CHANGES IN PROFCN VERSION 1.06 FROM VERSION 1.05 February 2015
1. Improved the logic that determines which of the several methods
available are used to calculate the radial functions of the second kind.
2. Changed some of the formating and content of the output files.
3. Improved the code for the calculation of r2 and r2d using the
Legendre function expansion.
4. Removed several lines of code that remained from earlier
versions but were no longer needed.
CHANGES IN PROFCN VERSION 1.05 FROM VERSION 1.04 October 2014
1. Replaced real*16 variable c with iqint(c) in a max function argument
to be compatible with some compilers that require this.
CHANGES IN PROFCN VERSION 1.04 FROM VERSION 1.03 May 2013
1. Corrected some typos in comment statements.
2. Added open statements for the output files fort.20, fort.30, fort.40,
and fort.50 to accommodate some fortran compiles that require them.
CHANGES IN PROFCN VERSION 1.03 FROM VERSION 1.02
1. Corrected several comments at the beginning of subroutine r2int.
2. Changed and improved the subroutine gauss that obtains the abscissas and
weighting factors for the Gauss-Legendre integration used in subroutine r2int.
CHANGES IN PROFCN VERSION 1.02 FROM VERSION 1.01
1. Changed subroutine conver so that the first ratio of d constants enr(1) is
calculated using backward recursion when l - m is equal to zero or one. This
was done by changing the definition of ilow to lm2 + 1 where it is calculated
prior to entering the do 200 loop. In version 1.01 ilow is set equal to
max[lm2 + 1, 2]. The integer lm2, given by (l – m)/2, is equal to zero when l
is equal to either m or m + 1. In this case ilow is now equal to one instead of
two. When c is small, especially when m is large, calculation of enr(1) using
the closed form eigval - glist(1) involves significant subtraction error and
will reduce the accuracy of the radial functions computed for l = m and l = m+1.
The value for enr(1) calculated from backward recursion for l = m and l = m + 1
is accurate.
2. Redefined jbes in subroutine r1bes to be max[jtop + lm2, jbot] to make sure
the correct number of terms are used for the next value of l.
3. Corrected the definition of dneg and idneg in comments at the beginning of
subroutines dalt and r2leg.
4. Changed format for writing the first derivatives of the angular functions to
the diagnostic file fort50. New format is f40.31 so that quadruple precision
values are written.
5. Made a number of improvements to comments as well as several cosmetic changes
that have no impact on the resulting function values.