Programs for ROtational SPEctroscopy

        This is a collection of programs for various aspects of the rotational spectroscopy problem. For those who are already using these programs an upgrade is possible from here. The site is also to serve as a documented archive for the IFPAN group.

        The main reason for creating this site is the experience that the advances in computer power have not been reflected in the quality of software available to a typical rotational spectroscopist. The codes of 1960/1970 are still in use, and many are unparalleled in the beauty of their FORTRAN. However, as handed down to current students, they suffer from very sparse documentation, lack of source commenting and cryptic input/output. Most of the programs listed here offer some advance on these counts.

        The level of the listed programs is very uneven. They range from 'state of the art' solutions to specific problems (e.g. QSTARK, STRFIT), to simple, yet useful tools (EVAL, PLANM). The various spectral fitting programs now face strong competition from the superb SPFIT/SPCAT package of H.M.Pickett. I am still surprised, however, that the older fitting programs have by no means been relegated from use. I hope that this is not entirely due to inertia and familiarity, but also to greater user friendliness and minimal learning curve. Nevertheless, the current importance of SPFIT/SPCAT is such that a separate section is devoted to various add-ons to get even more out of that package.

        The majority of the programs on this site have been written in our group although the number of donated programs is on the increase. Our programs are often based on preceding older code, for which I tried to acknowledge authorship when known. All of our programs are deeply rooted in the 1970's discovery that computing can have a 'human face' and programs can be INTERACTIVE! This is nevertheless not taken too far, as coding is, on the whole, rather conservative. Considerable effort has gone into commenting the source, and intelligible input and output, but interactivity rarely goes beyond simple ASCII menus. With some exceptions mice and GUI's are studiously ignored. The several graphics programs are keycode operated, although help screens are available. All programs contain extensive descriptive headers in the listings.

        Some older programs of other authors (like the classic STARK) are put out here without explicit consent of the authors, who may no longer be reachable. In keeping with the tradition of this field I assumed implicit consent of the original authors to such action. If this is not the case then please let me know. Several programs written by other people have had internal workings unaltered, while changes have been made to input, output, and commenting in the listing.

        Making these programs available publicly is a way of paying my debt to the many predecessors in programming for rotational spectroscopy from whose code I have been able to draw freely. I would be happy to accept programs to add to this site on a deposited basis. The required minimum would be the listing, specimen data, output for the specimen data and, as importantly, concise but clear documentation, including a suggested reference for example of use as well as citation.

        Comments on the programs already put up are also appreciated but I cannot promise a rapid reaction time unless some really grave error is spotted.

        Zbigniew Kisiel, kisiel@ifpan.edu.pl

        The use of these programs is free.

        There is also no objection to modification of these programs, but to avoid confusion please change the program name and insert appropriate comments at the top of the listing.

        If a program is found useful in publishable research and you want to acknowledge it, please cite the paper suggested in the description of the program - in particular if the program is not one of mine. If there is no suggested publication then please cite this site as:  "Z.Kisiel, PROSPE - Programs for ROtational SPEctroscopy, http://info.ifpan.edu.pl/~kisiel/prospe.htm."

        Version date: Each of my programs displays this on the header printed at the beginning of operation, and also carries it near the top of the listing. Many of the programs are in continuous use, and new capabilities are put in as required. It is therefore recommended to check these version numbers, especially as some of the .EXE files may not necessarily correspond to the latest listing.

        Internal documentation: Sufficient documentation for each program should normally be available on its web-page. However, please note that many programs carry additional/duplicate documentation at the top of their .FOR source listing. It is recommended that this be inspected, even if you are not FORTRAN literate.

       All programs are now available both as FORTRAN source (see the notes below) and as DOS/Windows executables. If an explicit link to the executable is not provided, just change the ending of the link to the FORTRAN listing from .FOR to .EXE.

       This web site is on a UNIX machine and all ASCII files (i.e. FORTRAN listings, data and output files) are in the UNIX ASCII format. In this format each line is terminated only with the LF character, whereas in WINDOWS/DOS, in particular, lines have to end with both LF and CR characters. Thus there may be compatibility problems on direct downloading to file onto WINDOWS/DOS machines. The easiest way to get the text files, such as program listings, is to read the file first into the viewer and then to save it using the FILE / SAVE AS options.

       If ASCII file type conversion is necessary it can be made with the unix2dos utility, which though native to UNIX environments, is also available for DOS. If you have no better ideas, try this University of Vaasa site - download the file dtuc11.zip and use the program UTDC.EXE contained therein.

General:

        All programs are in FORTRAN of basically F-77 vintage and, unless they contain graphics, compile in standard manner in DOS/WIN, UNIX, and MAC environments. The expected reply to YES/NO type of question is 1/0, unless Y/N is explicitly suggested in the prompt. Pressing just the ENTER key often suffices for 0.

       Most contemporary compilers no longer use default static memory allocation that preserves values of variables previously set in some subroutine. Many programs assume this to be the case so that you need to use an appropriate keyword to enforce static allocation. For example on the f77 compiler on SGI machines this is the -static option, whereas with Intel Visual Fortran you have to use option -Qsave. Note that with some compilers optimization options used not to be safe. With good compilers this is no longer the case, but if problems crop up it is still good to check by disabling optimization.

