Be careful to split with regular expressions - character set and
locale issues:
<pre>
<?php
$string = 'blårbærøl er greit';
$string = iconv( 'utf-8', 'latin1', $string );
setlocale( LC_ALL, 'nb_NO.iso-8859-1');
var_dump( preg_split( '/\W/', $string ) );
?>
Output
array(6) {
[0]=>
string(2) "bl"
[1]=>
string(2) "rb"
[2]=>
string(1) "r"
[3]=>
string(1) "l"
[4]=>
string(2) "er"
[5]=>
string(5) "greit"
}