That's very similar to a situation I was thinking about, except that I would have stated it more generally: I would consider (locally) using symbolic references in a scenerio when they would be convenient and the risk would be checked by having the complete list of possible values of the reference ennumerated en toto in the code. Additionally, I never bother with strict for very short use-once scripts (e.g., anything I type at the command line after perl -e).

Truthfully, I don't always use strict in other cases either, but I'm starting to do so more often that I used to do. I'm up to the point of using at the top of all files that get included by another file now, which is progress... and for the record I have yet to see strict catch a mistake that warnings missed, except in a conversation on Perlmonks where a monk contrived an example for me to demonstrate how it could happen. However, the example (though contrived) was sound and I've taken it to heart somewhat.

And that's why I don't think question 2 as it stands will get the information the test writer actually wants. I would have no trouble explaining the value of strict at this point, but I don't yet have a firm habbit of always using it, which is what the test really wants to know. (The best way to find out? Have 'em write a section of code to do some simple task and see if they declare all of their variables.)

Actually, my biggest problem with strict is that it doesn't test what I want it to test: whether a subroutine tromps on a (possibly lexical) variable from outside the subroutine. This is for me a *way* more likely scenerio than the ones strict does check.

{my$c;$ x=sub{++$c}}map{$ \.=$_->()}map{my$a=$_->[1];
sub{$a++ }}sort{_($a->[0 ])<=>_( $b->[0])}map{my@x=(&
$x( ),$ _) ;\ @x} split //, "rPcr t lhuJnhea eretk.as
o";print;sub _{ord(shift)*($=-++$^H)%(42-ord("\r"))};
[download]