The bogus doctype state does not work as a sufficient spec for
parsers that need to extract the kind of information that is needed
for browser-compatible doctype sniffing.
I suggest amending the bogus doctype state with substates that match
dbaron's doctype sniffing algorithm from Gecko, because in addition
to Gecko, WebKit uses that algorithm and Opera 9 uses something that
is mostly equivalent (subject to corner case differences induced
apparently by blackbox reverse engineering).
--
Henri Sivonen
[EMAIL PROTECTED]
http://hsivonen.iki.fi/
