#1062: TAG spec review of Stateless Bounce Tracking Mitigations
Discussions
Log in to see TAG-private discussions.
Comment by @MrPickles Feb 27, 2025 (See Github)
Also cc @Trikolon and @bvandersloot from Mozilla as FYI.
Comment by @lknik Apr 11, 2025 (See Github)
I support this proposal because this non-transparent method is being used since >10 years.
Discussed
Apr 14, 2025 (See Github)
Lola: I'm a bit confused, but you can assign me to this … is the spec to mitigate what they described? Hadley: There's two explainers about one spec Matthew: Not an expert, but interested in it … may have access to an expert soon Hadley: added @torgo as well
Discussed
Apr 21, 2025 (See Github)
Torgo: Missing Hadley and Matthew. Next week.
Discussed
Apr 28, 2025 (See Github)
Matthew: we did look at it - delta on the spec as a whole - explainer is missing [moved to https://github.com/privacycg/nav-tracking-mitigations/blob/main/explainers/bounce-tracking-mitigations.md] - re: the stateless part : recognizing that some types of bounce tracking do not use local storage. It seems like a good delta on what was there. We didn't see any architectural issues...
Jeffrey: the explainer moved to a sub-directory. Issue updated. I believe they have a specification ... in tracking mitigations ... there also is a dual-use site that is a site people go to but can also be used for bounce tracking... but that's not targeted here. One risk : now that google has cancelled work on actually removing 3rd party cookies I'm not sure what the shipping state is, but that shouldn't impact our review. .. should we say we're happy with it? please keep iterating?
Matthew: explainer is more detailed than just the bit about the stateless part. I haven't reviewed it all. It looks pretty good. It has a detailed alternatives considered section. No architectural concerns from looking at the spec. So if that's sufficient then I think it's OK to say "carry on with it".
Jeffrey: also we want to recognize that several browsers shipped bounce tracking mitigations without specifying, so it's good that someone specified it.
Matthew to write closing comment and post after doing quick check
Jeffrey: +1
Discussed
May 5, 2025 (See Github)
Hadley: sounds like we're happy.. DanA, any opinions?
DanA: Did we discuss how this is impacted by recent announcements about 3pc.
Jeffrey: Not in the brainstorming, but in the last breakout where this was mentioned. My suggestion is we give this a satisifed review, and ignore that Chromium may/may not ship it. We can encourage them to ship it by being satisfied.
DanA: OK
Matthew to post after call
Discussed
May 12, 2025 (See Github)
DanA: I think we should post Matthew's comment and close this.
Matthew will do so.
We agree to close as satisfied
<!-- PRs -->
Discussed
May 19, 2025 (See Github)
DanA: Pending comment from Matthew.
Hadley: seconded.
Jeffrey: Any objections to Matthew posting his comment?
No objections noted.
Matthew to post comment
Discussed
May 19, 2025 (See Github)
DanA: Pending comment from Matthew.
Hadley: seconded.
Jeffrey: Any objections to Matthew posting his comment?
No objections noted.
Matthew to post comment
Comment by @matatk May 21, 2025 (See Github)
Hi @MrPickles, and thanks for your review. We are happy to see this work proceed. We also appreciate that several browsers are shipping bounce tracking mitigations, so this spec work is welcome.
OpenedFeb 27, 2025
こんにちは TAG-さん!
I'm requesting a TAG review of Bounce Tracking Mitigations.
With browser vendors now actively working to remove third-party cookies from the web, some platform trackers are moving to bounce tracking. This technique involves navigating to a tracker domain at the top level of a browser tab, setting a first-party cookie or storing data in the HTTP cache, and then quickly redirecting away using a request that encodes the value of that first-party cookie or contents of the HTTP cache. Bounce tracking semantically functions like setting a third-party cookie. This spec outlines a proposal for mitigating the privacy impact of bounce trackers.
Further details:
You should also know that...
This is intended to only cover "bounce tracking mitigations" which is one part of the
nav-tracking-mitigations
repository. (The Privacy chairs asked for it to be included this repo and due to Bikeshed tooling support it became a single document. Please disregard other parts of the document other than the section on Bounce Tracking Mitigations.)This tag review is a continuation of https://github.com/w3ctag/design-reviews/issues/862. Since then, the spec has evolved to also look for "stateless bounces" (in other words, ignoring the requirement for cookie access) to prevent usage of the HTTP cache as a means to store data. Additionally, Mozilla is positive with the changes.
Note that there are two explainers: one for the original feature and another to explain a modification. Not all of the spec has not been merged and exists as a pull request at the time of writing. Apologies in advance for the inconvenience.