On 28 Jan 2012, at 9:45, Reindl Harald wrote:
Am 28.01.2012 07:00, schrieb Bill Cole:
to say it not polite: it is idiotic to remove "User unknown"
That's debatable. However, it is explicitly allowed by RFC5321 and
its ancestors.
not all things which are allowed are smart
the RFC allows you even a open-relay
so should we help anybody setup one?
it is only debatable because you have seen not enough of
the real world and thinking only the RFC is saving your
life and even if you make a big mistake it is OK as long
it is in the RFC
most spam-filters are not there but reality
Clearly we have a language barrier that probably won't be overcome. Even
with almost a decade working for US divisions of Daimler, VW, and DT my
German is worse than your English and it is obvious that I failed to
communicate my point. Maybe I can make it more simply:
I have used mail log analysis tools of my own designs for 6 different
MTA's over the past 16 years for mail environments as active as millions
of SMTP sessions per day. In my real world experience I have learned
directly that there's no marginal value to be gained from programmatic
interpretation or analysis of the text part of SMTP replies.
You should also be aware of the fact that RFC's are not written in
isolation from the real world. The revisions from 821 to 2821 to 5321
were distilled from the real world experiences of people running mail
servers, and it is instructive to look at what changed between them and
what did not.
See http://tools.ietf.org/html/rfc5321#section-4.2
as example we are doing fully automated bounce-managment based on
/var/log/maillog and remove adresses from ALL tables of all
customers
which contains "newsletter" in the table-name and having a field
"email"
If you relied on the text in that way in an SMTP client
it would violate a "MUST" statement in RFC5321.
wtf - who said that i rely only on the text?
No one said that, implied that, or inferred that. Please find someone
who is adequately fluent in both English and German to explain the
admonitions in the middle of page 47 of RFC5321 in light of the special
definitions of RFC2119. I think that is beyond my skills.
As this thread has moved away from topicality here and is burdened by
linguistic difficulty, I am done with it.
