On Don, 2010-02-11 at 18:26 +0000, Mike Cardwell wrote:
[...]
> I want you to describe a scenario where the sender or recipient are
> actually worse off because of the particular two features I've
The point is: The sender is worse off because he needs to invest time
for the workaround which is caused by the receiver. The receiver is
worse off because some senders plain simply give up when they are
expected to pass a Turing test.
No, I don't have numbers. But I'm pretty sure I'm not the only one.
> described. You've failed to even attempt that so far.
You see only two alternatives:
- keep these two features (and tell the senders that they should
actually be happy that they can invest time and effort to work around
FPs caused by the receiver spam).
- or deactivate it.
I proposed the 3rd solution:
- repair your spam-detection (change weight/limits, use Bayes,
greylistung, etc.) to not generate so many FPs that you actually need
an additional workaround.
That would actually remove the cause and not fiddle with the symptoms.
> I know this system works well because I've been using it for a long time.
To use your own words: Ridiculous. Just because someone uses something
for a long time doesn't make it good (or is even an indication for it).
Apart from that: You very probably don't know how many FPs didn't come
through because of people disliking Turing tests.
Bernd
--
Bernd Petrovitsch Email : be...@petrovitsch.priv.at
LUGA : http://www.luga.at
