Er, these all fall into the category of syntax errors. XML is just more permissive about the places where you can get away without handling these particular characters specially. Really, when you next tell someone this you'll be incorrect if you artificially call this "balancing" but the other stuff "invalid". Its equally invalid either way. This isn't a sort of "well ... it was mostly ok" operation. XML validity is either 100% one way or the other. This just happens to be a rule that can tip your XML into being invalid.