From: asg@severn.wash.inmet.com (Al Gilman)
Message-Id: <9511271614.AA15306@severn.wash.inmet.com>
Subject: Re: mid and cid URLs
To: masinter@parc.xerox.com (Larry Masinter)
Date: Mon, 27 Nov 1995 11:14:39 -0500 (EST)
In-Reply-To: <95Nov26.094022pst.2733@golden.parc.xerox.com> from "Larry Masinter" at Nov 26, 95 09:40:17 am
To follow up on what Larry Masinter said ...
I've re-reviewed all of the relevant documentation (scanned all RFCs
for the header "content-ID") and have come to the conclusion that in
all circumstances, 'content-ID' is required to be globally unique.
That is, 'cid:baz@host' is always unambiguous.
'mid:frob@host/cid:baz@host' == 'cid:baz@host'.
In such circumstances, they wouldn't need to cite a relative URI at
all.
I absolutely agree that they don't _need_ to. I really think
that they should _want_ to. The Content-ID value is
World-unique. It is a sufficient key. That doesn't mean that
the recipient doesn't _want_ to know that it happens to be
located in the current message. The recipient _does_ want to
know that.
I quoted you on the matter of the CID being a sufficient key to
Harald, just now, on the ietf-types list.
<double-negative warning>
I still think that we should not strictly limit our syntax to
carrying _only_ the information that is absolutely necessary. I
think that we can and should define a syntax which emphasizes the
necessary information, but makes room for useful additional
information.
Compare this with my discussion with Harald about the
advisability of annotation an MID URI for a FAQ with
"Newsgroups:" and "[archive-]Location:" info.
Al