Re: [spfbis] Proposed spf TXT record change

"John Levine" <> Thu, 11 February 2016 15:57 UTC

Return-Path: <>
Received: from localhost ( []) by (Postfix) with ESMTP id 4D4811B33CB for <>; Thu, 11 Feb 2016 07:57:10 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: 1.664
X-Spam-Level: *
X-Spam-Status: No, score=1.664 tagged_above=-999 required=5 tests=[BAYES_50=0.8, HELO_MISMATCH_COM=0.553, HOST_MISMATCH_NET=0.311, KHOP_DYNAMIC=0.001, SPF_PASS=-0.001] autolearn=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id NjAMNSGRQOAC for <>; Thu, 11 Feb 2016 07:57:09 -0800 (PST)
Received: from ( [IPv6:2001:470:1f06:1126::2]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 50C921B33DC for <>; Thu, 11 Feb 2016 07:57:08 -0800 (PST)
Received: (qmail 82833 invoked from network); 11 Feb 2016 15:57:02 -0000
Received: from unknown ( by with QMQP; 11 Feb 2016 15:57:02 -0000
Date: 11 Feb 2016 15:56:40 -0000
Message-ID: <20160211155640.5340.qmail@ary.lan>
From: "John Levine" <>
In-Reply-To: <>
X-Headerized: yes
Mime-Version: 1.0
Content-type: text/plain; charset=utf-8
Content-transfer-encoding: 8bit
Archived-At: <>
Subject: Re: [spfbis] Proposed spf TXT record change
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: SPFbis discussion list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 11 Feb 2016 15:57:10 -0000

>I am mystified that the bitterness on this topic remains impervious to
>things like evidence.
>The history of the SPF record type is documented in RFC6686.   I'm happy to
>let that document and the data it contains speak for itself.

There is a belief in parts of the DNS community that adding new RRTYPEs
is trivially easy.  Provisioning system problems don't matter either
because they don't exist, or because any zone file of any importance
is created with vi.

A decade ago when 4408 was published, the belief was pervasive enough
that its adherents demanded that the SPF record be added at the last
minute or the draft wouldn't be published, and it was added so quickly
that the language was botched and describes implementations that don't

Since all existing SPF implementations already used TXT records, and
type 99 SPF records involved considerable pain and no operational
benefit (I know about having the server automatically add the SPF
records, my provisioning system did it), it's not surprising that
nobody used them.

And since that experience showed a treasured belief to be wrong, well,
the reactions are not surprising.