Marc Perkel wrote:

> So - who wants to try feeding the system data? That helps make the 
> system more accurate for everyone. To feed me data you need to run a 
> small bash script that I will provide you all it does is send a string 
> to a port on my server.
> 
> To report a spam you run ip-report spam 1.2.3.4
> 
> To report a nonspam you run: ip-report ham 5.6.7.8
> 
> The idea being that you report only those that you are highly confident 
> in. Since the system runs on totals and percentages it can take some 
> error and still be accurate. I am especially interested in the nonspam 
> reports as this helps keep the blacklists accurate.
> 

Marc,

In case you have been reading only selectively, several folks have responded 
with rather less-than impressive results in first-round tests.

The early indicators are that best-case, your lists duplicate well-known acl 
tests or other established and trusted RBL.

Worst case, they return irrelevant data, or support actions not beneficial.

JMNSHO, but I see that as a conceptual or 'scoring' issue, not a lack of 
sufficient data.

Missing a white/yellow/black entry is 'invisible' to some parts of a testing 
phase.

Serving up entries that do not add value - or worse- increase falsing, won't be 
alleviated by more entries.

It is better *accuracy* and a reduction in duplication of other RBL in the 
entries you DO have that is needed.

Otherwise, one would be ahead to concatenate several of the present top-scoring 
RBL's and be done with it.

As is common already in Exim acl's

Time to pull back and do some basic work.

Bill



-- 
## List details at http://www.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://www.exim.org/eximwiki/

Reply via email to