Currently, the server doesn't need to parse the document at all.
This would be the first such case. (systems with server-side includes already
do some parsing, though).
What should it do for a document without a HEAD?
Having sounded so negative.. :-)... I support this suggestion.
Lee