Re: [apps-discuss] "X-" revisited

Mark Nottingham <> Tue, 12 July 2011 23:54 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 7BC1E11E80AF for <>; Tue, 12 Jul 2011 16:54:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -105.927
X-Spam-Status: No, score=-105.927 tagged_above=-999 required=5 tests=[AWL=-3.328, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id 8Ppo+It9JUhX for <>; Tue, 12 Jul 2011 16:53:59 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id 323F411E80AE for <>; Tue, 12 Jul 2011 16:53:59 -0700 (PDT)
Received: from (unknown []) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPSA id 959D2509EF; Tue, 12 Jul 2011 19:53:51 -0400 (EDT)
Mime-Version: 1.0 (Apple Message framework v1084)
Content-Type: text/plain; charset=us-ascii
From: Mark Nottingham <>
In-Reply-To: <>
Date: Wed, 13 Jul 2011 09:53:48 +1000
Content-Transfer-Encoding: quoted-printable
Message-Id: <>
References: <> <> <60B61428BA7C86CBCF4D236D@PST.JCK.COM> <> <>
To: Larry Masinter <>
X-Mailer: Apple Mail (2.1084)
Cc: John C Klensin <>, "" <>
Subject: Re: [apps-discuss] "X-" revisited
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: General discussion of application-layer protocols <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Tue, 12 Jul 2011 23:54:03 -0000

On 12/07/2011, at 11:26 AM, Larry Masinter wrote:

> Don't the considerations around "x-" also apply to "vnd." in media types?
> vendor-specific media types become standards. standard media types get vendor-specific extensions.

It's a somewhat different path, I think. A vendor who uses a vnd and then later standardises the format has demonstrated a willingness to do the right thing, and the incremental effort for them is small.

Contrast that with the common case for X-; IME it's usually started by someone who has an incomplete understanding of the standards process, and doesn't want to put much effort in. X- has become a sort of cargo cult whereby it's OK to do anything if you just put X- in front of it, so that's what people do -- even for things that end up getting widely used. 

We basically need a leaflet that we can drop on that island of cargo culters, to gently clue them in. Except that the island is really the rest of the world. But you get the idea.


P.S. A vnd that later gets standardised is likely to have differences between the versions, so a different identifier is necessary anyway.

> The additional "information" that we imagine is associated with "x-" or "vnd." is 
> * not testable
> * can evolve over time with no transition method, plan, policy in place
> * has no cost associated with mis-use  (calling something x- or vnd. when it isn't, or not using x- or vnd. when proscribed).
> In the general area of registry values, any lexical convention for registry values must have some actual implementation consequences for most of the community to take the restrictions seriously. 
> Larry
> --
> _______________________________________________
> apps-discuss mailing list

Mark Nottingham