ietf-822
[Top] [All Lists]

Re: 7/8-bit conversion vs. bouncing

1991-07-01 16:53:52
On Jul 1, 11:50pm, Keld J|rn Simonsen wrote:
Subject: Re: 7/8-bit conversion vs. bouncing


% Are somebody really insisting on bouncing 8-bit mails, which are intended
% for 7-bit land? 

  I think that there is a consensus that 8-bit SMTP data should be
bounced rather than converted when confronted with a receiving host
that can only handle 7-bit SMTP data.  All of the 8-bit content can be
sent via 7-bit transport and 8-bit transport should only be used
within appropriate local "enclaves".  

% This makes me think that the whole thing is unusable.

  Not at all.  8-bit transport is for local enclaves and use outside
such enclaves should not be expected to work.  All systems should rely
on EVFY to ensure that 8-bit transport is supported before sending any
8-bit data.  I fully support bouncing 8-bit SMTP data that cannot be
transmitted to a host because the latter host only supports 7-bit SMTP
data.  If one wants it to go through to all hosts, one should use 7-bit
transport.  Why the hangup with the 8-bit transport mechanism ?

  I have no confidence that 8-7 bit data conversions will be done
without mangling data in most cases.  This is based on experience with
the current Internet mail system and self-appointed pseudo-gateways.
        
% The IETF meeting in Copenhagen concluded that we should do 8<->7 bit
% conversion and this was not considered a transformation, as long as all
% the information was preserved.

  I wasn't there and would have argued against that interpretation 
had I been there.  How about other folks ?

  I think that the draft RFC might need to make itself unambiguously
clear about what is required of conformant systems in this regard.
There seems to be some confusion now as to what it says -- or is
it disagreement rather than confusion ?

Ran Atkinson
randall(_at_)Virginia(_dot_)EDU