Maybe this should be a nother top-level question, but can you tell me know you knew to promote my string to an array? I was playing with subroutines and arguments, and I found that all these generated the same answer (5):

my $a = 2; 
my $b = 3;
print &add($a, $b) . "\n";

my $opt = "$a, $b";
print &add(eval($opt)) . "\n";

my @opt = ($a, $b);
print &add(@opt) . "\n";

exit;

sub add {
   my ($x, $y) = @_;
   return ($x +$y);
}
[download]

Use split to split a string into a list.

#!/usr/bin/perl
use warnings;
use strict;

use List::Util qw{ sum };

my ($x, $y) = (2, 3);

my $opt = "$x, $y";
my @opts = split /,\s*/, $opt;
print sum(@opts);
[download]

The dashes are just a convention. They were somehow common at the time Perl/Tk was created, so there might be other modules using them. See the explanation of "unary -" in perlop.

map{substr$_->[0],$_->[1]||0,1}[\*||{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^*ARGV,3]

... [how] you knew to promote my string to an array?

A subroutine such as your  add() can be called with an argument list; choroba just made the arguments you wanted to pass into an expression (specifically, an array) that would be accepted as a list expression by a subroutine call; see perlsub. (BTW: There's no need to use the  & sigil for a subroutine call. Indeed, you should avoid this syntax because it has some oddball effects; see especially the paragraph beginning 'A subroutine may be called using an explicit "&" prefix.' in the DESCRIPTION section of perlsub.) (Also BTW: IMHO it's not, in general, a good idea to use the variable names  $a $b as lexical names even in short example code; these are Perl special variables and have special significance and usage; see perlvar.)

... arguments ... all these generated the same answer ...

I'll avoid using the  & sigil prefix in all discussion and examples.
    print add($a, $b) . "\n";
Pass a list of two scalars (list elements are always scalars) to the  add subroutine.
    my $opt = "$a, $b";
    print add(eval($opt)) . "\n";
First, eval the Perl string list expression  "$a, $b" and then pass the resulting list returned by eval to the subroutine.
    my @opt = ($a, $b);
    print add(@opt) . "\n";
Expand an array to the argument list of a subroutine call.

c:\@Work\Perl\monks>perl -wMstrict -le
"my ($x, $y) = (2, 3);
 ;;
 print add($x, $y);
 ;;
 my $opt_dq = \"$x, $y\";
 print add(eval($opt_dq));
 ;;
 my $opt_sq = '$x, $y';
 print add(eval($opt_sq));
 ;;
 my @opt = ($x, $y);
 print add(@opt);
 ;;
 my @ra = (99, @opt, 98);
 print add(@ra[ 2, 1 ]);
 ;;
 exit;
 ;;
 sub add { my ($x, $y) = @_;  return $x + $y; }
"
5
5
5
5
5
[download]

Questions:

• What's the difference between feeding the strings  "$x, $y" and  '$x, $y' to eval?
• What's  @ra[ 2, 1 ] all about? (Hint: see Slices.)

... dash notation ...

This is notation used in Tcl/Tk that was, AFAIU, intentionally imported into | implemented in Perl/Tk just to make it look more familiar. Otherwise, the  => is just a "fat comma"; see perlop.

---

Give a man a fish:  <%-{-{-{-<