Re: Every spam is sacred, back the first message
Mathieu Roy wrote:
> > In short: This is not the big problem you are trying to make.
> I'm not trying to make a problem but I questionned your proposal. Now
> you tell me that DNSBL's are ok at 99%.
No, I didn't say that.
What I said is that there are good DNSBLs and bad DNSBLs.
The DSBL (http://dsbl.org) is just one of them, one particularly good
in terms of high number of messages rejected and low false positive rate,
but I never said that all DNSBLs are as good as this one (see my
comment about SPEWS, for example).
You continue to put all DNSBLs in the same bag. You should not judge
DNSBLs in general, you should judge each of them separately.
> Maybe you should have started with this statistic at first. And not
> with the following question:
> "They have said "no" using (more or less) the following
> reasoning: Since Debian machines have been listed in several
> DNSBLs in the past, we should not use ANY of them ourselves
> (which is like saying: since we have sent tons of spam in the
> past via our mailing lists, we should accept ALL the spam we
> receive). Does somebody understand this?"
> I personally understand this. It means that they distrust DNSBLs
> because they suffered in the past or their errors - and know what
> troubles DNSBL errors may create.
That is putting all DNSBLs in the same bag, which I consider a
BTW: We are using a DNSBL ourselves, it's called rbl.debian.net.
See master's exim.conf.
If you are so much worried about false positives and such, you should
probably ask debian-admin about the listing and delisting criteria for
> If you are right when you said "the probability that a message sent
> from an IP in the DSBL is spam is about 99.95%", it's an argument.
> On 2000 messages (I receive about 2000 mails per month personally),
> it's potentially 1 mail inappropriately blocked or tagged. Which is
> not horrible.
> I would like also to remind everyone on the list that the first mail
> from Santiago was clearly talking also about blocking mails in the
> end, not only tagging.
Yes, I was talking about both things. About tagging and about
eventually blocking. So what? Can't one propose several different
things at the same time?
Each proposal should be considered by their own merits.