This is, for the most part, just a picky restatement of some points I was trying to make above. But references are tricky, so I want to make these points clearly.

The statement
    $removed = [ @removed ];
creates a hard reference to an anonymous array initialized by a shallow copy from the  @removed array. The statement
    $removed = \@removed;
also creates a hard reference, but to the named  @removed array. The nature of both these references is the same, but their referents, the things they reference, are different. Hence, the difference in behavior when using these two references.

#$removed =  [ @removed ]; #1. please notice : this makes a copy first of @removed . # changes to @{$removed} will NOT affect @removed

It is not true that changes to  $removed will not affect the referent. Because the anonymous array is initialized by a shallow copy, accesses by reference to lower levels of the referent, if such exist, can change the referent. E.g.:

c:\@Work\Perl\monks>perl -wMstrict -le
"use Data::Dump qw(dd);
 ;;
 my @ra = ([ qw(fee fie foe) ], [ qw(original stuff) ], [ qw(uno dos t
+res) ]);
 dd \@ra;
 ;;
 my $hardref = [ @ra ];
 dd $hardref;
 print '-------------';
 ;;
 $hardref->[1] = 'ZOT!';
 $hardref->[0][2] = 'KAPOW!';
 $hardref->[2][1] = 'WHAM!';
 dd $hardref;
 dd \@ra;
"
[
  ["fee", "fie", "foe"],
  ["original", "stuff"],
  ["uno", "dos", "tres"],
]

[
  ["fee", "fie", "foe"],
  ["original", "stuff"],
  ["uno", "dos", "tres"],
]

-------------
[["fee", "fie", "KAPOW!"], "ZOT!", ["uno", "WHAM!", "tres"]]

[
  ["fee", "fie", "KAPOW!"],
  ["original", "stuff"],
  ["uno", "WHAM!", "tres"],
]
[download]

Note that a change to the "top" level of the  $hardref reference ('ZOT!') does not change the original content of the  @ra array, but a more indirect or "lower" level access ('KAPOW!' 'WHAM!') does. As I say, tricky.