I guess it all depends on what you consider a 'significant problem'.
The metrics used to make this determination will vary from business to
business. For the several sites that I run here (small in comparison),
any false positives represent a potential loss of business or a hit on
business and we do everything we can to avoid. Other mechanisms used in
tandem seem to knock the spam levels down to a level, at least for us,
that are acceptable without the uncertainties associated with RBL's.
Best Regards,
-- Tim
On 5/11/2011 11:19 PM, Dave CROCKER wrote:
On 5/11/2011 6:51 AM, John Levine wrote:
The problem I have with this argument is that blackhole lists, in my
experience, cause a large number of legitimate messages to fail to be
delivered.
Having gone to a lot of MAAWG meetings, and talked to people who run
the mail systems at every large ISP in the country, and quite a few in
other countries, I can report that their experience with DNSBLs is
utterly unlike yours. There are plenty of incompetently run DNSBLs,
but nobody uses them so they don't matter.
Badly-run blacklists do cause problems. But, then, few people choose
to use bad lists.
Blacklists are an integral part of real-world email operations today.
When thinking about the scale and number of the major email service
providers, it is not reasonable to suggest that they would continue
that use were there a significant problem with false positives.
d/