Re:Two types of privacy
by atulsnj on Monday June 27, 2022 @02:38AM
(#31891310)
I always thought that since I opted for DO NOT TRACK I am not being tracked, now that it is a fingerprinting vector, it feels like being duped, and BTW if WebKit removed it then why not Firefox, I mean is there any good reason to have it anymore?.
[ Parent
| on HN ]
Some analytics software respect DNT headers, like Matomo by default [0] so it might still be useful.
[0] https://matomo.org/faq/general/configure-privacy-settings-in...
Re:Two types of privacy
by imwillofficial on Monday June 27, 2022 @03:35AM
(#31891624)
Relying on the good will of the advertising industry seems to be a poor choice.
posting from a throwaway account -- I worked at a few audience measurement companies, DNT was never looked at -- completely ignored. The reason was simple: all upside to ignore, no downside. Regarding fingerprinting -- I know firsthand the places I worked at considered it, and rejected it. Problems are: too cumbersome, they do not persist over long enough periods to be useful to the industry, and they cannot be shared with other companies in the data marketplace. An adversary can use them short-term, but if an adversary has the data to do that, they probably have additional more accurate means at their disposal. Also, in my view, academic papers on this subject always seem alarmist and naive, and methodologies have problems (lack scale, have selection bias, lack ground truth, overlook persistence over time, etc.)
{edited} - the way DNT worked in practice led to a very misleading, and harmful effect. Standards should be designed with stakeholder incentives in mind.
Re:Two types of privacy
by hyperdimension on Monday June 27, 2022 @01:45PM
(#31898031)
Yeah, the DNT header ultimately being used as another bit for tracking seemed to me the ultimate irony.
This, as for many things: Thanks, Microsoft.