- From: Kevin Babbitt via GitHub <noreply@w3.org>
- Date: Fri, 05 Dec 2025 21:25:28 +0000
- To: public-css-archive@w3.org
While I empathize with the notion that tainting is not a great author experience, it's at least a concept we have precedence for, in the `attr()` function and with `<canvas>`. Exposing a lower-granularity accent color seems to me like it would satisfy neither the privacy concerns (there's still fingerprinting entropy, just less of it) nor authors who want to derive colors from it ("this CSS function I wrote for making tints usually works great but gives weird results with AccentColor, why?").
I do think it makes sense to taint only when there's fingerprintable data present, and that authors should be able to escape tainting by setting `accent-color` to a fixed value.
@kyerebo when we discussed on the call this week, one of the asks was for more detail on how exactly the tainting would operate. Do you want to propose some specifics for that based on your prototype findings?
--
GitHub Notification of comment by kbabbitt
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/10372#issuecomment-3618652667 using your GitHub account
--
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 5 December 2025 21:25:29 UTC