4 ms·

> OK, so the "limited" set of information is already 5 pieces of information... Ehh...it'll probably be the useragent string and the result of geocoding the cl

by ilikepi 10y ago

> OK, so the "limited" set of information is already 5 pieces of information... Ehh...it'll probably be the useragent string and the result of geocoding the client IP address. This is fairly narrow (especially for obscure and/or out-of-date browsers), but I think calling it "unique" might be a stretch. I guess it would be highly dependent on the population of your city or town and on the popularity of the site itself. I might be one of just a handful of people who reads HN while laying in my bed at night on an iPhone 6S, but I'd expect quite a few people in a major metro area are hitting big media sites from Chrome on Windows. EDIT: trivial wording tweak

yellow_postit 10y ago
If it starts including things like display size, device, etc. then you get into the realm of browser fingerprinting which while maybe not truly unique does start to narrow the pool quite a bit: https://panopticlick.eff.org/about#browser-fingerprinting https://panopticlick.eff.org/about#browser-fingerprinting
thr0waway1239 10y ago
This is what I used to think, but the diversity of the devices in the developing countries (for example) adds more specifics. Even in countries where there is a lot more homogeneity of devices - you mention HN for example, but if I install ABP on my device and visit multiple sites using the browser, remember that the list of sites is another little trail of breadcrumbs. That could only become significant if the advertisers themselves share this data amongst themselves and combine it with user-agent + geocoded address, but you can count on a service which will pop up to facilitate exactly that (assuming one doesn't already exist). Not to mention, what if these ad networks consolidate at some point in the future?