John Lim from PHP Everywhere posted a benchmark, where he compared parsing the titles of a RSS feed using regex, DOM, XPath and Sax. He came to quite “shocking” results in that the regex approach was 60

times faster than the DOM approach. While the task was perfect for a regex job (get me the content of everything between and ), the difference seemed tooo big in my point of view. Therefore I looked at John's benchmark and tested it on Linux and Mac. The results were a little bit different, DOM was “only” 6 times slower

now than regex.. I can live with that ;)

More arguments, opinions and facts about that topic can be read in the comments of that post.