On 15/10/07 9:30 AM, "James M Snell" <[EMAIL PROTECTED]> wrote:
> Sam's position (which he can correct if I have misstated anything) is > that the validator should take a strict approach and mark all unknown, > non-namespaced attributes as an error unless there is a known > specification that is specifically updating RFC4287, as is the case with > the Bidi draft. It's not an error, but that doesn't mean that the validator couldn't issue a warning that the attribute is being ignored as foreign markup. e.
