"Kevin Davies" said in news:
[email protected]:
In emailing a small newsletter mostly within Australia I receive the
following message to a copy which goes to the UK -
Remote host said: 550-RFC2822 says you should have a Message-ID.
550 Most messages without it are spam, so your mail has been rejected.
My understanding of 'Message-ID' is very limited and I can find no
reference to it the properties of any SentMail (although it may well
exist)
Domestic newsletter copies are delivered OK. Current versions Outlook
and Word.
Suggestions ?
KD
The "Message-ID:" is added, if it is inserted by the sender's mail server.
It can be used to track who sent a message from there. For normal e-mail
accounts, it gets added. For anonymous e-mail accounts, it can still be
added because it doesn't specifically reveal the sender. However, spammers
never want it added since: (1) They aren't going to punish themselves by
tracking the message-ID for a reported spam coming from their mail server;
and, (2) They aren't going to bother tracking their spewage, anyway.
If you keep a repository of spam messages, you'll find a high percentage of
them do not include the Message-ID header. When you look at the non-spam
messages, a high percentage of them *do* have a Message-ID header. So it is
a good indicator to further bias a threshold of determining what is and
isn't spam. However, listservers often do not include the Message-ID so you
have to whitelist any newsletters to which you subscribe.
I've also seen some tech support replies that didn't have a Message-ID
header; however, in most of those cases (for me), the "support" was a
contracted provider doing the support rather than the ISP or product maker
themself. My ISP is Comcast and they contract E-Care (it was some other
name just a few months ago) to provide some of their support, especially
when you use their web forms to request help. They don't include a
Message-ID header and so they incur further risk that their messages get
detected as spam. They also would pretend to be Comcast in the content of
their e-mails but Comcast didn't provide a relay through which to send their
messages so E-Care's messages looked like a scammer fraudulently pretending
to be Comcast because they were sending supposedly "Comcast" communications
from a domain other than comcast(.net|.com). If Comcast cannot trust them
enough to provide them with an authenticated relay to send their purported
"Comcast" messages so they come from a Comcast domain then I don't trust
them, either. There have been several tech support replies that I've
received that did not have a Message-ID header because they used an
automated system to issue their reply rather than compose it themself to use
their own mail server.
The absence of the Message-ID header is only a partial indicator that the
message may be spam. I used to have a rule that deleted e-mails without
this header but now I just move the suspect e-mails into the Junk folder
(which has auto-archive configured to permanently delete e-mails older than
2 days). For others, their statistics may vary as to the percentage of
e-mails they receive are spam that also have the Message-ID header missing.
In my repository of retained spam messages, there were something like 2 to 5
e-mails out of 760 (I don't have a big repository of spam) that had the
Message-ID header, so it was missing in the vast majority of the spam that I
have received. Those with larger spam repositories, like thousands if not
tens of thousands of spam, might provide a more accurate analysis of the
percentage of spam in which the Message-ID header does appear.
As mentioned, listservers (for newsletters or e-mail style forums) and some
automated tech support systems (especially those that spew out an automated
response to which you have to reply to get a real response) do not include
the Message-ID header. But obviously such messages wouldn't be in a spam
repository because they aren't spam. You solicited them, so you have to
whitelist them if you use a Message-ID rule.
According to RFC 2822, "Internet Message Format", the Message-ID is an
*optional* header; this header can appear zero or one times within the
message portion sent using the DATA command in SMTP. Most legit (i.e.,
non-spam) e-mails will have it. The domain to where you sent your message
has setup a rule that they will not accept e-mails in which the optional
Message-ID header is missing. So check with your ISP or whatever e-mail
service you use as to why they are omitting this header.
