ietf-smtp
[Top] [All Lists]

Re: Getting the DSN standard used

1998-05-27 14:01:43
As a followup to Keith's message:

The DSN standards (RFC 1891, 1892, 1894) have now been
ready for more than two years. Yet very few mailers support
them. What can be done to get more support?

First a response to Jacob's comment: Last time I went to an
interoperability test, most major vendors of SMTP were supporting DSN.

I'm afraid there are two major obstacles to the widespread use of DSN:

1. the amount of User Agents supporting DSN is low

Yes, but this only explains the inability to request delivery
confirmation.  DSNs are very useful even if many clients don't support
them.

When I wrote this I was thinking about the support in UA's for the 
proper and decent presentation of DSN 'compliant' messages/bodyparts. 
For most 'ordinary' non-technical users the three-part messages are 
still not very helpful in determining what's wrong or what's OK with 
the message they had sent. I think this is a similar problem to the 
problem with the deployment of MIME back in 1992, when MIME 
just was released. User Agents then presented the user with the raw 
MIME structure. Now most UA's present MIME format messages properly. 

Note that there are two levels of support for DSN in MTAs:

1. if they generate nondelivery reports in DSN format
2. if they advertise the DSN SMTP extension, implying that they 
   support NOTIFY, ENVID, etc.

The first is easier to do than the second.  


2. from my experience I know that in a significant number of cases
   the problem why DSN is not supported by a company e-mail infrastructure
   the firewall that's being used in the real world. In many cases the
   firewall has an active mail component, thus limiting the support for
   SMTP and extensions to what the firewall provides. And more often then
   not the firewall does not support DSN or does not even ESMTP at
   all!

Yes, firewalls are a major problem.

We need a clear community statement to the effect that firewalls that
block arbitrary SMTP extensions are nonstandard and broken.

I support your statement here!

/rolf

--

Rolf E. Sonneveld

Sonnection
Schrijnwerkerlaan 16
3828 XC Hoogland
The Netherlands
Tel./Fax. +31 33 480 3968
Mobile:   +31 6543 70 20 5
E-mail: R(_dot_)E(_dot_)Sonneveld(_at_)sonnection(_dot_)nl
Homepage: http://www.sonnection.nl