Please note that I didn't pay a lot of attention to
keeping the "double" notation in the form of higher
precision floating point routines.  This means that
the Perl5 code won't be as accurate as the original
C-code.  It doesn't matter for my purposes.  If
anyone converts to Math::BigFloat for higher precision,
please send me the changes.  As it is I did a
quick check and found a difference of only 1.4 meters
between my Perl results and the results from a web-based
datum-shift calculator on the 'net.  -- Curt.