That looks like XML. I'd seriously consider using XML::Twig!

If you need some help show us some representative data and what you actually want to extract

It is of a fashion - impossibly malformed XML. The data now given above ( and specified output ) is now fully representative, this tool is effectivley a parser to take some poor output, to poor, but well-formed XML. Although Im keen to continue with this approach, is the solution on a differant approach ?

Actually that is a lot simpler to solve assuming one "element" per line:

use warnings;
use strict;

while (my $line = <DATA>) {
  # Assume one "element" per line
  if ($line =~ /<(\w+)>(.*?)<\/\1>/) {
    (my $midstr = $2) =~ tr/<>/[]/;
    $line = "<$1>$midstr</$1>\n";
  }
  print $line;
}

__DATA__

<sometag> my data here </sometag>
<anothertag> further text </anothertag>
<furthertag> not good as contains <this> in angle brackets </furtherta
+g>
<byetag> another possibility is <more> <than> one <angle pairs> in her
+e </byetag>
[download]

Prints:

<sometag> my data here </sometag>
<anothertag> further text </anothertag>
<furthertag> not good as contains [this] in angle brackets </furtherta
+g>
<byetag> another possibility is [more] [than] one [angle pairs] in her
+e </byetag>
[download]

---

DWIM is Perl's answer to Gödel

This violates your "no code" premise, but if your tags are all on one line, this seems reasonably robust and should be fairly efficient.

#! perl -sw
use strict;

while( <DATA> ) {
    s[^(<([^>]+?)>)(.*)(</\2>)]{
        (my $x = $3) =~ tr[<>][[]];
        "$1$x$4";
    }e;
    print;
}


__DATA__
<sometag> my data here </sometag>
<anothertag> further text </anothertag>
<furthertag> not good as contains <this> in angle brackets </furtherta
+g>
<byetag> another possibility is <more> <than> one <angle pairs> in her
+e </byetag>
<a tag with spaces> and content containing a false </a tag with spaces
+> and a real </a tag with spaces>
[download]

Produces

C:\Perl\test>junk2
<sometag> my data here </sometag>
<anothertag> further text </anothertag>
<furthertag> not good as contains [this] in angle brackets </furtherta
+g>
<byetag> another possibility is [more] [than] one [angle pairs] in her
+e </byetag>
<a tag with spaces> and content containing a false [/a tag with spaces
+] and a real </a tag with spaces>
[download]

Of course, it doesn't attempt to deal with attributes, or multi-line elements or nested tags or any of that good stuff.

Thanks - Considering all the responses, I think attempting to do this +without+ a small amount of code, ie an a single 'replace' expression is not the best approach. Thanks for this solution.