In message <9402211509(_dot_)AA01936(_at_)ptpc00(_dot_)cern(_dot_)ch>, Tim
Berners-Lee writes:
Dan,
1. We make the following correspondence between the terms of the SGML
standard and the MIME RFC:
...
SGML SYSTEM identifier => MIME Content-ID
I proposed in an earlier note that the SGML SYSTEM identifier
should be a URI and that a special form of URI (cid:) should
be used to specify a content identifier within a MIME
multipart message.
SGM SYSTEM identifier => URI
This allows SGML to reference anything in the universal
syntax, inclusing now URLs and later URNs, and
other parts of a MIME object using the cid: form.
I think this is important, as external references
are important too, and one will need to mix them.
So far we've only been talking about MIME and SGML. Neither MIME
tools nor SGML tools currently grok URI's. I see no reason to
introduce that much new technology just to allow a MIME message to
hold multiple SGML entities.
Besides: MIME already specifies a way to reference bodies outside
the message: message/external-body. e.g.:
Content-Type: multipart/x-sgml; boundary="cut-here";
document=id1
--cut-here
Content-ID: id1
Content-Type: text/x-sgml
<!DOCTYPE HTML [
<!ENTITY % htmldtd SYSTEM "id2">
&htmldtd;
]>
<html>...</html>
--cut-here
Content-ID: id2
Content-Type: message/external-body;
access-type="anon-ftp";
site="info.cern.ch";
directory="pub/www/doc";
name="HTML.dtd"
Content-Type: text/x-sgml;
--cut-here--
Dan