Attributes are cool whenever out-of-band data is required. Some examples :

• I have a toy application here, where I use attributes to specify access levels for subroutines.
• If your application consists of a chain of list-munging subroutines, you can easily add tracking information as which routine touched which item without having to change/adapt every routine and without conflict with the output.
• The always present compiler hints or runtime type enforcement can also be implemented via attributes (like the :Integer attribute), but of course, this is also out-of-bound information about your "data" (this time, your program is regarded as data by perl)

perl -MHTTP::Daemon -MHTTP::Response -MLWP::Simple -e ' ;    # The  
$d = new HTTP::Daemon and fork and getprint $d->url and exit;#spider
($c = $d->accept())->get_request(); $c->send_response( new   #in the
HTTP::Response(200,$_,$_,qq(Just another Perl hacker\n))); ' #  web
[download]

What attributes are for? Uhmmmm... why you ask me? The camel book (3rd edition, page 745) shows my num PI : constant = atan2(1,1) * 4 as an example and tells briefly about other possible uses.

Why would I use them. I came around them in 180455 and replies. Yesterday night I read perldelta, today I was digging into the code and docs of Attribute::Protected. I found it interesting, but since it relies on Attribute::Handlers it would be interesting to know if it is safe to write production code with it.

attributes reports the usage attribute declaration for variables is still experimental. What I would like to have is, say, a measure of how much experimental attributes are: is it safe to play with them in subroutines now? Is it not? Code that uses Attribute::Handlers is reliable or is expected to break in the short/medium term?

Ciao!
--bronto

# Another Perl edition of a song:
# The End, by The Beatles
END {
  $you->take($love) eq $you->made($love) ;
}