RE: Request to close ISSUE-78

I like Sid's approach now that the use of "no header" has been explained. It certainly seems like no header should be the proper default.

JC
Twitter


-----Original Message-----
From: Sid Stamm [mailto:sid@mozilla.com] 
Sent: Tuesday, December 20, 2011 4:58 PM
To: JC Cannon
Cc: David Singer; Tracking Protection Working Group WG
Subject: Re: Request to close ISSUE-78

Here's the way I think about the difference between DNT:0 and no header:

DNT:0 means "you can track me", regardless of the society or jurisdiction, it is explicit consent.  In places where the social norms accept tracking for people who haven't been asked (USA), DNT:0 is equal to no header.  In places where social norms reject tracking for people who haven't explicitly provided informed consent (UK), DNT:1 is equal to no header.

Because header seems to be socially-centered or perhaps based on jurisdiction, I'm advocating that lack of header means "whatever society accepts because the individual hasn't weighed in" and the presence of the header is an explicit request one way or the other.

If it clears things up, perhaps we could add a third value that means the same thing as no header.  Then we would have:

0: User accepts tracking
1: User rejects tracking
2: User's preference undefined

But if that's the case, it seems to me that implementing the third option (2 here) should be optional for user agents -- to save bytes we could just not send the header if the user's preference hasn't been established or discovered.

Until we all settle on a standard, here's Mozilla's current position on the topic:
http://blog.sidstamm.com/2011/11/firefox-wont-activate-dnt-by-default.html


-Sid

----- Original Message -----
> From: "JC Cannon" <jccannon@microsoft.com>
> To: "David Singer" <singer@apple.com>, "Tracking Protection Working Group WG" <public-tracking@w3.org>
> Sent: Tuesday, December 20, 2011 4:00:18 PM
> Subject: RE: Request to close ISSUE-78
> This sounds odd to me as well. I thought DNT:0 was the same as no
> header.
> 
> JC
> Twitter
> 
> -----Original Message-----
> From: David Singer [mailto:singer@apple.com]
> Sent: Tuesday, December 20, 2011 3:58 PM
> To: Tracking Protection Working Group WG
> Subject: Re: Request to close ISSUE-78
> 
> 
> On Dec 20, 2011, at 12:50 , Kevin Smith wrote:
> 
> > I see. So, if we stick to this definition, clearly no DNT header
> > does NOT mean the same thing as DNT:0 because DNT:0 actually means
> > DNT is turned on. However, if DNT:0 means DNT is on, but there is a
> > local exception...
> 
> It's a little odd to have a 'protocol' where the default behavior
> indicated by the absence of the header cannot be explicitly signaled
> in the header. Generally, it's possible to say "I want default
> behavior" explicitly. It's odd enough to notice, but not (IMHO)
> serious enough to add a value (-1?) to say "whatever, default
> behavior, please".
> 
> 
> David Singer
> Multimedia and Software Standards, Apple Inc.

Received on Wednesday, 21 December 2011 01:01:22 UTC