- From: Rigo Wenning <rigo@w3.org>
- Date: Tue, 4 Nov 2003 01:00:52 +0100
- To: "'public-p3p-spec@w3.org'" <public-p3p-spec@w3.org>
On Wed, Oct 22, 2003 at 03:31:48PM -0400, Dobbs, Brooks wrote: > 1) We need be careful between providing tools on which to make policy > and making policy itself. While it may be a useful tool to have the ability > to state that collection covered by separate PRFs is under the control of > the same entity, ultimately it is up to the UAs to decide how they want to > implement these tools. I think, the beneficial in this proposal is, that it allows to have a means by which the double control on PRF's is checked. In fact, the same entity requires that the issuing host's PRF mentions the same entity AND the same entity has to refer to the issuing host. As configuration of both needs privileged access to configuration files, hacking without control of those files will be harder. Therefor, I think the addressing mechanism is good, but I have issues with the semantics like I expressed them during the last calls. > > 2) We use the term entity (and later in the agent relationship > proposal - "agent") without providing sufficient guidance on what this > means. Rigo suggests there may be ways around this but I think they would > be unclear at best and difficult to explain to the end user. This suggests > to me that point 1) may mean that UAs don't act on this tool if they had it. I referred to the rather clear definitions in the EU Directive, where a data controller is the entity that controls data processing, means has power to order, manage and prohibit things. The agent for me here is a data processor, means an entity that does some processing on behalf of somebody else (i.e. with a contractual background that allows that control). But I agree, that this is rather complex. > > 3) I think that there is an issue in that entity is today defined at > the policy level. A PRF for a host can reference 2 distinct policies each > of 2 distinct entities (data controllers) - how then can there be meaning to > say at the PRF level this entity is the same as anything else? The same as > which of the 2? Don't forget, that the entity is not defined at the PRF-level but at the Policy-level. This means that a same Policy with the same entity can be referenced by multiple PRF's. We extend that proposal now to multiple PRF's by the way of allowing third party PRF's to reference other PRFs. > > 4) There would need to be a mechanism to expire same entity > relationship. This may be handled through the expiry in policy files but > there is currently no such mechanism for CPs Advantage and disadvantage of CPs is, that they have a lifetime for exactly the request that is made. So no need for expiry on CPs. CPs are out of scope here. I know that the issue came from the CPs and from the third party cookies. But we would be rather unwise to remove the incentive to create policies in the third party context. BUT: Full policies in the third party context were a nuisance so far and this proposal might remedy the performance issues attached to it, if we get it right. > > 5) Headers are limited practically by the number of same hosts that > they can declare. Also declaring multiple "same-entity" relationships would > presumably require a validation against each full policy prior to accepting > the accuracy of the header declaration. This is somewhat illogical, as, IF > we where to accept the performance hit of policy prefetching, then we would > have no need for CPs! Jack rightly points out that we may only need to > validate against the PRF and not the policy - but I think that I may have > brought that into question with point 3). As far as I remember our discussion, the same host has to reference the PRF of the referring host. (apart from CP, where we are in a stateless context). This means that we might obtain sufficient performance improvements to get rid of CPs and avoid the false declarations inherent to the summary-nature of CPs. > > 6) Again on the topic of CPs, once a same entity relationship has been > established between cookie (set by A) and site B from which an asset > appears, how they can a same entity relationship be established from a > subsequent replay on site C (provided that PRFs or policies) for all these > sites do indeed declare each other as same entity? This shows the issues which were discussed the most. It means that browsers would have to switch to pay attention to cookies replayed. At the moment, the performance improvement is that we only evaluate cookies at set-cookie. If the same cookie is re-used in different contexts, either you declare up-front all the potential uses or you evaluate on the replay-time. While 1/ is good for browsers, 2/ is good for content providers using the same cookie in different contexts. This might be easier with the full format, where specific data has a specific statement attached to it. In the context of CP, this just turns into nightmare, as the potential overstatement is huge. The same entity proposal does not remedy the situation. But it would perhaps allow the browsers to get sufficient performance to do full policy. Full policy remedies as it allows for fine-grained expression of data concerned. Best, Rigo
Received on Monday, 3 November 2003 19:01:01 UTC