@sorted = sort { length($b) <=> length($a) } @unsorted;
[download]

Mikhail Gorbachev, Lucille Ball, and I all thank you.

-clay

Well, I don't really know what you mean by "smarter" (I thought, at first, that you meant more efficient or, perhaps, with less bugs than your version but then you go on to say that your version is "needlessly messy" which makes me wonder...) but, except for the bug where any two lines that happen to have the same length would get concatenated together, yours was close to being pretty efficient. Perhaps you'd like the Schwartzian Transform?

@arr = ('medium', 'longest of all', 'short');

@sorted = map  { $_->[0] }
          sort { $b->[1] <=> $a->[1] }
          map  { [$_, length] } @arr;

print "$_\n" for @sorted;
[download]

Of course, that's perhaps also "needlessly messy," since Big Willy's version certainly will work, albeit perhaps a little more slowly (depends on how many lines you're sorting and what order they're in).

*shrug* HTH

bbfu
Seasons don't fear The Reaper.
Nor do the wind, the sun, and the rain.
We can be like they are.

"The Transform" makes sense only when the function is expensive to compute. Last time I looked, length wasn't all that expensive. {grin}

-- Randal L. Schwartz, Perl hacker