Re: [w3ctag/design-reviews] Specification review for fenced frames (Issue #838)

lolaodelola left a comment (w3ctag/design-reviews#838)

Hi folks,
Thank you for your patience here. We've discussed this extensively and have decided resolve this proposal as unstatisfied.

We appreciate the effort to preserve user privacy on the web, however we think all of the proposed [use cases](https://github.com/WICG/fenced-frame/blob/master/explainer/use_cases.md) involve sending too much information between sites, which risks re-enabling cross-site tracking.

While Fenced Frames does marginally improve Protected Audience, PA's unsatisfied review was only in part due to not enforcing fenced frames, so it doesn't justify shipping fenced frames. 

We can see that you’re trying to block cross-site tracking for `SelectURL`, however, since the goal is for the URLs to cause users to behave in observably different ways, observing that difference in behaviour can reveal the cross-site data that drove the URL choice.

Additionally `window.fence.notifyEvent()` allows sending too much information cross-site without a reasonable expectation that users will understand that's what's happening.
We’ve also been following the conversation in WICG/turtledove#990 and while we can see the provision for rate limiting the number of fenced frames, and the [Intent to Experiment for Partitioning :visited Link History Phase 2](https://groups.google.com/a/chromium.org/g/blink-dev/c/qbHl0W-75zA/m/GzieAsJJAAAJ) was resolved as satisfied by the TAG, we would like to see mitigations for the captcha style attack as we believe this to be an important attack vector to protect users from abuse.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/w3ctag/design-reviews/issues/838#issuecomment-3253510774
You are receiving this because you are subscribed to this thread.

Message ID: <w3ctag/design-reviews/issues/838/3253510774@github.com>