I am writing here as MHTML WG Chair...

The Montreal MHTML WG Meeting did not decide to keep the text from
Jacob's draft!  It decided that the text must not confuse people as to
what is going on, and thus needs to be fixed.

The discussion in this list will now decide the text to be included,
and that includes tossing the original text if we decicde to do so
here.

Please note that it is the mailing list consensus that MUST govern!
Not what a subset of the WG decided in a 2.5 hour face to face meeting
where we were under heavy pressure to close on a list of 28 agenda
items, which we did, pending further resolution editing of the text to
reflect the sense of the meeting.

The issues of line-feeds and charsets/encodings were left as editing
tasks to be completed as we go into a WG last call of approx 4 weeks
duration.  From our discussion here, it appears to me that we are
rapidly closing on the view that --

   "Whatever HTML does within the specifications for composing HTML,
    MIME must convey it safely and without distortion to the
    recipient, whose UA environment must be capable of handling the
    raw HTML content, after MIME transfer-decodes it and hands it over
    to an HTML rendering engine."

It must not become any kind of MIME responsiblity to understand any
aspect of HTML content.  Period!  Full Stop!

Cheers...\Stef


From Martin J Duerst's  message Fri, 5 Jul 1996 11:22:43 +0200 (MET DST):
}
}>My objection is based simply on the fact that I cannot see why we need
}>to even mention numeric character references or entity references, and
}>*especially* numeric character references. These are entirely separate
}>from the issue of tranfer encodings.
}
}I agree that there are lots of things in this area that we don't have
}to mention. My proposal (sent a few days ago) for rewriting what is
}currently in the draft is just to get the stuff correct in case it gets
}included (as it seems to have been decided at the Montreal meeting).
}
}Regards,        Martin.