- From: <Frederick.Hirsch@nokia.com>
- Date: Fri, 7 Oct 2011 17:41:05 +0000
- To: <cantor.2@osu.edu>
- CC: <Frederick.Hirsch@nokia.com>, <mnystrom@microsoft.com>, <public-xmlsec@w3.org>
Right, I thought this would be the case.. I think we should go with the element. Same effect, avoid a number of issues. Anyone disagree? regards, Frederick Frederick Hirsch Nokia On Oct 7, 2011, at 12:44 PM, ext Cantor, Scott wrote: > On 10/7/11 8:35 AM, "Frederick.Hirsch@nokia.com" > <Frederick.Hirsch@nokia.com> wrote: >> Note that this is not the 1.1 schema but the 1.0 schema. However, this >> seems appropriate and as the new attribute is optional should not hurt >> existing implementations. > > Ordinarily, no, you can't do this. The only caveat that might make it ok > here is that the Algorithm is new. A 1.0 implementation wouldn't handle > the algorithm, so you're moving a failure from one layer to another. > > That said, this is generally just not done. There are assumptions built > into the whole house of cards of XSD aware software, and it just doesn't > fly. > > There's a simple fix, just use an element. EncryptionMethod has a wildcard > in it, it's just not an attribute wildcard. > > <xenc11:MGF Algorithm="..."/> > > -- Scott >
Received on Friday, 7 October 2011 17:41:53 UTC