>I know I'm not making any friends here
<G> Debating is OK, just so long as everyone realizes that different
people have different opinions, and there is no one "right" answer.
> - but you would say that you're selling a product that filters mail.
Yes.
>Out of interest what is your source for those stats.
The 18 billions spams sent per year comes from a variety of sources (most
estimates seem to be somewhere around 10-20 billion; a GartnerGroup survey
from a year or two ago claimed 15 billion), but I don't have the source for
that stat handy. The 4 seconds to delete an E-mail is just a rough
estimate (it may be quicker for obvious spams, but could be quite a bit
longer for others). The 6,000 people working full time to delete all the
spam is just a calculation (18 billion E-mails times 4 seconds each is
800,000+ hours).
We've got more stats at http://www.declude.com/JunkMail/pros.htm that
covers some pros and cons of spam scanning. Of course, it's geared towards
the pros, but includes some cons and our thoughts about them.
>Don't get me wrong I think spamming of all kinds is 100% wrong. I just
>believe that blocking email is wrong.
And you are entitled to your opinion (you aren't the only one who thinks
blocking mail is wrong). That's actually the main reason why we offer
per-user configurations in our anti-spam software.
>The Question is: Without being the recipient, how do you know something is
>spam unless you are the recipient?
>
>Can anyone actually pin down a common denominator?
Spam detection is a very tricky art. It is impossible to have 100%
accuracy unless you know what the recipient wants.
One of the methods that we use is a header analysis (the BADHEADERS test,
for those who are familiar with Declude) that looks at the headers of spam
for broken headers. A lot of spam (perhaps 50% or so) has
non-RFC-compliant headers, typically because the spamware developers aren't
very familiar with the RFCs, and a lot of spamware allows the spammer to
enter in their own made-up headers in an attempt to obscure their
identity. The test will *only* catch spam and other illegitimate
E-mail. Of course, even this test will have false positives (a few ASP
components that send E-mail, for example, send out an illegal "Date:"
header; at least one foreign version of Eudora translated the header name
"Date:" to "Fecha:", so it would not include the required "Date:" header).
Another good test is our SPAMROUTING test, that will detect E-mail routes
that are very poorly chosen (possibly against RFCs) for normal E-mail
delivery. For example, if someone in the U.S. sends an E-mail to another
U.S. destination, but sends it through an overseas open relay, it is almost
certainly spam. Of course, on rare occasions, there will be exceptions,
and you can't have 100% accuracy. This, along with the header analysis,
can catch about 80%+ of spam.
One option is to filter E-mail aggressively, but not delete the
spam. Instead, move it to a separate folder (most E-mail clients can do
this), and check the folder once a month or so to make sure nothing you
wanted came in. That way, you can have great spam control, while also
getting all your the mail.
-Scott
Declude: Anti-virus, Anti-spam and Anti-hijacking solutions for
IMail. http://www.declude.com
Please visit http://www.ipswitch.com/support/mailing-lists.html
to be removed from this list.
An Archive of this list is available at:
http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
