RE: draft-gellens-on-demand-04.txt

[Randall Gellens <randy@xxxxxxxxxxxx>]

At 2:52 PM -0700 6/26/98, Jeff Stephenson (Exchange) wrote:

>We still seem to disagree on the use of a separate port for this command.
>While doing this on port 25 might be architecturally difficult for some
>implementations, it's actually much more straighforward for us than using a
>separate port.
>
>I propose a compromise in which both ports 25 and 366 are allowed, with the
>provision that support for ATRN on port 25 MUST only be offered if the host
>supports a full SMTP service on that port; an ATRN service on port 366 MAY
>be the restricted profile of SMTP that you describe.

For interoperability, I think it is best to stick with 366 in the
document.  Using a separate port doesn't preclude offering ODMR
integrated with SMTP; the server can listen to both ports.  If a
provider wanted to offer ODMR only on port 25, as long as the customer
could configure its end to also use 25 (which shouldn't be a problem)
then it's no business of anyone else.

>Also, I think that the fourth paragraph of 3.2.1 could use a bit of
>clarification.  Say I issue the command
>
>	ATRN foo.com, bar.org
>
>but I've only got rights to access foo.com.  My understanding of this
>paragraph is that the service is to issue a 450 response and send no mail,
>even if there is mail queued for foo.com.  However it seems to me that the
>paragraph could be interpreted as saying that you don't send mail for
>bar.org, but that sending it for foo.com is allowed.  I'd strengthen the
>paragraph to say "... the provider MUST NOT send mail for _any_ domains to
>the customer...".

I agree; thanks for pointing this out.
--
Randall Gellens
Opinions are personal;    facts are suspect;    I speak for myself only
-------------- Randomly-selected tag: ---------------
Drunk Gets Nine Months in Violin Case
--Newspaper headline