The sense I am gathering is that the proposal is not obviously insane, and in fact is a bit novel in that such a narrowly scoped adoption of XML syntax -- i.e., only to the extent that it both reflects the web as widely practiced and only to the extent that doing such does not introduce ambiguity into the grammar -- had not been considered before.
FWIW, it sounds sane to me to align validation as much as possible with the UA parsing in a way that issues that aren't really problems for the UA aren't flagged as invalid. Closing slash on void elements sounds like a good example of "this is invalid because we're sticking to our fixed ideas"[1] rather than "this is invalid for technical reasons like causing ambiguities in DOM parsing". So I support Sam's approach. That said, HTML5 must see <input type="checkbox" checked/> as a checkbox input with a "checked" attribute. Finding a "checked/" attribute and not checking the checkbox is not compatible with the web (learnt the hard way!). Perhaps finding a slash in "attribute name" mode on void elements should be a parse error if the next character is not > ? (Pretty certain the specs already disallow attribute names starting with forward slash.) [1] disclaimer: not intended as a flame-bait but probably is one.. :-ΓΈ -- Hallvord R. M. Steen
