UNEX 1.3.0 beta.


1. Introduction.
2. Installation.
3. Capabilities.
4. Usage.
5. Last changes.
6. Feedback.


1. Introduction.
----------------
UNEX - program for molecules study. It can all (or almost all) that needed for 
a conventional GED research, from initial processing of experimental materials 
to structural analysis. Also with UNEX one can investigate molecular geometry
using rotational constants. UNEX does not based on any version of original 
Kced or other programs.

The primary aims of developing of the program:
1. Experimental study of the molecular structure.
2. Development and automation of all stages of the GED method.
3. Expend capabilities of the GED method due to the development of the joined 
   structural analysis methods and usage of the information from other physical
   and chemical methods.
4. To provide the capability to carry out "very accurate" researches due to 
   extended capabilities and flexibility of the program interface.
5. Development of the method of molecular geometry study by rotational 
   constants obtained from MW spectroscopy and high resolution IR and Raman 
   spectroscopies.
6. Multiplatform support.
7. Compatibility with other programs.



2. Installation.
----------------
UNEX works in a command line. It doesn't require any special operations for 
installation. If you have got it as the archive file, simply unpack this 
archive to any comfortable place on the hard disk.


3. Capabilities.
----------------
Some of current UNEX features and capabilities:
1. Study of a molecular geometry by rotational constants obtained from 
   MW spectroscopy and high resolution IR and Raman spectroscopies.
2. Isotopomers of the molecule under study can have different geometry, so
   it is possible to work behind Born-Oppenheimer approximation.
3. Study of the molecular structure by GED data. Static model of gas electron 
   diffraction theory. 
4. Dynamic GED model. Parametrical form of the potential function. Usage of the 
   relaxation of the geometrical parameters, mean amplitudes and vibrational 
   corrections.
5. Modeling any mixtures of molecules with static and/or dynamic GED models.
   Modeling as mixtures of conformers with identical topology, and mixtures of 
   molecules with various topology is possible.
6. There is no need to write "spesr" subroutine. Geometry is set by 
   z-matrices. UNEX z-matrix  format is the same as Gaussian z-matrix format.
7. Support of dummy atoms. With the dummy atoms and the approach of z-matrices 
   it is very simple to set any symmetry for the molecule.
8. Powerful methods of minimization and criteria for minimum finding. It is 
   possible to do Robust-minimization (iteratively reweighted least-squares 
   method).
9. Variation of the geometrical parameters, mean amplitudes, scale's, 
   concentration of molecules, parameters of the potential function. A group 
   variation.
10.Automatic calculation of the errors for dependent parameters.
11.Automatic division the mean amplitudes into the groups.
12.Multivariate scanning of the GED R-factor in any parameters. The amount of 
   dimensions is not limited. The amount of points on each parameter is not 
   limited. Scanning by groups.
13.Smoothing model-dependent background by cubic splines or polynoms.
14.Smoothing model-dependent background with sector function usage.
15.Automatic calculation of g-functions and atomic scattering inside the 
   program. It is possible to input g-functions from outside the program.
16.Unlimited number of points of GED intensity. The non-equal step on 
   intensity is allowable. Unlimited number of nozzle-to-plate distances. 
   Unlimited number of intensity curves from each nozzle-to-plate distance.
17.First-order local sensitivity analysis.
18.Work with GED gas standards. Estimation of the electron wavelength and 
   sector function. Support for benzene and CCl4 as a gas standards.
19.Work with image files. Support for uncompressed 8/16 bit TIFF format files.
20.Calibration the scanner. It is possible to do the calibration "on the fly".
21.Initial pocessing of a GED experimental data. Extracting intensity curves 
   from scanned diffraction patterns.
22.Methods for the scanning quality control.
23.Flexible and convenient input format.
24.Efficient usage of the SMP (multiprocessor) systems.
25.Versions of program for OS/2 Warp, eComStation, FreeBSD, Linux, Windows.



4. Usage.
----------------
In command line:
Linux/FreeBSD: ./unex <your_work_file>
Win/OS/2: unex.exe <your_work_file>

There is no full English documentation at present time.
For the help about usage of the program see example/test files.



