RE: Parties and First Party vs. Third Party (ISSUE-10)

It's only a compromise if user expectations were a viable option.  If we only consider implementable solutions, I would probably redraw the lines like so:

Domain > Branding > Affiliation > Corporate Partnership

As you stated, the domain is too inaccurate, and in Brussels I do think we had consensus that corporate affiliation was a minimum requirement.  So it seems to me that a compromise would have to be somewhere between Branding and Corporate Affiliation.  Easy discoverability has been offered as one such compromise, although I have not yet heard a great mechanism for that.  What is the current proposal for an easy discoverability mechanism?

From: Jonathan Mayer [mailto:jmayer@stanford.edu]
Sent: Wednesday, March 14, 2012 11:46 AM
To: Kevin Smith
Cc: Jeffrey Chester; Tracking Protection Working Group WG; Shane Wiley; Sean Harvey
Subject: Re: Parties and First Party vs. Third Party (ISSUE-10)

Privacy advocate preferences: user expectations > branding > affiliation.
Industry participant preferences: affiliation > branding > user expectations.

That's how branding is a compromise.

Jonathan

On Mar 14, 2012, at 10:17 AM, Kevin Smith wrote:


I am not sure what branding would be a compromise between.  User expectation was never a feasible option because there is no technical representation of user expectation, and user expectations will vary by region, culture, experience, and can even vary wildly across users in a single home.  It is purely subjective.

From: Jeffrey Chester [mailto:jeff@democraticmedia.org]
Sent: Wednesday, March 14, 2012 6:48 AM
To: Tracking Protection Working Group WG
Cc: Shane Wiley; Sean Harvey; Jonathan Mayer
Subject: Re: Parties and First Party vs. Third Party (ISSUE-10)

We are willing to live with branding, which offers some guarantee that consumer/users can make more informed consent choices regarding their use of the W3C's Do Not Track system.  Discoverability is insufficient (and would require, for example, ensuring discoverability of the techniques of optimization, prediction, and automated audience buying that are increasingly a factor in a consumers online activities).

Branding is a compromise, as far as we are concerned.








On Mar 14, 2012, at 1:50 AM, Jonathan Mayer wrote:



Time we can deal with - I would support a phase-in period from corporate affiliation to branding.

Expense is a different matter.  It would be helpful to hear from the other multi-brand businesses in the group how much they believe engineering support for DNT would cost.  Specifics would greatly assist in understanding; bald assertions like "[a]ny other solution will cost industry 100s of millions of dollars" do us little good.

I'll be very disappointed if industry participants newly decide, six months into this process and over a year into defining DNT, that *any* shift from status quo party boundaries is unacceptable.  But if that happens, we'll have to balance economics against privacy.  We would, after all, be considering a Do Not Track standard that necessarily allows information flows that violate user expectations and cross brand boundaries.

Jonathan

On Mar 13, 2012, at 9:54 PM, Shane Wiley wrote:



Jonathan,

We may need to take the pulse of those in industry again as I believe through further discussion that Industry is firmly on the side of corporate affiliation + easy discovery (single click).  Any other solution will cost industry 100s of millions of dollars globally to either rebrand all of their online efforts and/or reengineer back-end systems to develop separation between brands.  These appears to be a non-starter for an "easy implementable" goal as small and mid size publishers will be looking to larger publishers to provide the tools to implement DNT on their properties.  And most of the larger publishers in the world operate under a multi-brand structure and would likely not implement DNT due to the expense.  I personally wish it were less expensive to unwind several hundred years of branding strategy across the globe, but I don't believe that will be possible in the timeframe of this working group.

- Shane

From: Jonathan Mayer [mailto:jmayer@stanford.edu]
Sent: Tuesday, March 13, 2012 9:46 PM
To: Sean Harvey
Cc: Tracking Protection Working Group WG
Subject: Re: Parties and First Party vs. Third Party (ISSUE-10)

Sean,

I don't doubt that many industry participants would greatly prefer a corporate affiliation standard, just as many civil society participants would greatly prefer a user expectations standard.  But we're now operating in the zone of compromise, where the relevant question is what stakeholders will accept.  And many participants, from both industry and civil society, have indicated they would accept branding.

