Re: diff of TPE editing since the FPWD

On 1/12/12 4:43 PM, Bjoern Hoehrmann wrote:
> * David Wainberg wrote:
>> In either case, we need to define (cross-site) "tracking", right? Let's
>> just define the data collection and use we're concerned with, and skip
>> defining the parties. As I've argued before, we are doing this
>> backwards. We've been trying to back into a definition of tracking by
>> defining parties and use exceptions. But if we just define the
>> collection and use that constitutes tracking, first the need refinements
>> via exceptions, etc. will become quite clear, and second, most
>> everything else will fall into place.
> The Working Group is expected to produce something that allows people to
> say something is contrary to the specifications produced by the group.
> Ed Felten's question was whether there is anything that would be contra-
> ry to the group's specifications under a party-based definition but not
> under a "cross-site" based definition, or vice versa.
Ed's question was "Is this "cross-site" discussion a debate about 
substance, or only about terminology?" Perhaps I could have been more 
direct, but my answer is that there is a substantive difference, and not 
one of mere terminology. If it were mere terminology the terms would be 
interchangeable in the spec, but they are not. The key difference I see 
largely goes to the complexity of definition and implementation, rather 
than to the expected end result for users, but to me that's a important 
difference. A party-based definition will generate unnecessary and 
avoidable ambiguity and complexity for the companies trying to adhere to 
the standard.