# construct the template in a more self-documenting way.
# this is a list of name => width values.  you can add
# more info here if you like; just adjust the loops
# that pull information out of it.  having all this
# information in one place means that you only have to
# change one place...

my @INPUT_COLUMN_INFO =
  ( unknown1 => { width => 12 },
    unknown2 => { width =>  4 },
    source   => { width => 20 },
    dest     => { width => 20 },
    ... );

# you can also use this structure to do clever things
# such as building a hash out of each input line to
# extract values by name instead of by numeric index.

# =================================================
# derived constants

my @INPUT_NAMES;    # names in original order
my $INPUT_TEMPLATE; # build up an 'unpack' template
my %INPUT_POS;      # map from name to index

for ( my $i = 0; $i < @INPUT_COLUMN_INFO; $i += 2 )
{
    my ( $name, $info ) = @INPUT_COLUMN_INFO[$i, $i+1];
    push @INPUT_NAMES, $name;
    $INPUT_TEMPLATE .= "A" . $info->{width};
    $INPUT_POS{$name} = @INPUT_NAMES-1;
}

# =================================================
# use during processing

while (<>)
{
    # construct hash using slice
    my %input;
    @input{@INPUT_NAMES} = unpack $INPUT_TEMPLATE, $_;

    # now access info from the hash
    my $src  = $input{source};
    my $dest = $input{dest};

    # ...
}