W3C home > Mailing lists > Public > public-tracking@w3.org > February 2013

RE: ISSUE-10 First party definition, ISSUE-60, ACTION-?

From: Alexander Hanff <a.hanff@think-privacy.com>
Date: Wed, 27 Feb 2013 20:46:30 +0100
To: "'John Simpson'" <john@consumerwatchdog.org>
Cc: <public-tracking@w3.org>
Message-ID: <012f01ce1523$245df8a0$6d19e9e0$@think-privacy.com>
Hi John,


RFC's relating to cookies (where the talk of 1st and 3rd party entities
began in 1997) are RFC 2109 (1997), RFC 2965 (2000) and RFC 6265 (2011)
where it is made pretty clear that a first party is an entity that exists
within the requested URI (either directly or a sub-domain of the requested
URI) anything else is third party.


It is actually quite interesting that we have come full circle to the RFCs
because there was an incredibly aggressive debate at the time whether or not
to block 3rd party cookies completely because of the privacy concerns (back
in 1997) - predictions which have since come true.  The 2000 discussion was
more relaxed and agreed to allow third party cookies but to give users
controls to block them if my memory serves me correctly (which is why we
have 3rd party cookie options in web browsers to this day).  The 2011
discussions were equally concerned about the privacy issues with 3rd party
cookies but allowed leverage on how browsers handle them.


However when talking about first and third party, this was first discussed
in the 1997 RFC with the requested URI definition to my knowledge - I am
willing to read any earlier definitions if someone has one to hand?




Alexander Hanff


From: John Simpson [mailto:john@consumerwatchdog.org] 
Sent: 27 February 2013 20:25
To: Alexander Hanff
Cc: public-tracking@w3.org
Subject: Re: ISSUE-10 First party definition, ISSUE-60, ACTION-?


Hi Alexander,


Can you please point me to the RFC definitions to which you refer?




John M. Simpson

Privacy Project Director

Consumer Watchdog

2701 Ocean Park Blvd., Suite 112

Santa Monica, CA, 90405

Tel: 310-392-7041

Cell: 310-292-1902




On Feb 27, 2013, at 8:48 AM, Alexander Hanff <a.hanff@think-privacy.com>

The issue in question is not whether or not people will be aware that by
clicking on a Like button it will post something to their timeline - that is
not the purpose of Do Not Track.  The issue in question is whether or not
someone accepts or consents to Facebook tracking their online behaviour if
they click on a like button and do so across all web sites where those
buttons exists - furthermore, just clicking on the button is not an accurate
description of how this tracking works.


My understanding is that if a user is currently logged in to Facebook or has
any Facebook cookies on their machine, merely loading a page with the "Like"
button script embedded is enough for Facebook to be able to track that user
across sites with the widget.


This redefinition is not within the scope of TPWG  - TPWG's purpose is not
to redefine existing RFCs (1st/3rd party definition has existed in RFC for
ooo about a decade or more) - TPWG's purpose is to come up with a standard
and compliance specification to deal with 3rd party tracking consent


How any of you can sit there with a straight face and say it is ok to
redefine a technical term that has existed for over a decade is beyond me.
This "new" definition goes against the very premise of DNT which is to send
a signal about not being tracked; by making some of the most invasive and
widespread tracking technologies immune to the standard simply by redefining
1st party to put them out of scope, is reprehensible.


Alexander Hanff


From: Justin Brookman [mailto:justin@ <http://cdt.org> cdt.org] 
Sent: 27 February 2013 17:34
To:  <mailto:public-tracking@w3.org> public-tracking@w3.org
Subject: Re: ISSUE-10 First party definition, ISSUE-60, ACTION-?


There is no consensus definition of "first party" --- there are three
separate ones in the text.  I believe they all say much the same thing and I
was merely trying to merge them. :)

I believe the group is at consensus that if someone clicks a "Like" button,
then it is reasonable to expect that Facebook is going to receive
information that falls outside the scope of Do Not Track (namely, that the
user 'likes' some particular page or pbject, and now FB can display that in
Newsfeed and Timeline consistent with the user's privacy settings).  If
anyone in the working group disagrees with that, feel free to speak up.
Alexander, if you want to comb through the mailing list to see our previous
exhaustive discussions on this, you may find them informative.  Or you may
not, I don't know.

However, you do, obliquely, get to a relevant point --- that perhaps the
definition should include be revised to say "clearly branded" before
"embedded widget" in order to make sure that the user knows what she's
clicking on.  I believe the group had discussed something similar
previously.  I would be fine with a discussion on what constitutes clear
branding (I would say things like the Like, Tweet, and +1 buttons qualify)
in an appendix.

Justin Brookman
Director, Consumer Privacy
Center for Democracy & Technology
tel 202.407.8812
 <mailto:justin@cdt.org> justin@cdt.org
 <http://www.cdt.org> http://www.cdt.org

On 2/27/2013 11:01 AM, Alexander Hanff wrote:

Why is the group second guessing what consumers think?  The definition of
first party already exists, there is no need to redefine it in a light which
makes it easier for exceptions to be made for tracking widgets.


Many users will not be remotely aware that a "Like" button is actually
hosted by Facebook, they would assume it is hosted on the domain they are
visiting.  To assume otherwise is absurd and further weakens the validity of
this DNT process.


Alexander Hanff


From: Justin Brookman [ <mailto:justin@cdt.org> mailto:justin@cdt.org] 
Sent: 27 February 2013 16:52
To:  <mailto:public-tracking@w3.org> public-tracking@w3.org
Subject: ISSUE-10 First party definition, ISSUE-60, ACTION-?


Peter asked me to try to combine the three definitions of "first party" in
the current text in consultation with Heather.  The existing definitions are
all very close, and I don't think there are major substantive disagreements
here.  Anyway, here is my best effort (Heather provided feedback, but she's
not around this morning, so I don't know if she blesses this):


In a specific network interaction, if a party can reasonably conclude with
high probability that the user intends to communicate with it, that party is
a <dfn>first party</dfn>.  In most cases on a traditional web browser, the
first party will be the party that owns and operates the domain visible in
the address bar.  A first party also includes a party that owns and operates
an embedded widget, search box, or similar service with which a user
intentionally interacts.  If a user merely mouses over, closes, or mutes
such content, that is not sufficient interaction to render the party a first


Rob Sherman is separately working on text regarding multiple first parties.


Chris Pedigo and Vinay Goel are separately working on text regarding data
processors that stand in the shoes of their controllers, party-wise.

Justin Brookman
Director, Consumer Privacy
Center for Democracy & Technology
tel 202.407.8812
 <mailto:justin@cdt.org> justin@cdt.org
 <http://www.cdt.org> http://www.cdt.org


Received on Wednesday, 27 February 2013 19:46:57 UTC

This archive was generated by hypermail 2.3.1 : Friday, 3 November 2017 21:45:04 UTC