>>24
Required fail only makes people annoyed when pages break because of something stupid. Yes, it's bad programming that's usually to blame for bad XHTML, but sometimes it's people screwing the code up because they changed something without knowing how XHTML works. There's a *lot* of that in the corporate world for some strange reason, and many of these people aren't sensible enough to stay out of things they shouldn't be messing with because it "looks like HTML" ... and then they go and test their changes with MSIE which doesn't support XHTML properly so the server's almost certainly sending it as text/html which defeats the purpose. Then you have a bunch of clients who can't access the site because *they* are sensible enough to use a browser that fails on bad XML.
This wouldn't be an issue if they'd implemented death on syntax errors from the very start, but HTML has traditionally been an unfortunately permissive environment, leading people into this sense of security that their code won't cause immediate death to the whole site.