Jonathan

On Mar 13, 2012, at 9:32 PM, Sean Harvey wrote:




Thanks Jonathan. I have been far from alone in espousing a corporate affiliation plus discoverability approach. As Shane from Yahoo and others have indicated on this list and in direct meetings, it is not the job of this standards committee to break up the multi-brand approach of many companies on the web. This is not an issue of my and Heather's objection, there is a broad disagreement with you on this topic that we can discuss further in a weekly meeting.


On Wed, Mar 14, 2012 at 12:30 AM, Jonathan Mayer <jmayer@stanford.edu<mailto:jmayer@stanford.edu>> wrote:
Sean,

I've heard both you and Heather express hesitation to adopt a branding approach.

To situate the discussion, we've had (for some time) four options for delineating parties and first parties vs. third parties: domain names, corporate affiliation, branding, and user expectations.  See http://lists.w3.org/Archives/Public/public-tracking/2011Oct/0343.html.

Domain names have been, I think it's very fair to say, thoroughly rejected as over- and underinclusive.  Corporate affiliation is a deal breaker for many privacy advocates given how it has been abused in other privacy regulatory regimes.  Many industry participants view a user expectations approach as unworkable.  (I disagree, and despite persistent grousing I *still* have not seen a concrete example of how the approach is unworkable.)  Branding is the only option that remains, and the discussion surrounding ACTION-123 and ACTION-124 both on- and off-list was very positive.

Given that context, could you please explain your concern and propose a better option?

Jonathan

On Mar 13, 2012, at 9:13 PM, Sean Harvey wrote:




Just to be very clear we absolutely do not have consensus on 2 or 3, nor are we near consensus on those points. Easy discoverability was the main issue to my knowledge.

On Wed, Mar 14, 2012 at 12:10 AM, Jonathan Mayer <jmayer@stanford.edu<mailto:jmayer@stanford.edu>> wrote:
We agreed in Brussels that:

1) If two entities are not related by corporate affiliation, they are not part of the same party.

>From discussion on the mailing list, I think we are very close to consensus on three other points:

2) Branding should determine party boundaries.

3) Branding should determine first parties and third parties.

4) An entity must make "discoverable" the other entities that it considers part of the same party.

We do not have consensus on a final issue:

5) If two entities are related by corporate affiliation, are they part of the same party?

I've taken a stab at text that captures these five points.  It is based on the current TCS document, the DAA principles, my proposal with Tom, and the CDT proposal.

--------------------------------------------------

I. Definitions

A. Network Interaction
A "network interaction" is an HTTP request and response, or any other sequence of logically related network traffic.

B. Entity
An "entity" is any commercial, nonprofit, or governmental organization, a subsidiary or unit of such an organization, or a person.

C. Affiliation
If an entity holds significant ownership in or exercises significant operational control over another entity, they are "affiliated."

D. Party
A "party" is any group of entities that:
a) consistently presents common branding throughout each entity, and
b) is related by affiliation.
[there is debate over whether to flip the "and" to an "or"]

E. First Parties and Third Parties
A "first party" is any party, in a specific network interaction, that brands content that occupies the full window.
A "third party" is any party, in a specific network interaction, that does not brand content that occupies the full window.

II. Transparency Requirement

A. Operative Text
A party must make reasonable efforts to ensure users can discover which entities it encompasses.

B. Non-Normative Discussion
A list of entities in a privacy policy would ordinarily satisfy this requirement.






--
Sean Harvey
Business Product Manager
Google, Inc.
212-381-5330<tel:212-381-5330>
sharvey@google.com<mailto:sharvey@google.com>




--
Sean Harvey
Business Product Manager
Google, Inc.
212-381-5330
sharvey@google.com<mailto:sharvey@google.com>