Installation of EPRFT The program has been compiled and tested on several flavors of Unix systems (Sun SPARC with SunOS 4.1.1, SGI Iris Indigo with IRIX 4.0.5, IBM AIX, CDC-4680 with EP/IX), IBM AT (DOS) using BC, Atari ST using TC, and VAX/VMS. On Unix systems, both gcc and the default C compiler (cc) should work. For other systems, some of the conditional defines may have to be modified. Most part of the "system definition" is done in "sys_def.h". Unix: On Unix systems, there is no graphical output of spectra unless a separate program for that purpose is installed, e.g., XGRAPH by D. Harrison (available from FTP servers). If XGRAPH is available, define XGRAPH in the Makefile. Settings in "eprft.h" (such as maximum number of data points, parameter sets etc.) may be edited. (The constants FILETYPE, ADD_ONE, NPMAX and VIEW can also be modified in the Makefile. Vide infra.) Look through the Makefile and edit path names, compiler (cc or gcc), compiler options etc. On any of the Unix systems tested, the program (in its standard configuration) should compile by simply running make and installing the binary (eprft) and man page (eprft.1) at their appropriate places. (Or run "make install" and "make install.man" after checking the path names in "Makefile".) IBM-PC: The program has been compiled by means of Borland C++ (3.1), but Turbo C (2.0) should also work. Atari ST or MegaST (monochrome): The program has been compiled by means of Turbo C 2.0, but version 1.1 should also work. VAX/VMS (not tested thoroughly): @eprftmak.com FILES: eprft.h sys_def.h eprftx.c amoeba2.c ffread.c fourier.c get_fnam.c getchr.c Configurable constants (eprft.h or Makefile): XGRAPH Define, if the file viewer "xgraph" is available FILETYPE 0 Default output file type (plain simulations) 0: binary LITTLE_ENDIAN (IBM-PC) data file (I2 integer) 1: Asyst file 2: ASCII file (I6), 1 number per line 3: ASCII file (8I7) 4: File from HP 1000 5: binary BIG_ENDIAN (Atari) data file (I2 integer) 8: x, y ADD_ONE 0 (0 or 1). Note: The abscissa scaling "units per point" is calculated according to units_per_point = scan_width/(number_of_points - 1 + ADD_ONE) NPMAX 4096 max. number of data points (power of 2). Note: On "small" systems such as IBM-PC/DOS the maximum value of this number is restricted.