On Wed, Apr 21, 2004 at 03:32:38PM -0600,
administrator(_at_)yellowhead(_dot_)com wrote:
| It is wrong to assume that broadband lines are home users. We service our
| customers using 2 X 4 Mbps ADSL circuits operating in a balanced mode. It
| is 1/4 the cost of T1/T3 and just as reliable.
Do those customers control their own PTR?
If so, no problem.
Even if they do not control their own PTR, if they can put in SPF
records, again, no problem.
That is the goal of the algorithm.
Besides, suppose you take a human going through a mailbox which is 95%
full of spam. Suppose the human accidentally deletes a non-spam message
with probability p1. I take this probability as the baseline for
measuring antispam methods.
Any antispam technique that has a lower false positive rate than p1 is,
in my mind, acceptable. The "looks like broadband" rule is very
acceptable.
The three-step algorithm above is even more acceptable because it gives
the sender a way to exert more control --- by doing SPF first.