Re: [ISSUE-5] What is the definition of tracking?

In parallel with this thread, and following on from the "cross-site" discussions in Brussels, I sketched this definition of tracking, which I think is quite like Roy's.  Critically, it treats all servers the same - no-one has to guess or know or be in agreement about "what the user thought they were visiting", what is 1st or 3rd party.  And for the vanilla site that wants ordinary transactions and logs, it's easy to comply.


"Tracking is the retention or use by a site, after a transaction is complete, of data records that can or do associate that user with either any other site, or with data not derived from the user's direct transaction with that site."

So this allows:
* knowing about another site during a transaction (e.g. 'please supply an advert to thew New York Times')
* keeping logs of what happens on your site ('Dave was served an ad for dishwashers') (which can be used for story-boarding, frequency capping, etc.)

It does not allow:
* remembering I was on sears.com and that's why I was served a dishwasher ad
* remembering the source URL when that conveys information from, or identifying, another site (e.g. resources specific to a host, or info passed about the user)
* remembering referer information
* combining the data from my transaction with other data to work out who I am or facts about me

It makes life hard for re-direction services, since almost any logging would involve identifying another site (so they had better not remember data that can identify users).

This would mean, however, that the current basis of 1st vs. 3rd party, and exceptions on those grounds, would need re-visiting.  Which is a big change to the model.


David Singer
Multimedia and Software Standards, Apple Inc.

Received on Tuesday, 27 March 2012 23:47:33 UTC