Comments on: DKIM: what it's not

By: laura

laura — Fri, 09 Oct 2009 19:54:37 +0000

Steve,

You can get some instructions on http://DKIMcore.org/. Step by step depends on what MTA software you are using, so you’ll want to talk to your vendor or look at the opensource tools available for your particular mail software.

By: steve urow

steve urow — Fri, 09 Oct 2009 15:52:52 +0000

I send out a newsletter and want to implement DKIM, but I do not see the nuts and bolts, step by step of doing so.

By: laura

laura — Wed, 30 Sep 2009 23:50:04 +0000

I have been told, in no uncertain terms, that s= should not be used for reputation and if it is that Internet Greybeards will show up on your doorstep and beat you with standards documents.

By: J.D.

J.D. — Tue, 29 Sep 2009 19:56:24 +0000

It’s unlikely that s= will have any impact on deliverability at all.

By: Nazzareno Gorni

Nazzareno Gorni — Mon, 28 Sep 2009 14:45:05 +0000

I agree with Laura and I have a question.
Which is the impact of the DKIM selector parameter (s=)?

By: Jim Fenton

Jim Fenton — Sat, 26 Sep 2009 17:56:15 +0000

I have a slightly different take on these:

1. It’s the other part of the ESP’s statement, “…and will benefit from the positive reputation they have,” that I disagree with. With domain-based reputation it’s the domain, not the mail server, that has a reputation, as you point out.

2. The presence of a DKIM signature does allow a domain to start establishing a (hopefully positive) reputation with the ISP, which in turn can result in increased delivery to the inbox without further action on the domain’s part. It’s true that it’s not a direct consequence of signing though.

3. I completely agree with you on this; there’s no way that I can see any ISP requiring signatures, except possibly if the author domain publishes an ADSP record requesting this.