dec  8 09:48:57 EET 1998
- Exp. resolution can now be upto 64K. + maximum sample size
  enhanced to 64 K.

Tue Dec 22 10:42:15 EET 1998
- Added a simple programming example using LIBEPR (examples/programming)
  and one example using it.

Fri Jan  8 14:41:12 EET 1999
- Fixed convolution routines (hopefully everything is ok now..)
- Added cross/auto correlation 

Thu Jan  14 10:45:14 EET 1999
- Numerical solution of SH now recognized transitions which have
  reversed energy dependence on field strength. (For example nuclear
  transitions in atomic Cu)

Fri Jan  15 15:32:58 EET 1999
- Correlation and convolution axis fixes.
- Windowing in Fourier space
- Cu atom example + BR graphs

Thu Jan  21 19:20:03 EET 1999
- Major modifications to measurement interface
- ESP300 (CW) interface appears to work now (still problems with DMA)

-- 0.7 released --

Thu Feb 11 14:14:10 EET 1999
- Cepstral analysis and power spectrum added.
- Fourier space axis is now correct (for left/real part only).

Sun Feb 14 21:20:56 EET 1999
- Added more windowin functions
- there are still numerical stability problems with cepstral analysis.
  However, the matlab gives exactly same result so libepr works ok.
  The number of spectral points appears to give better results.

Tue Feb 16 20:16:14 EET 1999
- changes to windowing functions
- fixed signal phase correction (given in degrees - cos works in radians)

Thu Feb 18 18:08:49 EET 1999
- Major speed up for er200d endor scanning.
  (affects other systems to small extent as well)

Fri Feb 19 18:08:04 EET 1999
- Changes to varian measurement setup (the intensity is now read directly
  from behind the signal channel - should have done this in the first place...)
- The internal gain register in varian (DAS8) can now be adjusted
  from the program.
- Corrected typos in varian interfacing documentation

Tue Feb 23 21:43:41 EET 1999
- more Varian fixes

Fri Feb 26 09:09:46 EET 1999
- added ang. mom. related stuff (B.L, aS.L and oct/tetr crystal fields 
  + distortions)
- ang. mom. is now calculated instead of using arrays
- L+ and L- operators added (and I identity)
- fixed a bug in numerical SH: if the same spin was operated twice with
  an operator that changed the spin state then this was not accounted for!
  (this error produced incorrect results, for example for: I.P.I and S.D.S
  systems)
- added so, cra, and crb to variables
- load/save now keeps the selected directory

Sat Feb 27 12:35:14 EET 1999
- some of the baseline stuff rewritten
- INDEX macro was just the opposite as it should have been
  (it now gives the correct array ordering required by fortran convention)
  This may reverse indexing for simulation sets saved before this.
  (for example, g_ij will now be g_ji... - fortunately these are
  mostly symmetric). This appears to cure the Heinzer exchange simulation
  routines.
- Oh well. The numerical SH matrix construction was somewhat rewritten.
  Maybe I got it right this time ;-)

Mon Mar  1 11:22:47 EET 1999
- spin-orbit and crystal field params are now in cm-1.
- fixed parameter printing problems
- numer. sh. based epr spectrum simulation routine rewritten so that
  it consumes much less memory. It might be faster too (at by factor of
  two for "the Knight example").
- fixed FFT xbeg and xend
- esp2epr converter had some inconsistencies...

Tue Mar  2 17:38:37 EET 1999
- Active space definition: one can now tell which energy levels are considered
  when calculating the spectrum. However, the eigen solver still produces
  all the levels - they are just excluded from the transition search.

Wed Mar  3 08:27:24 EET 1999
- preloader for powder / numerical spin hamilton simulations (speed up)
- for tetragonal distortion one can now specify x, y and z
  distortions separately
- BR graps: GHz or cm-1 energy units
- the fictious spin now also affects orbital ang. mom. - canceled for
  now. The crystal field stuff becomes very tiresom to program...
- some additions to the documentation (numer. spin Hamiltonian)
- fixed couple of "core dump" problems when operations were issued
  on an empty page

Fri Mar  5 09:30:42 EET 1999
- documentation updates on numer. spin hamiltonian
- new example: quenched boron atom
- everything can now also be specified in cm-1
- variables dialog now updates parameter headings during editing
- ENDOR B was sometimes stored as blank which confused the simulation parameter
  file reading... - fixed

--- 0.7-2 released ---

Sun Mar  7 22:21:01 EET 1999
- a factor of half speed up for numer. spin Hamiltonian simulation
- variables dialog is now smaller

Thu Mar 18 18:34:13 EET 1999
- Added Boltzmann populations for energy levels which affect the
  transition intensities.

Tue Mar 23 10:10:31 EET 1999
- Large arrays are now allocated only when needed in hamilton.c
  (via malloc()). This solves start up problems on small memory machines
  (ie. core dumping when xemr was started)

Thu Mar 25 09:13:49 EET 1999
- Better support for axial symmetry powder spectra
  (nphi = 1)
- B2 (triplet Sigma) example

Fri Apr  2 17:07:06 EEST 1999
- Updated html examples section
- Documentation updates

Tue Apr 13 07:57:34 EEST 1999
- Changed numer. diagonalization routine to a faster one
- Xemr now also works with Intel (pro and II) optimized BLAS library
  ("ghenry version") - b2 triplet from 54 s to 44 s...
- fixed a bug in powder spectrum generator: igloo and one dim modes were
  sometimes set to incorrect values.

