W3C home > Mailing lists > Public > public-xmlsec@w3.org > January 2011

Re: ACTION-764: Review placement of base64 alg in 1.1/2.0, should it be under transforms?

From: <Frederick.Hirsch@nokia.com>
Date: Tue, 11 Jan 2011 02:52:59 +0100
To: <cantor.2@osu.edu>
CC: <Frederick.Hirsch@nokia.com>, <bal@microsoft.com>, <public-xmlsec@w3.org>
Message-ID: <6A7BFF75-1B54-4F4D-87A0-498D4314BF41@nokia.com>
I agree it is essential to list under transforms, but as I mentioned since it also is used for encoding it should also be listed there with a note that it is the same URL and that encoding is used with Object.

Does that make sense to you ?

regards, Frederick

Frederick Hirsch

On Jan 10, 2011, at 8:25 PM, ext Cantor, Scott E. wrote:

>> ACTION-764 asks whether we should re-classify Base64 as a Transform
>> Algorithm from an Encoding Algorithm in the required algorithms list (in,
>> e.g., XMLDSIG 1.1 Section 6.1).
>> Looking at XMLDSIG 1.1 and XMLENC 1.1, it appears that we use the same
>> Base64 URI in both ³encoding algorithm² and ³transform algorithm²
>> contexts, and the URI is a valid in both.
>> For example,  in XMLDSIG 1.1 Section 4.6, there's an
>> example in the text where we talk about putting a PNG image in an
>> <Object> element and identifying it with encoding=<the base64 encoding
>> URI>.  And we have a defined Encoding attribute on Object in the schema.
>> (We use a similar example in XMLENC Section
>> 3.1 EncryptedType.)
> At least in the XML Signature case, Encoding there is advisory only and
> does not connote processing semantics. OTOH, the base64 transform does. So
> at the very least, that algorithm URI is perhaps both an "encoding" (which
> does not seem to be normative concept) and a Transform, and ought be
> listed as both. If I were to pick one, that one would be the Transform.
>> I havenıt checked in 2.0 yet, but given the existing usage in 1.1 Iım not
>> sure Iıd move it under Transforms.
> I think I would do that or duplicate it and provide more context for what
> an "Encoding Algorithm" means. I'm not as concerned about that for 1.x,
> it's an old issue there.
> -- Scott
Received on Tuesday, 11 January 2011 01:53:43 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:55:15 UTC