At 19.21 2/26/95, Harald(_dot_)T(_dot_)Alvestrand(_at_)uninett(_dot_)no wrote:
you are very close to what I think of doing for UNINETT too:
Defining a service (?) that is able to convert all 8-bit mails
into MIME using EMIL, and saying that a site may become compliant
with the UNINETT recommendations by defining this site as its "smarthost".
This is needed, we all agree about this. What I want is to try to define where
this conversion is going to take place. This is definitely one of the
services a network provider can sell to a customer, but we also
have to try to make a MS-Mail user aware of the problem and
try to convince him that by running the conversion on his
host, he can have full control of the conversion process
(and he must run a UNIX host there anyway because he might need
to rewrite the addresses the MS-Mail gateway is using).
It would also be possible to define a service converting Q-P to
8bit, but I am more doubtful about offering this.
I agree.
I think we are in agreement that 8-bit untagged is an evil, but we
need to define how to deal with it.
If I say to someone "uninstall Microsoft Mail", they will not listen
to me.
This is correct, they will not listen, but if you say first,
if you pay me a million bucks, I can run this for you, and
for $1000 I can install it on your "you-will-need-it-anyway"
_own_ UNIX relay mailer.
If I say to the same people "pay me USD 5000 a year, and do this", they
MAY listen. That is what I wanted to see in Julian's document, which,
unlike the SMTP-AS document, seems to be a guide to how to survive in
practice, and not a preaching of the official IETF dogma.
I would like his document to, in some way, say that 8bit is the wrong
way of going. Explain what actually happens when this strategy is
used, and also explain how easy it is to install for example
Emil on the Relay Mailer. Either run it yourself or buy this
service (like buying NNTP, secondary MX, XNTP) from the network
provider.
What do you think about writing a paper as "rules" for being a
network provider?
Patrik