On 2012-10-20, Jeroen Geilman <jer...@adaptr.nl> wrote: > > DNSBLs are recommended by just about everyone who is serious about > email,
There are a couple ways to use DNSBLs. There are those who are "serious" but either incompetent or on a cost-saving agenda, and then there are those who are "serious", and have enough budget to use DNSBLs competently. The incompetent use of DNSBLs: This group uses DNSBLs as originally intended - to *block* connections. This reckless approach (in effect) guarantees denial of service on the sole basis of IP address, neglecting more effective criteria. Not only is the judgement as to whether the message is spam or ham cheapened, it violates EFF principles. The competent use of DNSBLs: This group uses DNSBLs not to block, but rather to aggregate DNSBL tests with other more effective characteristics of the message. Instead of foolishly allocating all weight of the message treatment to one single factor, the finding is appropriately weighted with other factors. And even if the message is judged to be spam through a more careful process, /it is still delivered/, and rightly so. Keep in mind that the OP called for the "ultimate" mail server, not the cheapest one. To me this implies that quality *trumps* revenue and cost-savings (as opposed to being one of many profit-driven factors). > and a proper EHLO is actually an RFC requirement. You should read the requirement. The RFC certainly does not insist that senders buy a domain name. The RFC allows for senders who do not own a domain name to supply their literal address (aka IP address) for the EHLO. Such a message is RFC compliant, but blocked by those who are uninformed about this and implement reject_non_fqdn_helo_hostname. It is indeed a common misconception that the RFC requires a hostname for the EHLO.
