Re: Multiple Content-Location headers

Einar Stefferud (Stef@nma.com)
Wed, 14 Jan 1998 21:23:23 -0800 

I want to strongly endorse Jacob's positions on why HTTP and SMTP
transport of MIME should not require different MIME headers for the
same purposes, and add the point that not all users are going to
always have full IP access to the whole Internet so they can at all
times just reach out and grab any web page that is wished.

This is going to be especially true for archived information which by
definition is a recording of what existed at the time of the archiving
action, and MUST NOT be subject to version changes whether such
changes be intentional or accidental.

I also want to add that WEB technology is analogous to Libraries,
where the User Goes to the Materials, while Mail is orthogonal where
the Materials go from a sender to a receiver by means of transport by
a third party.  If the EMail SENDER will not let go of the sent
object, the EMail Transfer Agent Promises to not DELIVER it to the
Recipient.

I note that both Postal Systems and Libraries have existed for many
centuries, and that neither has yet replaced the other.  Further, I
note that it has been a very important aspect of both libraries and
postal systems that most objects in libraries may easily be
transferred via postal services, and the objects sent through postal
systems can also most often be retrieved from libraries.

Would that Internet Mail and Internet WEB services could work together
so well.  I do not look forward to the day when EMail (the ultimate
Push Technology) is replaced by the need to always go OUT to the
library to fetch my mail.  Using the web for mail reminds me of the
concept of agreeing on which rock in a field, under which to leave
messages for each other.  I call it RockMail.  Primitive at best...

So, back to our objectives.  They is simple: To agree on MIME headers
and their definitions between WEB, and MAIL, and any other related
transport technologies so that we may go forward with the knowledge
that when we create MIME objects, the composer need not know by what
transport they will be moved.  This of course includes the notion of
not having to rewrite or restructure objects because they are about to
be transported via a different transport than the one in which they
arrived.

I also wish to note that this kind of separation to achieve
indepenence between protocol layers is the great value of the IP/TCP.
Almost all protocols above IP do not need to know what underlying
media will be used to move the bits.  And. the more the merrier!

Cheers...\Stef