Re: High-level text on third-party responsibilities (ACTION-38, ISSUE-19)

>> It is both possible to uniquely identify a user agent but not a user
(example: fingerprinting a specific web browser, but no idea which
household member is using it at a particular time)

Now that internet is becoming more and more accessible from mobile
devices, it is safe to argue that it is possible to uniquely identify a
human with fingerprinting, for the specific web browser will be tied to
the mobile device you are carrying.

Rob

Aleecia M. McDonald wrote:
>
> On Dec 13, 2011, at 8:56 PM, Shane Wiley wrote:
>
>> Jonathan,
>>
>> If you feel by not adding user agent that this somehow creates a
>> loop-hole that companies will attempt to thwart to avoid DNT, then I'm
>> okay (reluctantly) putting user agent back in the language although this
>> feels wasteful.
>
> Really quickly for those playing our home game:
> 	A user -> a human
> 	A user agent -> usually a web browser (ex: Internet Explorer)
> 	A user-agent string -> a text description a user agent can send about
> itself (ex: "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)")
>
> It is both possible to uniquely identify a user agent but not a user
> (example: fingerprinting a specific web browser, but no idea which
> household member is using it at a particular time) or to uniquely identify
> a user but not a user agent (example: a user logs in to a website over
> time from various computers in a public library, Starbucks, etc.).
>
> The people contributing to this thread seem to understand, but even with a
> little prior discussion there may be a few readers who did not.
>
> 	Aleecia
>
>
>

Received on Wednesday, 14 December 2011 08:07:33 UTC