W3C home > Mailing lists > Public > public-tracking@w3.org > May 2012

Re: tracking-ISSUE-150: DNT conflicts from multiple user agents [Tracking Definitions and Compliance]

From: Chris Mejia <chris.mejia@iab.net>
Date: Wed, 30 May 2012 19:34:19 +0000
To: W3C DNT Working Group Mailing List <public-tracking@w3.org>
Message-ID: <CBEBDE0C.1B2E1%chris.mejia@iab.net>
I believe new Issue-150 is closely related to open Issue-143. If the user's intent in turning on/off DNT is not clear (especially in cases where the user doesn't even know they are specifically sending a DNT:1 header), there is no way for publishers to understand how to accurately "honor" any consumer's DNT header flagó it's a fundamental flaw with this scope of this proceeding.  I laid out the concern in some detail in my previous email to the group ("In Support of Issue-143"); so I'll just give the brief version here: if publishers do not understand the context of the user's DNT expression (was the user properly informed about what setting does/means, before it was set) how are publishers to determine what the user actually intended, or if they user is even aware that a DNT flag is being sent?  If any question/statement in any UI can lead to the sending of DNT:1 or DNT:0, where is the integrity of the system/solution?

To give just one example (there are many) of how a DNT mechanism that lacks a uniform informed consent requirement might be abused, consider the theoretical yet plausible scenario where an email is sent to (millions of) users informing the users that they should "click here to prevent evil doers from knowing who you are" or even worse, "click here if you think blue is a pretty color" (replace with a variety of malware tactics), the user's click leading to a programatic setting of DNT, without the user's informed consent under uniform compliance rules.  When that happens (some zealot decides to abuse the system), I'm sure we'll eventually learn about it, after some amount of damage being done.

When it becomes known that users were deceived into sending a DNT expression (no uniform informed consent), here's what the end-game of publishers might be:  without a way of discerning how DNT was set (which program; who owns the program; being able to inspect the program), and under which auspices it was set (what did the user agree to when they clicked?), when learning of a set of users who were deceived into setting DNT, publishers may be forced to consider if they should honor any DNT header requests at all, in an effort to protect the web experience of all users.  Under this scenario, publishers may be compelled to issue public statements outlining the fatal flaws of this W3C DNT mechanism, citing the specific abuses, and walking away from compliance on the grounds that being "compliant" with such a system would be harmful to the majority of its users.

Is that really the result that this working group is looking for?  If not, I strongly suggest that we all get on board with defining a system where the actual intent of the user is absolutely clearó the only way I can think to accomplish this is to require compliance with a uniform requirement to properly educate/inform the user about their choice, at the point user choice is made.  Of course I'm open to hearing other suggestions for solving this problem, but I feel that "it's out of scope/Charter for this project" is not an acceptable solutionó that answer does not solve the problem described here and in open Issue-143.  Please, let's solve the actual problem.

Chris Mejia, IAB/DAA


On 5/30/12 1:35 PM, "Tracking Protection Working Group Issue Tracker" <sysbot+tracker@w3.org<mailto:sysbot+tracker@w3.org>> wrote:

tracking-ISSUE-150: DNT conflicts from multiple user agents [Tracking Definitions and Compliance]

http://www.w3.org/2011/tracking-protection/track/issues/150

Raised by: Aleecia McDonald
On product: Tracking Definitions and Compliance

Due to multiple addons that support Do Not Track, there could be conflicts. For example, a user could turn off DNT (not unset, actually off, sending DNT:0) in Firefox, yet install Abine's "Do Not Track Plus" addon (which sends DNT:1). More fun, users could have three different addons, each with a different value. Do we have either best practices or requirements for user agents here?

Created from original issue-148, with actions taken by ifette and jmayer to write proposals.
Received on Wednesday, 30 May 2012 19:35:15 UTC

This archive was generated by hypermail 2.3.1 : Friday, 21 June 2013 10:11:28 UTC