Since you're dealing with fixed length fields, substr works well:

$field_start  = 8;  # 0-based
$field_length = 8;

# Get current value:
$old_value = substr($s, $field_start, $field_length);
$old_value =~ s/\s+$//;

# Change the value:
$new_value = 'bla';
substr($s, $field_start, $field_length) =
   substr($new_value . (' 'x$field_length), 0, $field_length);
[download]

And if you're converting from fields of one size to fields of another size:

foreach (@old_records)
{
   my $name = substr($_,  0, 8); $name =~ s/\s+$//;
   my $desc = substr($_,  8, 8); $desc =~ s/\s+$//;
   my $date = substr($_, 16, 8); $date =~ s/\s+$//;

   $name = ...reformat name...;
   $desc = ...reformat desc...;
   $date = ...reformat date...;

   push(@new_records, 
      sprintf("%-7s %-13s %s\n", $name, $desc, $date)
   );
}
[download]

Well you have totally got me on the logic! Why not just print "NAME DESCRIPTION DATE\n" for 1..$n; anyway to convert what you have to what you say you want:

while(<DATA>){
    s{(NAME)\d*\s*(DESC)\d*\s*(DATE)\d*\s*}
     {$1 ${2}RIPTION $3\n}g;
    print;
}
__DATA__
NAME010 DESC025 DATE009
NAME011 DESC035 DATE009
[download]

In general there is nothing that perl *can't do* when it comes to munging text.

my %xref;
$xref{'NAME'} = 'NAME';
$xref{'DESC'} = 'DECSRIPTION';
$xref{'DATE'} = 'DATE';

while (<DATA>) {
#    s/(NAME|DESC|DATE)(\d*)/$xref{$1}/eg;
    s/(NAME|DESC|DATE)(\d*)/$2/g;

    print;
}

__DATA__
NAME010 DESC025 DATE009
NAME011 DESC035 DATE009
[download]

s/NAME\d{3}\sDESC\d{3}\sDATE\d{3}/NAME    DESCRIPTION   DATE/g;
[download]

Thanks that was the trick ... sorry about the confusing description I was trying to come up with example, but wasn't quite sure how to phrase it.

If your data is fixed width then unpack can be quite handy to chop out the data.

while(<DATA>) {
    @bits = unpack "A4A4 A4A4 A4A4", $_;
    print "[$_] '$bits[$_]' " for 0..$#bits;
    print $/;
}

__DATA__
NAME010 DESC025 DATE009
NAME011 DESC035 DATE009
[download]