Two things that help in deducing what's going on with a split function are to tell perl to keep the delimiter matches as elements of the list it returns, and to print out the array line by line rather than with a print "@t\n";

This is how I would attack the problem:

#!/usr/bin/perl
print "Entrez votre ip :";
$ip=<STDIN>;
@t=split(/(.)/,$ip);   # Surround the delimiter in (), then the
                       # regexp matches become array elements
print scalar @t,"\n";  # Don't waste a variable
print "[$_]\n" for @t; # Surround each match by brackets,
                       # print each match on its own line.
[download]

Then your output becomes:

Entrez votre ip :21.23
11
[]
[2]
[]
[1]
[]
[.]
[]
[2]
[]
[3]
[
]
[download]

"Oh, each byte was a match," says I. Changing @t=split(/(.)/,$ip); to @t=split(/(\.)/,$ip); gives:

Entrez votre ip :21.23
3
[21]
[.]
[23
]
[download]

OK, good. Now we need to get rid of that line feed, and take the delimiter out of the array. Final code:

#!/usr/bin/perl
print "Entrez votre ip :";
$ip=<STDIN>;
@t=split(/[.\n]/,$ip); # Don't need \. for period in [ ]
print scalar @t,"\n";
print "[$_]\n" for @t;
[download]

Which outputs:

Entrez votre ip :21.23
2
[21]
[23]
[download]

Was that what you were looking for?