        Minimal tweaking may be necessary for several trivial issues such as

• the edit descriptor in FORMAT statements used to halt carriage control - I generally use the $ character for this, but the occasional backslash character \ may have slipped in and this is less widely recognized
• the bell signal is hard coded as CHAR(7) and many systems will probably just ignore it

Graphics:

        Some graphics is clearly better than none and after early experimentation I selected the graphics package that appeared for the first time with Microsoft Fortran 5.0 (MSF5). This is regrettably tied to the PC DOS/WIN world, but works well enough. Further development has been to combine these graphics with large memory requirements, and the solution came with Microsoft PowerStation Fortran 1.0 (MSPS1). For the NT/WIN95+ systems the same graphics became available in Microsoft PowerStation 4.0 (MSPS4). Microsoft then sold off its Fortran to DIGITAL who were, in turn, acquired by COMPAQ. Another change of ownership then took place, and the current version (10) is under the INTEL brand name. Various compatibility issues are as below:

        MSF5 (Microsoft Fortran 5): this code seems to be the most troublefree, but it has to be run in a full-screen MS-DOS window (or just DOS). Programs use only the low DOS memory, so cannot be too large and screen resolutions higher than VGA are not supported. This version was the only one tested on an Apple (PowerMac) and worked very acceptably on the W3.0 emulator found there, and is able to run on PC's ranging from 286/Hercules to PIV/Windows 2000+.

        MSPS1.0 (Microsoft PowerStation Fortran 1): programs have access to much more memory and to resolutions higher than the VGA. Full screen operation is still by far the best mode, although (extremely slow) operation in a window is possible. A 32-bit DOS extender, DOSXMSF.EXE, is required and the programs are best launched from DOS. For W95/W98 it is necessary to block detection of Windows from DOS, which is done in the properties/advanced section of the MS-DOS window. Furthermore resolutions higher than VGA, which work fine within W3.1, are subject to various problems in W95+. Also note that compilation with the optimization options produces code that crashes very nicely.

        This was for a long time our preferred version, but there are portability problems. I am not clear on the legalese concerning the DOSXMSF.EXE extender (and the associated DOSXNT.386). They may, or may not be available from the Microsoft www page, but they seem to be available from various sites. They are, for example, included in the DOS/Windows3.x version of the ICON-EDiT package available from http://iacrs1.unibe.ch/members/iconedit.html#iconedit. Note that there are various sizes and dates of these files and some are not compatible with W95+.

        MSPS4.0 (Microsoft PowerStation Fortran 4): this produced native NT/W95+ applications which was fine but recompilation of older code with minimal modifications produced results that were inferior in many respects to the code from MSPS1.0. In addition there were many evident bugs in the compiler.

        CVF6 (Compaq Visual Fortran 6): this turned out to be a much more reliable implementation of Fortran for the Windows environment, and most of the old graphics inherited from the MSF5.0 still works. The OpenGL graphics (available since MSPS4.0) is well documented so this version may form a springboard in that direction. Much of the behaviour suspect with MSPS4.0 has been understood. Many of the changes since MSPS4.0 have the robust feeling of DEC Fortran about them, and there has been useful manufacturer support for this version.

        The downsides are relatively few, one being that links with the past have to be broken as the executables will no longer run in Windows 3.x or clean MS-DOS. Also the executables have become three times larger, and the graphics runs many times slower than with MSPS1. It is a sign of the times that nowadays nobody seems to care about such things, since they are usually more than adequately compensated for by advances in the hardware. Finally the number of changes in the code are just a little bit too many to keep a single source compatible with running on older systems, so older graphics versions of my programs are being successively discontinued. This version (6.6B) is currently our preferred version, akthough the next incarnation (see below) is also used interchangably.

        IVF10 (Intel Visual Fortran Compiler for Windows): This is the latest incarnation of this indestructible compiler following transfer of the CVF compiler line to INTEL. It is assured to be backwardly compatible with the CVF6 version.

        If you are confused, here's a recap - the ownership trail for this compiler is: MICROSOFT->DEC->COMPAQ(+HP)->INTEL

        Some of the graphics programs, in particular ASCP, SVIEW and PMIFST, provide a route to high-quality printouts by outputting files for gle, which can then generate PostScript output.

        gle stands for graphics layout engine, and was originally written by Chris Pugmire, up to version 3.3, in the early 1990's. This is a freely available program and its appeal for many computer users is that the control file for the drawing is plain ASCII, so it can be manipulated with any editor. The language is fairly straightforward, yet powerful, so that very sophisticated diagrams can be created with even limited computer resources. gle is reasonably popular in academia and various subversions are available on many sites.

        It is currently recommended that you install gle 4.0.7, but you may also check these background notes on gle. You also need to preinstall the GSview/Ghostscript package. To obtain publishable quality PostScript printout from, say ASCP, first generate the appropriate files using the G option of ASCP. This will generate several files, only one of which has the extension .gle, and has the name, say, xxx.gle. Providing gle 4 (version gle 4.0.7) has been installed, open the command prompt window on the directory containing the .gle file and type

        gle_ps xxx.gle,

which generates a file called xxx.ps. This file can then be viewed and printed using your preferred way of handling PostScript, typically with GSview. The xxx.gle control file can be modified with any ASCII editor to obtain a customized diagram.

        More recent versions of gle 4 are available but, for the time being, only the version 4.0.7 (and earlier) are 100% compatible with the scripts generated by the PROSPE programs.

Use the refresh option of your browser to make sure you are reading the current version.