Validator misleads content negotiation; prefers CSS

Hello,

I've noticed a problem with the W3 Validator. When I tell it to
validate a page (not a CSS document), it sends the request as
user agent "Jigsaw/2.2.0 W3C_CSS_Validator_JFouffa/2.0.", with the
following Accept header:

text/css,text/html,text/xml,application/xhtml+xml,application/xml,image/svg+xml,*/*;q=0.

I believe prefering text/css over text/html for HTML page validation is
wrong.

Also, Accept-Language is set to static "en-us,en;q=0.5". This might
cause the W3 Validator to receive different content than the user's
browser, thus verifying a _different page_. I think it'd be desired,
in order to improve the user experience, to replicate the
Accept-Language header which the user's browser sends in W3's requests.

(I'm not subscribed to the list, so if it's not too hard, please Cc me
the response.)

Otherwise, thanks for the invaluable tool.