I love perlmonks because it constantly reaffirms that there's always more than one way to do it. When you have a question, there's usually more than one 'right' answer. The question is: when I call some encapsulated piece of code, how will I know whether it succeeded or failed? Was there an error and what should I do?

The reason why I like the PBP 'rule' is because it's an easy to remember answer to a question that should be considered on every function call. Defensive programming means checking return values. The concept of what to return on failure is like a mini-protocol. "Use a bare return to return failure" is such an easy protocol....

For all the folks who think it's stupid, and hate the idea of typing 'return 1;', make up your own mini-protocol and a call it Perl's SecondBest Practices. LOL :-)

"Don't return undef when you mean an empty list" -- this is some serious wisdom, too. I'm really glad you said that. :-) I think it's worth mentioning that PBP also recommends using Contextual::Return to deal with that exact 'complication'.

IMO The Best Practice here is a combination of using "return;" anywhere something goes wrong and at the bottom of your subs using a block like this:

  use Contextual::Return;
  sub items {

    ## Something failed!
    if ($failure) return; #TODO: Consider croaking!

    ## Being explicit!
    return (
        LIST     { qq(Item1 Item2) }
        HASHREF  { {Item1 => 'big', Item2 => 'small'}}
        SCALAR   { 2 }
        );
  }
[download]

I think knowledge sharing and promoting 'the right way' is what will allow the Perl Community to continue to flourish. I think Damian was Noble in pursuing this ideal.

Kurt

Updated: Added points about Contextual::Return