Now lets pretend for a minute that I'm a parser that doesn't understand
<METADATA>. So I have no idea if <METADATA> is a HEAD tag or a BODY tag.
Perhaps the current spec should be more specific about the treatment of
such a non-conformant document. For example, it could say:
1. An unknown element encountered within HEAD shall be interpreted as an
implicit termination of HEAD and shall be interpreted according
to the content model:
<!ELEMENT unknown - O ANY>
2. An unknown element encountered within BODY shall be treated similarly
except that it shall not terminate BODY.
Alternatively, the DTD could be changed to require an end tag for HEAD in
which case the implied termination of HEAD could be dropped. Such a change,
however, would probably do more harm than good.
Regards,
Glenn Adams