So after thinking a lot about FLOC and how it impacts and changes the advertisement environment, I have more questions than answers.
And at least from my current understanding, it becomes worse for organisations that want to advertise, which might includes your government trying to educate citizens about things, while not really making things much better for individual privacy, due to the ability to correlate cohort IDs to identify users.
Mhm ._.
While a lot of people decided to set a header to opt-out from FLOC tracking for their users, I decided to embrace it and let people make up their own mind.
For those interested in the banner, here are the implementation details:
It's kept simple and efficient. It'll only show up, when your browser implements the API. Enjoy!
@sheogorath can you do both or would the opt-out change typeof document.interestCohort?
@silmathoron that's something I wasn't able to test yet. It's quite hard to come by a browser than actually does FLOC when being in Europe and not even using Google Chrome in first place.
(So far all browsers I tested were negative in both, my own test and EFFs amIFloced.org)
@silmathoron I didn't have the time yet, but if you want, you should be able to verify it using the method from this toot:
https://microblog.shivering-isles.com/@sheogorath/106075090545371556
@sheogorath Actually, I stumbled on that piece which made me wonder whether the banner is actually worth it...
https://seirdy.one/2021/04/16/permissions-policy-floc-misinfo.html
An it seems to answer the previous question by a negative (can't do both your test and the permission header)
@silmathoron that's one of the reasons I decided against adding the header for my page. At least according to the standard, it's not needed. However the banner on the other hand, helps to make people aware of the browser behaviour and this way might help people to make a sane decision if they want to participate in this experiment.
The banner itself is designed to only check if the browser implements FLoC, it doesn't use it or estimate whether it's active.
@sheogorath ah, you're right, it's calling the function that would be a problem
