RE: doNotTrack attribute in API

With the exception API, it does vary. This is why the change was made. IE11 supports this API and this is why we made the change. We originally proposed it on navigator exactly to avoid polluting window but since it now varies we chose to accept the associated disadvantages.

-----Original Message-----
From: Marcos Caceres [mailto:mcaceres@mozilla.com] 
Sent: Tuesday, January 14, 2014 12:31 PM
To: Mike O'Neill
Cc: Adrian Bateman; 'Matthias Schunter (Intel Corporation)'; public-tracking@w3.org
Subject: Re: doNotTrack attribute in API



On Tuesday, 14 January 2014 at 20:16, Mike O'Neill wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>  
> Compared with” Do Not Track means do not track me beyond this network interaction” or “Do Not Track means do not tell me I am being tracked” what object the property lives on is of very little importance. The TPE says window so that’s fine.
It matters for platform consistency and it matters to developers (as we need to know where to find this stuff predictively). It also matters to implementers, as we need to know where to put stuff so we can have an interoperable platform (as you can see already, we have a serious problem where we now have shipping products exposing this property in different places). Because this is now in the wild, it means it's going to be very difficult to change: as it would likely break content that is now relying on this being on the navigator object in the case of Chrome and Firefox - and on window in the case of IE11.   

If no user agent today is providing per-origin control over do not track (are there any UAs that support this or plan to support this?), then I would urge the group to leave it on the Navigator interface. Doing the more sophisticated per-origin DNT should maybe be done in a future version of the spec.  

You may also ask the TAG to weigh in about the above.

--  
Marcos Caceres

Received on Wednesday, 15 January 2014 00:37:08 UTC