Same for list

#!perl
use strict;

my @list = ('BULLOCK JOE A','SMITH, A DOE');

for my $name (@list){
  my ($surname,$first) = $name =~ /(\w{3,})/g;
  print "$first : $name\n";
}
[download]

But I don't have my data in a hash

The hash is used as a clear and concise way to list both the data which you have (the keys) with the results which you want (the values) and to test that an operation on the former yields the latter. There is no need for you to use a hash in your implementation. See How to ask better questions using Test::More and sample data for a recent discussion on why using a test in a question is a good idea and why (to my mind) using a test in an answer is equally beneficial.

The fact that my example iterates over a list (keys %names) should help you to use a similar list in your own code as do the examples from pod and ExReg.

You could just as easily do it with an array or list then. Modify the code hippo gave to use an array instead:

#!/usr/bin/env perl 
use strict;
use warnings;

my @names = (
    'BULLOCK JOE A',
    'SMITH, A DOE',
);

for my $fullname (@names) {
    my ($sname, $fname) = $fullname =~ /([A-Z]{3,})/g;
    print "Forename $fname extracted from $fullname";
}
[download]

There are three assumptions that need to be met: both first and last names must be at least three letters long, the last name must come first, and the middle name must be an initial (or at least less than three letters long. If you cannot meet those conditions, you will have a hard problem.