Re: [OAUTH-WG] Proposal: OAuth 1.0 signature in core with revision

Lukas Rosenstock <> Mon, 27 September 2010 21:43 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id A3CDF3A6D97 for <>; Mon, 27 Sep 2010 14:43:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.378
X-Spam-Status: No, score=-1.378 tagged_above=-999 required=5 tests=[AWL=0.598, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id Xr44D7kh5C6x for <>; Mon, 27 Sep 2010 14:43:31 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id 394CE3A6D5D for <>; Mon, 27 Sep 2010 14:39:45 -0700 (PDT)
Received: by qyk1 with SMTP id 1so5573604qyk.10 for <>; Mon, 27 Sep 2010 14:39:57 -0700 (PDT)
MIME-Version: 1.0
Received: by with SMTP id n19mr5930042qah.248.1285623596947; Mon, 27 Sep 2010 14:39:56 -0700 (PDT)
Received: by with HTTP; Mon, 27 Sep 2010 14:39:56 -0700 (PDT)
In-Reply-To: <90C41DD21FB7C64BB94121FBBC2E72343D45D80139@P3PW5EX1MB01.EX1.SECURESERVER.NET>
References: <90C41DD21FB7C64BB94121FBBC2E72343D45D80139@P3PW5EX1MB01.EX1.SECURESERVER.NET>
Date: Mon, 27 Sep 2010 23:39:56 +0200
Message-ID: <>
From: Lukas Rosenstock <>
To: Eran Hammer-Lahav <>
Content-Type: multipart/alternative; boundary="00c09f899702cf9c5804914491b2"
Cc: "OAuth WG (" <>
Subject: Re: [OAUTH-WG] Proposal: OAuth 1.0 signature in core with revision
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: OAUTH WG <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Mon, 27 Sep 2010 21:43:42 -0000

In my opinion, there could be two arguments against including 1.0 signatures
into the core:
1) The spec will be too long.
2) The signature mechanism isn't good, so it would be replaced with
something else anyway.
Is 1) a concern? For 2), I think the adoption of OAuth 1.0 speaks for the
signature mechanism being both feasible to implement and secure. Also as
it's already written it should not delay finishing the spec.

I agree with Dick Hardt that if the signatures are included, there should be
a mechanism to allow for different signature schemes who might be defined in
the future.

There's another question: What will be required for OAuth compliance? Is
only-bearer-tokens fine or must the signature mechanism be supported? Is it
fine to not allow both but only a third signature algorithm?

By not including signatures some people (those who strongly feel about
signatures) will feel OAuth 2 is somewhat "less" than OAuth 1 and stick with
OAuth 1 instead of the whole web world moving forward to a new standard.

2010/9/27 Eran Hammer-Lahav <>

> Building on John Panzer’s proposal, I would like to ask if people have
> strong objections to the following:
> - Add the 1.0a RFC language for HMAC-SHA-1 signatures to the core
> specification in -11
> - Discuss the signature language on the list and improve both prose and
> signature base string construction
> - Apply improvements to -12
> Keeping the 1.0a signature in the core specification makes sense and builds
> on existing experience and deployment. If we can reach quick consensus on
> some improvements, great. If not, we satisfy the need of many here to offer
> a simple alternative to bearer tokens, without having to reach consensus on
> a new signature algorithm suitable for core inclusion.