- From: Lars Nyman <larsnyman@14designs.com>
- Date: Wed, 19 Sep 2001 12:03:38 -0700
- To: "Martin Presler-Marshall" <mpresler@us.ibm.com>
- Cc: <www-p3p-policy@w3.org>
> What you've got here is two conflicting policy reference files. Each > one declares a policy which covers all cookies on the site...this is a bad > thing. P3P's non-ambiguity rules require that the site only declare one > policy for a given cookie or URL. Ah, thanks for pointing that out. I had failed to realize that the Significance of order (section 2.3.2.1.1) only applies to POLICY-REF elements in the same file. I went back to the spec and looked at 2.4.1 Non-ambiguity. That section also states "If a policy reference file at the well-known location declares a non-expired policy for a given URI, this policy applies, regardless of any conflicting policy reference files referenced through HTTP headers or HTML link tags." So, I guess in my particular example, even though there are two conflicting policy reference files, the user agent should use the policy specified by the policy reference file at the well-known location. But, regardless, that was not my intention - the intention was to let each store specify their own cookie policies and to do that I have to use more carefully crafted policy reference files. > Now, let's assume that the two policy reference files weren't > conflicting. Imagine that the main site sets one cookie on every single > page under mall.example.com, and the shoe store sets a second cookie for > its pages. Imagine further that the two policy reference files > differentiate this correctly, perhaps by naming the cookies to include or > exclude. In this case, the user-agent would need to consult both policy > reference files in order to find the policy for the two cookies. Thanks, I think this answers my original line of questions. Lars > > -- Martin > > Martin Presler-Marshall - Program Manager, Privacy Technology > E-mail: mpresler@us.ibm.com AIM: jhreingold > Phone: (919) 254-7819 (tie-line 444-7819) Fax: (919) 254-6430 (tie-line > 444-6430) > >
Received on Wednesday, 19 September 2001 15:03:36 UTC