Re: Deciding Exceptions (ISSUE-23, ISSUE-24, ISSUE-25, ISSUE-31, ISSUE-34, ISSUE-49)

Alan, 

while your argumentation may be compelling within undeclared neutral grounds, 
it isn't as convincing after the user has expressed the clear will to not be 
tracked/recorded/followed/profiled/classified/targeted. And this user has 
received a response saying: "We honor your expressed preference". If this 
acceptance now means: "We except with our fingers crossed behind our backs and 
will take advantage of an exception to have you 
tracked/recorded/followed/profiled/classified/targeted" then the average user 
will find that a bit confusing. An exception isn't an exception anymore if it 
is the default IMHO. 

If we accept that the exceptions we are discussing in a DNT=1 scenario are 
still exceptions, than those exceptions derive from the general rule: not 
tracking. And it is the nature of exceptions that they have to be justified as 
a departure from the general rule. And they have to be interpreted in a narrow 
way. Because otherwise, they aren't exceptions anymore but a change to the 
general rule. A change to the general rule would mean here that DNT=1 means 
DNT=unset which in turn would make all our efforts pretty futile. And I 
wouldn't want Shane to go through his excellent list of things a company must 
do for DNT compliance without any need to just have a fig leaf for further 
collection.

So I'm sorry to say that the burden of argumentation in a DNT=1 scenario is 
with those claiming the exception and wanting to collect and use data. Sean 
and Shane have done a good job on why they want the data. I tried to explain 
we shouldn't collect that data (arguments to be improved). Ninja and Roy 
hinted at a solution by retention limitation and Jonathan hinted at a client 
side solution. I would rather like to discuss those concrete solutions than 
fundamentally question the exercise we are doing.

Best, 

Rigo

On Thursday 02 February 2012 12:17:25 Alan Chapell wrote:
> I agree with much of what Shayne said - but will add the followingŠ.
> 
> It seems like much of the discussion (on this topic at least) is a bit
> one-sided. If we're going to ask industry to granularly explain why
> certain data uses pass Jonathan's Compelling need test, then it seems fair
> to ask Jonathan (and/or others) to be able to granularly demonstrate (for
> example) how and to what extent client side frequency capping approaches
> work. The Stanford team has clearly done some fantastic work here, but a
> test using a relatively small game network may or may not translate
> perfectly outside of the testing environment. And I think its fair to ask
> for a clear demonstration of how these concepts can be applied by both the
> MSFT's and Googles, as well as tier 2 and tier 3 companies.
> 
> On that note, I'm concerned that much of what we're talking about
> implementing will create all kinds of technical and logistical issues for
> companies who don't have the resources of a Yahoo or Adobe.
> 
> Has the group considered bringing in some invited experts from the
> long-tail of industry to help ensure that this is something that they can
> implement without hiring an army of tech consultants?
> 

Received on Wednesday, 8 February 2012 14:08:34 UTC