9 January 1998: 3j, 6j, and 9j routines have been split out of ulib.f and placed in cnjj.f. Mod(i,2) has been changed to IAND(i,1) in cnjj.f and spinv.f 19 April 1997: Fixed bug which caused intensity errors for strengths except for dipoles with TYP+I1=0. Fixed bug in strengths when component dipoles are calclated. Both bugs have existed since 9 March 1997 revision. 18 April 1997: Fixed minor bug which caused incorrect phases for imaginary sine parameters. Made improvements in way rho was initialized. 3 April 1997: Fixed minor bug which caused wrong parameter names in SPFIT when the number of vibrational states was greater than 9. 14 March 1997: Fixed minor bug in SPINV.FOR which did not allow first pair of vibrational states to be l-doubled. Output file .cat from SPCAT now will include lines with uncertainties greater than 999.9999, but the uncertainty will be printed as 999.9999. Previously lines with a large uncertainty were not included. Calmrg has been modified so that lines with such large uncertainties will not be output to the .mrg file. 9 March 1997: Fixed bug in CALFIT which did not use absolute value of F for negative F values in the line file to calculate the maxmimum F value. Added option to calculate component elements for transition dipoles in the .str file. Program Stark has been updated to accomodate this new kind of .str files. New extensions for l-doubled states has been included in SPINV.F These extensions allow calculations to be made for l-doubled states (or E states in a methyl rotor). ESYMWT on the option cards of the .par file have been modified to accomodate this option. See spinv.txt for details. Documentation is now found in .txt files rather than .doc files, since the latter extensions had bad side effects with MS-Word aware browsers. 9 December 1996: A bug in the intesity part of spinv.f has been corrected. This bug was due to phase conventions which made transition dipoles anti-Hermitian with respect to v. This bug only affected intensities where there was (1) interaction between vibrational states and (2) both permanent and transition dipole moments. 21 June 1996: Maximum number of vibrational states is now 98, and maximum number of J values has been increased to the range -260 to 360. Improved code for constrained parameters so that derivatives are calculated as a single unit. Temperature can be specified for intensity calculation. 23 May 1995: Fixed bug in spinv subroutine SETOPT which gave incorrect spin weights for the I(tot) representation. Improved code for constrained parameters so that the error bars scale with the constraint. Also, changed options for suppressed F quantum number (i.e. when F is negative). As before, negative F will signal that the levels will be calculated using a selected F and then the hyperfine contribution from the last spin will be removed to first order. Now if F=-1 then the program will make a guess for a good F to use. If F < -1, then the calculation will use an F which has the absolute value of the F specified. 24 March 1995: Modified spinv to include higher order centrifugal distortion terms to the transition dipoles. 1 March 1995: Modified spinv so that negative ESYMWT now diables ESYMWT feature. Default value of ESYMWT is now -1. 28 December 1994: Fixed bug in calmrg.f which made it ignore lines from the .lin file with N"=0 or N'=0. 14 December 1994: calpgm32.exe is a self extracting file which contains the IBM PC 386 executables for spfit,spcat,dpfit, and dpcat. These were compiled with MS Fortran Power Station as a 32 bit DOS application. The extender and notes for installation are included. This file replaces calpgm32.zip. 4 December 1994: Modifications were made to allow dumping of a Hamiltonan matrix from SPCAT or DPCAT. It can be used to check the definition of a given parameter or operator by running the cat program using a test var file that only includes the parameters in question. Documentation has been revised and (perhaps) clarified. 19 November 1994: calpgm32.zip contains the IBM PC 386 executables for spfit,spcat,dpfit, and dpcat. These were compiled with MS Fortran Power Station as a 32 bit DOS application. The extender and notes for installation are included. 7 November 1994: A new version of spinv is now available. This new version will handle vibration-electronic states with different spin multiplicities or different Kmin. Intensities can be calculated for transitions between states of different spin multiplicity. There is also an optional way of using spin multiplicities which are greater than 9.