W3C home > Mailing lists > Public > xml-dist-app@w3.org > March 2002

Re: Need new MEP for SMTP binding

From: Jacek Kopecky <jacek@systinet.com>
Date: Thu, 14 Mar 2002 19:04:15 +0100 (CET)
To: Mark Baker <distobj@acm.org>
cc: xml-dist-app@w3.org
Message-ID: <Pine.LNX.4.44.0203141853560.26446-100000@mail.idoox.com>
Mark,

 > >  how do you accomplish acknowledgement in email?
 >
 > Stuart explained it well.  Basically the response you get to your
 > attempt to send email to a SMTP server is a hop-by-hop acknowledgement.
 > A "success" response doesn't mean "the recipient received your message",
 > it just means "this hop accepted your message for forwarding".

My point is that this is not an 'ack' that would be useful for 
the sender node. An ACK is an acknowledgement of receipt, while 
here it's only saying "I haven't failed on the first stop." 
Therefore the sending node cannot be certain if the receiver got 
the message - the whole idea behind acks - and we cannot call 
this a one-way-with-acks MEP.

 > > As far as I know
 > > all the standardized ways are optional and usually unimplemented
 > > or even ignored for security reasons.
 > >  If you mean hop-by-hop at the transport level (transport
 > > intermediaries),
 > 
 > s/transport/transfer 8-)

I'm afraid I don't see a reason for this change. Might be a 
subtle difference between the two words that I don't see here.

 > > I think in case of email where you logically
 > > never have a single hop transfer, this hop-by-hop ack is useless
 > > to the sending SOAP node.
 > 
 > I don't follow.

See above. I understand that an SMTP error like "message
rejected" is useful, but it's useful in the same way as the
operating system's error "cannot open socket". The availability
of the first-hop status does not constitute an ack as in
"one-way-with-acks MEP".

 > > Oh, I forgot to add that I'd in fact like to see a one-way MEP, 
 > > but without the ACKs.
 >
 > I was considering mentioning this.  It's really the degenerate MEP,
 > because it's the pattern that the envelope uses without a binding.
 > I agree that giving it a URI would be a good thing though.

 I wouldn't use the word degenerate, I think the common name is 
fire-and-forget. As UNIT_DATA in the abstract model, this is the 
building block, the basis, not a degeneration of something more 
complex.

 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)
                   http://www.systinet.com/
Received on Thursday, 14 March 2002 13:04:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:09 GMT