++ for spreading the truth about Barney! ;)

... but why do you search for both upper and lowers in RtoD() when you're already made everything uppers with tr/a-z/A-Z/?

As for the unweildiness and uncombined searches, you could simplify the script a lot by using a hash instead of a regexp:-

#!/usr/bin/perl -w

use strict;

my %numeral = (
    'I' =>    1, 'V' =>    5, 'X' =>   10, 
    'L' =>   50, 'C' =>  100, 'D' =>  500,
    'M' => 1000,
);

print "Enter name: ";

while ( chomp( my $phrase = <STDIN> ) ) {
     my $sum = 0;

     $phrase =  uc($phrase);
     $phrase =~ tr(U)(V);

     for my $letter ( split // => $phrase ) {
         $sum += $numeral{$letter} || 0;
     }

     markOfTheBeast(), last if $sum == 666;

     print "The number is $sum ... blessed be!\n",
       "Enter name: ";
}

sub markOfTheBeast {
    print <<END_OF_DAYS;
*********************************************************************
*YOU HAVE EXPOSED THE SIGN OF THE BEAST.  PLEASE REDEEM YOUR GET OUT*
*  OF HELL FREE CARD AT THE NEAREST STARBUCKS OR PORTAL TO HELL TO  *
*                  AVOID THE ONCOMING APOCALYPSE.                   *
*********************************************************************
END_OF_DAYS
}
[download]

However, if you did want to do that with regexps, you could combine them by keeping the hash and doing something like ...

sub RtoD {
      # ... or uc(shift) to be paranoid. :-)
    my   $n =  shift          || return 0;

    if ( $n =~ /([IVXCLMD])/ ) { return $numeral{$1} }
    else                       { return 0            }
}
[download]

    --k.