XHTML in practice was too strict and tended to break a few other things (by design) for better or worse, so nobody used it...

That said, actually writing HTML that can be parsed via an XML parser is generally a good, neighborly thing to do, as it allows for easier scraping and parsing through browsers and non-browser applications alike. For that matter, I will also add additional data-* attributes to elements just to make testing (and scraping) easier.