Re: Spammers Want You To Believe That (Challenges from challenge-response systems qualify as unsolic
am 16.05.2005 08:53:05 von bonomiIn article
Alan Connor
>On 31 Jul 2004 13:50:18 GMT, Frank Slootweg
>>
>
>
>
>
>> Or some idiot decided that (starting) to use a C/R system is a smart
>> thing to do. Nah, you're right, can't happen!
>
>Funny. "Idiots" that use C/R systems are the only people on the planet
>that don't have a spam problem.
Alan, is, as usual, "talking nonsense" on
One point of disproof:
The mailserver I use (nor do I personally) doesn't use a C/R system.
The mailserver I use (nor do I personally) doesn't use 'greylisting'.
The mailserver I use (nor do I personally) doesn't use DNSBLs.
I'm the only one on the server who posts to USENET,
and I don't use a munged (or otherwise invalid) email address.
[Surprisingly, Alan doesn't trust his own systems enough to post
with his _real_ address. ]
So far, this year, there have been a grand total of _three_ pieces of junk
email make it to _any_ in-box (half a dozen users, with several dozen mail-
boxes) on the mail-server. Each of those three messages resulted in the
mailadmin spending about 40 seconds updating mail-filtering definitions.
In the same period, the system has rejected a grand total _two_ messages
that were legitimate. (One managed to trigger a previously undetected bug
in the filter-code -- there was a clock-based problem that could only be
triggered during a 5 hour window that occurs only once a month. The other
was a user in the midst of testing their outgoing mail configuration, and
selected a subject line that happened to match one that an older virus uses.)
A grand total of 5 errors out of more than 10,000 spam delivery attempts.
That's a "gross" error rate of under 0.05%
On average, *system*wide*, about 1 piece of spam gets delivered every 6 weeks.
On average, a _person_ getting mail here sees about 1 piece of spam every
9 months,
The "false-positive" rate is anomalous. 2 failures system-wide in the last
45 days; the same two failures in the last 400+ days. With the detected
bug fixed, users can expect to lose about one piece of valid mail every
_five_years_.
Users see:
one piece of spam every 9 months,
one lost message every 5 years.
"Not too shabby," I'd say.
*BUT*, this system does *not* use C/R, so, according to Alan Connor all the
users of this system _must_ have a spam problem.
>I think it is pretty obvious who the real idiots (or malicious liars)
>are.
H*ll, I might have to actually _agree_ with him about *that*!