5. Last changes
----------------
[1.3.0 beta]
1. The output information on command PRINT=COMPINFO about accuracy of 
   calculations has been corrected.
2. The errors hiperellipsoid now printed after command MINIMIZE. 
3. Condition number now printed after each MINIMIZE job.
4. New important possibility to do a Robust-minimization has been added. 
   See command ROBUSTM.
5. Leaking of memory was corrected at work of command IMAGE=INTSCAN.
6. Procedure of extraction of GED intensity curves from diffraction images 
   especially for pictures with narrow rings (for example from polycrystals) 
   and for distant area of usual pictures of gas samples has been improved. 
   The quantity of iterations accepted by default has been increased up to 20.
7. Key IntExclBadPts is not necessary now and removed from field BASE.
8. Key WriteWeightsImg has been added to the BASE field for management of 
   writing of a file with diffraction image points weights.
9. Now it is possible to change "on the fly" UNEX settings by means of 
   repeated typing of command BASE in any place of the working file.
10.Atomic weights of the basic isotopes of elements has been refreshed using 
   NIST data.
11.There is essentially new operating mode of command SEARCH, i.e. search of a
   global minimum of functional - method of Monte-Carlo. Details see in the 
   documentation.
12.The possibility of restriction the time of command SEARCH work has been added. 
   For this purpose the key SrchTime has been added to field BASE. Pay attention 
   to its value by default.
13.Calculation of concentration of molecules (in case of mixture model) in 
   command SEARCH has been corrected.
14.Now after work of command SEARCH the parameters of model automatically 
   accept the best values found by command SEARCH.
15.The printing of some geometrical parameters after command PRINT=ALLGEOM and 
   commands PRINT=BOND, PRINT=ANGLE, PRINT=TORSION has been corrected.
16.The possibility to print out a matrix of correlations between estimated 
   points of GED intensity by command IMAGE=INTSCAN has been added. See key 
   ImgPrintIntCorrs in field BASE.
17.More convenient printing of a matrix of correlations between estimated 
   parameters by command MINIMIZE or ROBUSTM.
18.The algorithm of work of command STANDART has been improved.
19.Now errors of points of sector function estimated by command STANDART are 
   printed.
20.By means of key StdPrintSecCorrs in field BASE it is possible to print out 
   a matrix of correlations between estimated points of sector function by 
   command STANDART.
21.Now by means of command AUTOGROUP=AMPLITUDES only those amplitudes grouped 
   atoms of which do not contain hydrogen.
22.Crashing of the program during printing of theoretical and experimental 
   sM(s) curves for standards has been corrected.


[1.2.0 beta]
1. Added possibility to study molecular geometry using experimental rotational
   constants from any vibrational level.
2. Now molecule can have a set of isotopomers.
3. Added possibility to input rotational constants, their errors and 
   corresponding corrections in molecular fields.
4. Added possibility to manual input of atomic masses in z-matrix field.
5. Added key (PrintMainInertXYZ) to BASE field that manage printing cartesian
   coordinates of all atoms to log.
6. And again changed syntax of the functional scanning command. Now, if used only
   GED data for scanning, one should use the command like this: 
   SEARCH=SCAN,GEDINT,<scan>,</scan>,1-1
7. Improved strategy of combined method of minimization (LSQ + golden section).
8. Added possibility to choose minimization method with the MinMethod key.
9. Fixed not critical bug: if molecules with GED dynamic model are
   present than it was not possible to input intensity curves before all 
   potential curves will be typed.
10.Fixed scan of the R-factors by parameters of molecules with GED dynamic
   model.
11.Fixed memory leaks when orthogonal parameters are minimized.
12.To the table of estimated parameters, that printed after minimization, added
   previous values (before minimization) of all parameters.
13.More optimal usage of the RAM memory. Now UNEX use less memory, especially 
   on jobs with molecules with GED dynamic model.
14.Fixed bug: for command SEARCH=SCAN UNEX didn't convert input values of angle limits to the radians.
15.Fixed not critical bug: after command SEARCH=SCAN UNEX didn't convert output values of angle parameters to the degrees.
16.Significantly improved usage of sector function in SBGL command.
17.Documentation (for the present in Russian) was improved and several test
   graphs added.


[1.1.0 beta]
1. Big part of code was changed.
2. Dynamic model behavior was changed. Now concept of "dynamic model" is 
   related to the specific molecule instead of whole system (like in prev. 
   versions). As a result, it is now possible to make models of a various 
   mixtures of dynamic and static molecules with different or identical 
   topology (see tests). For the same reason the way of input of the 
   information concerning of dynamic model has a little changed (see the 
   documentation, examples, tests).
3. In connection with dynamic model input change for the commands 
   PRINT=POTENTIAL and PLOT=POTENTIAL names of molecules are required.
4. Syntax of command MINIMIZE was a little bit changed (the type of 
   minimization is required now). Now, for minimization on GED sM(s) it is 
   necessary to use MINIMIZE=GEDINT, 1-1...
5. SMP-support is added. Now on multiprocessor computers minimization goes 
   much more quickly.
6. The key CpuNum for the compulsory task of quantity of processors was added 
   to the BASE field.
7. In connection with a works above new ways of R-factors minima searching 
   (except for SCAN), SCAN command syntax was changed. Now general command 
   SEARCH with corresponding parameters (see the documentation, examples, 
   tests) is used.
8. Now each Scale-factor of all GED intensity curves for a variation should 
   have unique number of group. Though by default the program itself sets 
   different numbers of groups to Scale-factors, it is recommended to enter 
   them manually, otherwise automatic numbers can depend on presence of other 
   groups of parameters.
9. Added parameter ImgPrintIntR which supervises the intensity received on 
   command IMAGE=INTSCAN depending on GED parameter s or from distance up to 
   the center of a photographic plate will be printed.
10.Now for command PRINT=FURTERMS name of molecule is required.
11.Now there is also Linux version of the program.


6. Feedback.
----------------
Any comments and suggestions are welcome.
Program homepage: http://molstruct.chemport.ru/mykced_en.html

Yu. V. Vishnevskiy
Moscow State University
Department of Chemistry
E-mail: y_v_vishnevsky@rambler.ru
