The DOM-Level-1 spec is quite clear in the definition of the Element interface that tagName should be returned as uppercase: http://www.w3.org/TR/REC-DOM-Level-1/level-one-core.html#ID-745549614 This means that Node::nodeName should also return uppercase. MSIE returns element names uppercase. I can't get hold of an attribute node in MSIE, so I can't test that. So this would be good for compatibility. However, since XHTML is lowercase, there are some thorny questions. Obviously, if something is sent as text/xml or text/xhtml, then it's XHTML. However, if it's sent as text/html, what is it? I would say if it has either an xml declaration or an XHTML doctype, then it's xhtml. This is one of the lousy things about going lowercase - the DOM has assumed uppercase normalization and the HTML vs. XHTML boundary is fuzzy. You may want to consult with W3C folks on what to do here (or perhaps you already have and they suggested a lowercase normalization for HTML...)

We've talked about the xhtml issue in the DOM WG, but haven't come to any firm conclusions. You're right - we need to figure out the right direction for it. Element::tagName does the right thing w.r.t. casing. I'm assuming the bug is stating that Node::nodeName doesn't do the right thing for Elements, right?

Right. Sorry - I was getting various tests mixed up.

Fix for Node::nodeName casing for HTML elements checked in on 5/11/1999.

Verified fixed.