Tue Apr 20 10:13:46 EEST 1999
- fixed a bug in preloader - transition moment operator cannot be preloaded!
- fixes to ENDOR simulation (numer. spin hamiltoninan)
- integration intervals can now be specified
- clean up powder integrator

Sat Apr 24 21:21:37 EEST 1999
- Fixed naphthoquinone endor simulation example - the difference in line
  intensities was due to a bug...
- minor changes to powder integrator

Wed May 19 17:19:02 EEST 1999
- Switched to portland c-compiler: b2 triplet is now at 39 s ;-)
  (the code still compiles with gcc - see the main makefile)

Sat Jun  5 10:41:33 EEST 1999
- Obtained portland group patch for redhat 6.0 - things appear to work now!

Sat Jun 12 07:28:56 EEST 1999
- One can now visualize: transition moments or transition eigen vector
  difference (ie. to characterize the transition). It calculates the
  differences between squares of basis coefficients between given two
  states.
- Added "measurement in progress..." indicator
- minor cleanups for xemr/main.c
- changed the function of arrows: up & down = scale, left & right = page
  prev & next
- fixed parameter setup problem (dlevel & co. were only updated after
  define simulation dialog)
- fitting can now also be performed in the freq. space (not tested)
- Implemented Hooke-Jeeves optimization method


Thu Jun 17 22:55:16 EEST 1999
- Removed % stuff from mc_limit - it was rather unpractical!
  The mc_limit is now given in difference value.
- minor speed up to numer SH in misc.c (preload).
- more speed up for transition moment calculations:
  The b2 (1500 G) example went from 30 s -> 13 s!
  For some ENDOR simulations the run time dropped also to one third.

Sun Jun 20 16:29:37 EEST 1999
- Changed the input scheme for g, A, D, and P tensors. THe user should
  now input tensor principal values, and the rotation matrix that 
  provides the relation between the fixed axis system.
  NOTE: the aliases still use the old notation!
- fixed powder integrator bug when regular mesh was used.
- added a check for illegal angle values
  
Thu Jun 24 22:02:33 EEST 1999
- rewriting of powder integration routines. Many different integration
  algorithms were added.
- the powder integrator params are now saved to sps files.
- Added manual RMS spectrum evaluation for spectra on two different pages.
- Transition to comedi library started (both das8 and 1600 are now supported).

Mon Jun 28 20:50:31 EEST 1999
- Added linear and cubic spline transition interpolation (transition position
  and transition moment). This is similar that sophe uses (except that
  it does cubic spline and linear interpolation at the same time)

Thu Jul  1 17:59:45 EEST 1999
- Implemented eiven vector following ("level follow"): The eigen ordering 
  problems should be gone now. However if the spectral range is large then the
  eigenvectors may have weak correlations to the starting states ->
  orders may be incorrect.
- Added peak "padding" in hamilton.c: it now returns the same number of
  peaks so that the peak positions in peak arrays stays at the same place.
- BR-gaph now respects level follow. Watch out for this when doing
  trmom or transition analysis (these do not respect level following!)

Sun Jul  4 12:04:08 EEST 1999
- the level following works now better. The previous point is used in
  calculating the correlation matrix.

Fri Oct  1 08:37:34 EEST 1999
- fixed PEEK_ORDER_HACK bug which sometimes missed peaks.
- fixed a bug in crystal field code - it was sometimes operating on
  wrong spin!!!
- more changes to orbital angular momentum: now it can handle many
  electrons (the same spin-own-orbit constant for all electrons - 
  although this can be generalized easily so that each electron
  could have its' own spin-own-orbit constant).
- removed value fields in variable dialog: one could accidentally
  change the values with this!
- equivalence editor now displays the index texts "on the fly"

Fri Oct  9 08:37:34 EEST 1999
- fixes to variables and equivalences dialogs
- updated device drivers to kernel 2.2

Tue Oct 12 20:14:56 EEST 1999
- fixed a crystal field bug (operators negative multipliers were silently
  ignored!)
- crystal field document updates

Tue Oct 22 20:14:56 EEST 1999
- linear interp. powder integrator typo fix
- minor clean ups in hamilton.c and powder.c
- libf2c is no longer needed if portland cc is used (argh... arpack needed it)
- fixed a preloader bug in misc.c - the number of field indep. hamiltonian 
  was sometimes wrong.
- adding empty page as overlay caused core dump
- added cnvall script which converts all ESP type spectra to xemr format
  present in the current directory

Sat Oct 30 10:29:34 EEST 1999
- added magnetic field direction dependend linewidths (for EPR)

Fri Jan 28 10:53:48 EET 2000
- pressing numbers in xemr main window no longer crashes the program

Fri Feb 18 07:59:19 EET 2000
- larger peak arrays
- minor clean up in sp_hamilton_mode (modes now defined in common.h)

Wed Feb 23 12:27:13 EET 2000
- fixed hfc printing bugs

Tue Mar 28 08:03:59 EEST 2000
- A new spectrum add mode