WINDOWS Utility Programs for use with PARAM

A set of utility programs are provided for preparing data-sets for PARAM and for analyzing the results. Because they work within individual folders these programs should be run at the command prompt level, i.e., using "cmd.exe".  All the programs and utilities can be found in the folder "%PARAM_Source%Source_code\Bin_and_cmd" Most programs are command scripts, i.e., they end in ".cmd", the rest are executables, i.e., they end in ".exe" but for simplicity both types will be referred to as "programs."

PARAM is exported with all the files needed to carry out parameter optimization work, and a folder structure, named "PARAM_for export" that allows the utilities to be used easily.  A convenient way of regarding this folder structure is as a rigidly-structured quarantined domain that has only minimal interaction with the rest of the computer.  One advantage of this rigid structure is that the utilities can be used easily because the relative positions of the various folders is hard-wired. Another is that the entire structure can be easily copied to make a duplicate which can then be used in parameterizing either an existing semiempirical method or a new method.    

Environment programs

SETUPDOS (Sets up environment for a project involving PARAM)

This sets the environment that allows

Within SETUPDOS, the PATH is extended to include the folder  "%PARAM_Source%Source_code\Bin_and_cmd" Finally, the command prompt program cmd.exe is started.

SETUPDOS is the only program that needs to be accessible from the desktop. Copy "setupdos.cmd" from "PARAM_for_export\Source_code\Bin_and_cmd" and paste a shortcut on the desktop.  When this shortcut is clicked, a command-prompt window opens.  The first time SETUPDOS is run the window has default settings, these are not convenient for the type of work to be done.  They can be changed by right-clicking on the blue bar at the top, and selecting "Properties" from the drop-down window. Within "Options" change "Edit options" to put a tick in both "QuickEdit Mode" and "Insert Mode" For "Font", "TT Lucida Console" and "Size" 14 are easy on the eyes. In "Layout" if there is a tick in "Let system position window" remove it. Set the window position to "Left" = 6, "right" = 4, i.e., the start is at the very top left point on the screen.  Set the window size to fill the entire screen, and set the "Screen Buffer Size" "Height" to a large number, 6000 is good - that allows scrolling back a long way.  Click on "OK" and exit the "Properties" window.  The next time the shortcut is clicked the window that is made should look better than the default. 

Initially, the domain consists of the folders "Data_normal", "HTML files", "Optimize parameters", "Source_code", "Survey", and "Working".

Utility programs for preparing data-sets for PARAM

FILES (Constructs a file "files.out" that contains the names and empirical formula of the reference data files.)

FILES creates a list of all reference data files in a folder and writes it to a file "all.txt", It then uses a program "files_program.exe" to work out the empirical formula for each entry in that list and uses that formula to put the file names in chemical order, so H+ is first and Bi2I6 is last.  This list is then written to a file "files.out" which is needed by other programs. 

SELECT (Runs in a reference-data folder that contains "file.out" and generates a new file "bits.txt".)

Program SELECT selects a set of systems to be used in a parameter optimization run. This set is written to a file named "bits.txt"  SELECT is run in a folder that contains a collection of reference data but before it can be run, a file "files.out" must be created in the folder, this is done using utility program "FILES"

To run SELECT, navigate to the folder "Data_normal" This contains a set of reference data, each entry ends in ".mop" and a file "files.out"  Use SELECT with one of the following options: ALL, AND, NOT, ONLY.  SELECT reads "files.out" and uses the options provided with SELECT to construct the file "bits.txt"  As soon as a good version of "bits.txt" is made, it should then be moved to a different folder, likely "Working" or "Optimize parameters", or re-named.  This will prevent it being over-written if SELECT is run again. The easiest way to run SELECT is to use one of the following options.

ALL (Runs SELECT with the "ALL" option.)

Selects all the files in "files.out".  This option would normally be used only when a new method is being prepared for release.  It allows operations, such as surveys and statistics, to be performed on a very wide range of chemistry.

AND (Runs SELECT with the "AND" option.)

Selects the files in "files.out" that contain all the elements specified; other elements are allowed.

NOT (Runs SELECT with the "NOT" option.)

Selects the files in "files.out" that do not contain any of the elements specified. Only useful for selecting sets such as the inorganic compounds or non-lanthanides.

ONLY (Runs SELECT with the "ONLY" option.)

Selects the files in "files.out" that contain one or more of the elements specified; no other elements are allowed.

Utility programs for analyzing results of PARAM

STATS (Constructs statistical analyses of the accuracy of different methods.)

Run program STATS in a folder that contains the results of survey jobs run using PARAM.

STATS generates statistical information on the accuracy of methods that can be run using MOPAC.  Up to 12 different methods can be compared in one run.  There is no simple measure of accuracy for computational methods, so four basic quantities are used - heats of formation, geometries, ionization potentials, and dipole moments.  The accuracy of a method also depends on the systems being used.  By default, all systems are used, but this default is the least informative.  A better option is to compare how well the different methods work with specific groups of elements, for example H and C for hydrocarbons, H, C, N, and O for some simple organics, etc.

Each run of STATS will generate entries in the folder "HTML files" If there are no statistics files already in "HTML files", new files will be created.  If there are files, then the summary files will be updated to include the latest results from STATS.  After several runs of STATS these summary files can start to look untidy so an option is to delete all the STATS files from "HTML files" and start over.

SETUPMAP (Constructs a HTML file listing all the reference data, and individual HTML files for each reference datum.)

SETUPMAP generates a list of reference data and a local URL for each of the individual reference data files.  The graphics are best viewed with Firefox, but first JSmol must be installed and Firefox needs to be edited to allow local URLs to be opened using JSmol, see Installing JSmol for instructions on how to do this.