You only think that both versions of your code use the same variable, but the second version does not. What happens in fact is, that your sub print_summary uses the global variable $type ($main::type, to be verbose), while you later on declare a lexical variable $type. The strict pragma tells you about that, which is why you should use it:

use strict;

sub print_summary {
    if ($type =~ /w/) {
        print "Hello\n";
    };
};

my $type = "w";
print_summary;
[download]

Here, Perl raises an error that you're using an undeclared variable $type in print_summary, which tells you that declaring the variable too late prevents Perl from seeing it.

If $type is a variable that really needs to be available to your entire program, then you should put the declaration at the top of the file. However, in this case, I'd suggest that you consider passing it as an argument to your function:

sub print_summary {
    my ($type) = @_;
    if($type=~/w/) {
        print "hello\n";
    }
}
# ... later ...
print_summary("world");
[download]

Actually, what I'd probably do is more along these lines...

sub summary {
    my ($type) = @_;
    if($type=~/w/) {
        return "Hello, ";
    } elsif ($type =~ /h/) {
        return "world!\n";
    }
}
# ... later ...
print summary("world");
print summary("hello");
[download]

Notice here that the function returns the string to be printed, rather than printing it. This makes the function more reusable in the event that you need to send the results somewhere other than standard output, e.g., write them to a file, send them over a socket, or capture them to a string for subsequent processing.

If you're going to pass $type into the function I'd do this:

sub summary {
    my $type = shift || return '';
[download]

To avoid warnings later when someone forgets and tries to

print summary();
[download]

I think in that case I'd prefer to *have* the warnings.

• use <code> tags;
• use strict; # and perl will help you to understand;
• use warnings; # as well!