If you're sure there's always going to be whitespace between names, you could do something like this:

sub parse_name {
    my ($namestr) = @_;
    my @names = split(/\s+/, $namestr);
    my ($last, $first, $middle) = ($names[0], $names[1], $names[2]);
    $last =~ s/,$//;
    $middle ||= "";
    return [ $first, $middle, $last ];
}
[download]

Note that the subroutine also trims any comma from the end of the last name, and returns a blank middle name if one wasn't defined.

Now call the subroutine parse_name() with a name string, and you'll get a reference to a list containing the first, middle, and last names.  For example:

                                                                      
+    
use strict;
use warnings;
                                                                      
+          
my @data = (
    'Smith, John',
    'Thompson, Frank A',
    'Smith, John A JR',
    'Smith, John A III',
    'Smith, John A (Johnny)',
);

foreach my $name (@data) {
    my $p = parse_name($name);
    printf "First(%10.10s)   Middle(%5.5s)   Last(%10.10s)\n", @$p;
}

__END__

Output:
First(      John)   Middle(     )   Last(     Smith)
First(     Frank)   Middle(    A)   Last(  Thompson)
First(      John)   Middle(    A)   Last(     Smith)
First(      John)   Middle(    A)   Last(     Smith)
First(      John)   Middle(    A)   Last(     Smith)
[download]

Update:  Added test code.