> How do you plan on doing this in HTML+? Sounds like the first part
> will be text/html, where the anchors reference other parts. The reference
> can be done by defining a URL access type of "part", e.g. HREF=part:XXX
> where XXX is a Content-id.
I think that the HREFs should use the standard URL for the document, and the
MIME subheader for the image (or whatever) should include a field specifying
the URL for the subpart, e.g.
Source: http://info.cern.ch/hyptext/WWW/lep.jpg
The current spec for the HTTP doesn't give a suitable field, although
Message-id is a near miss (defined according to RFC 850). I don't like the
idea of HREF=part:XXX etc. as this means you couldn't send the main document
by itself.
I would also like a try out simple approach to embed foreign data directly,
as this would much more convenient for inline equations etc.
Regards,
Dave Raggett
