- From: Joseph Reagle <reagle@w3.org>
- Date: Tue, 18 Jun 2002 12:58:40 -0400
- To: Ari Kermaier <arik@phaos.com>, "Takeshi Imamura" <IMAMU@jp.ibm.com>, merlin <merlin@baltimore.ie>, Aleksey Sanin <aleksey@aleksey.com>, Jiandong Guo <jguo@phaos.com>
- Cc: xml-encryption@w3.org
On Tuesday 18 June 2002 11:39 am, Ari Kermaier wrote: > Frankly, I'm not sure how the text in [1] actually helps implementors. I'd agree, it doesn't help an implementor with respect to how to implement it. That's supposed to be a feature, though the result should be interoperable. > What does "copying" nodes from node-sets X and Y to node-set Z mean -- > aren't X's and Y's nodes in different underlying documents? Yes. Create a node as defined in [a] akin to that in the existing document (unless decrypted). [a] http://www.w3.org/TR/xpath#data-model > It seems that > what we're trying to do is help implementors fix up the resulting XML > document, using language that remains consistent with the XPath data > model. This may be a futile effort. It is a difficult task, but our critical requirement is to be able to "decrypt-and-replace". If we can specify it abstractly and still have interoperable results: good. What would you like to see Ari? Do you feel the present Candidate RECs are approriate and we should add necessary warning/limitations text? Do you have more specific text in mind?
Received on Tuesday, 18 June 2002 12:58:41 UTC