Errors-to: owner-drums@CS.UTK.EDU
Message-Id: <199505312026.QAA21750@CS.UTK.EDU>
Date:        Wed, 31 May 1995 22:16:38 +0200
From: Eric Thomas <ERIC@SEARN.SUNET.SE>
Subject:     Re: getting started
To: "Perry E. Metzger" <perry@imsi.com>
cc: drums@CS.UTK.EDU
In-Reply-To: Message of Wed, 31 May 1995 16:14:03 -0400 from "Perry E. Metzger"
             <perry@imsi.com>

On Wed,  31 May 1995  16:14:03 -0400 "Perry E.  Metzger" <perry@imsi.com>
said:

>The internal systems do not need to have all the properties. The gateway
>*does*.

Lots of good  it does when it  can't be implemented because  of a missing
property in the target mail system.

>> Well, guess what happens when you try to gateway this to the Internet?
>> The gateway  has to  put *something*  in the  one-size-fits-all origin
>> field.
>
>The problem is not in outgoing mail from these pathetic gateways.

Read  again. I  was talking  about what  the gateway  has to  put in  its
internal one-size-fits-all field. This suggests  it is creating a message
in the target network,  not in the Internet, ie that  the problem is with
mail that  is *incoming* to  the gateway.  Outgoing mail is  easy because
Internet mail  supports all  the necessary  functionality to  comply with
Internet standards. In general when you gateway  from A to B, some of the
functionality  in network  A  is likely  to  be lost  because  it is  not
available in B. Internet users lose  their bounce behaviour and LAN users
lose their guaranteed positive delivery/read/discard acknowledgement, and
so forth.

>It  is the  fact  that  they routinely  bounce  to  header addresses  on
>incoming mail,  or send  you helpful  messages, or  fail to  relay error
>messages with <> as  the origin address, or fail to  accept mail with <>
>as an origin address, or ....

If they fail  to accept mail that  reads MAIL FROM:<> they  are buggy and
that's a totally  different issue. If they fail to  bounce where you want
them  to bounce  it's usually  because they  can't do  it. Read  my other
messages.

  Eric