Re: [apps-discuss] Questions about Structured Syntax Suffixes (SSS)

Mark Nottingham <> Thu, 24 May 2012 10:58 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 1020921F8606 for <>; Thu, 24 May 2012 03:58:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -102.999
X-Spam-Status: No, score=-102.999 tagged_above=-999 required=5 tests=[AWL=-0.400, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id RHuAFaxVmSrb for <>; Thu, 24 May 2012 03:58:09 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id 5A79621F85BB for <>; Thu, 24 May 2012 03:58:09 -0700 (PDT)
Received: from (unknown []) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPSA id 8CC8E22E25B; Thu, 24 May 2012 06:58:01 -0400 (EDT)
Mime-Version: 1.0 (Apple Message framework v1278)
Content-Type: text/plain; charset="us-ascii"
From: Mark Nottingham <>
In-Reply-To: <>
Date: Thu, 24 May 2012 20:57:59 +1000
Content-Transfer-Encoding: quoted-printable
Message-Id: <>
References: <> <>
X-Mailer: Apple Mail (2.1278)
Cc: Apps Discuss <>
Subject: Re: [apps-discuss] Questions about Structured Syntax Suffixes (SSS)
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: Thu, 24 May 2012 10:58:10 -0000

On 24/05/2012, at 8:39 PM, Henry S. Thompson wrote:

> Mark Nottingham writes:
>> Question: Is this actually useful in practice? I.e., what are the
>> real-world use cases for SSS?
>> We started this experiment with +xml, and I'm not aware of much
>> software that uses that suffix to great advantage (please educate me
>> if I'm overlooking something).
> Good question!  I did a quick check, and the four browsers I looked at
> did _not_ treat material served as application/foo+xml as XML (unless
> its URI ended with ".xml").

Ah, that.

> So _currently_ (aside from application/html+xml, which is recognised
> as a unit, not specifically as allowing generic XML processing because
> of the '+xml') the suffix serves only, as you suggest, as
> human-readable confirmation of expectations.
> _But_ I think there is at least a potential upside going forward, in
> that as we see both ...+xml and ...+json variants of particular types,
> I _would_ expect software to dispatch on this information, rather than
> sniffing.  Or, to turn that around, if we _don't_ support SSS, we are
> in practice requiring applications to sniff.

What software? If it knows about the specific format -- whatever convention is used -- it's not necessary to have a convention. The convention is only good for software that DOESN'T know about the specific semantics/syntax of the format in use, but can derive some value from knowing a few generic conventions about it.

The only one I can come up with is a generic test tool, like the one that I run at <>; with this sort of information, it'd be able to colourise the syntax of a few more formats that pass through it. 

To me, that's pretty thin justification; I'm not asking for this, and I run one of these tools!

What I'm concerned about is that this may encourage people to register both XML and JSON variants of a format -- in fact, I suspect that this is part of the underlying motivation. I've already given my thoughts about that here <>.


Mark Nottingham