>It must not become any kind of MIME responsiblity to understand any
>aspect of HTML content. Period! Full Stop!
Which is, as I said long ago, one reason for using a single document
character set in the first place. This provides an extra level of
robustness to UA's and gateways that perform blind translation and/or
mislabel the content. One other point is that we really can't expect
gateways and whatnot to have anything other than MIME parsers in
them (which would be a requirement without a single document character
These are points I made more than a year ago in HTML-WG...