3 ms·

> It is clear that browsers are doing everything they can to prevent user abuse. If that was true in general, my browser would block all cookies and all conten

> It is clear that browsers are doing everything they can to prevent user abuse. If that was true in general, my browser would block all cookies and all content from third party domains by default. It would systematically misreport Referer and User-Agent. It would disable APIs typically used for fingerprinting. It would disable all redirect-through-advertiser links. All managed by a simple and unintrusive whitelist (per first party, per third party or temporarily for a whole browser session or for a site session) so that you can opt in to these things where there is a legitimate use. It isn't because it has a much greater interest in browsing being a seamless experience and the web being a powerful application platform, and actual privacy measures are buried in some configuration menu or in third party plugins so that whiny power users can enjoy privacy while the unwashed masses go about their browsing as though the web standards aren't completely messed up from a privacy perspective. > This provides a mechanism to do the required attribution without also tracking or compromising a user’s privacy. Without that, there isn’t a way for the current ad model to work. Is that a good or bad thing?

olliej 7y ago
Safari already segments cookies, storage, etc such that the ability to use those to track people is limited. I see no reason to believe it’s not going to continue making tracking harder. The user agent string does not provide as much information as people seem to believe. Whitelists aren’t usable, and don’t work at scale. The whole point of breaking tracking in general is so that the “unwashed masses” get privacy. Every study and survey has found that users value their privacy, but have been told (by advertisers and I assume you) that privacy is dead, or that they cannot have the internet without it. It is absolutely not something that just “power users” care about. It’s just power users are the only ones who currently get any. Finally, the attribution isn’t attributing to a specific user - the entropy is heavily limited, the reporting is time delayed by a random amount, and is made with a stateless session so there isn’t any way to directly tie it to a single specific user, leaving just the IP address, which is now typically shared, and frequently changes.
- boomlinde 7y ago
  > Safari already segments cookies You said browsers in plural, though. * >The user agent string does not provide as much information as people seem to believe.* How much information do people seem to believe that it provides? > Whitelists aren’t usable, and don’t work at scale. I don't need it to "work at scale". I need to be able to turn blocking off occasionally. Also, "at scale" and "usable" are vague to the point of meaninglessness. Why aren't they usable? What scale does a whitelist need to work at that makes you say that it can't? > Every study and survey has found that users value their privacy, but have been told (by advertisers and I assume you) that privacy is dead, or that they cannot have the internet without it. It is absolutely not something that just “power users” care about. It’s just power users are the only ones who currently get any. That's my point exactly, because privacy isn't the browser vendors' top priority, nor are they generally doing "everything they can" to improve it. The standards don't even allow it. > Finally, the attribution isn’t attributing to a specific user - the entropy is heavily limited, the reporting is time delayed by a random amount, and is made with a stateless session so there isn’t any way to directly tie it to a single specific user leaving just the IP address, which is now typically shared, and frequently changes. "It's not so bad, really, for most users" is not the same thing as "browsers are doing everything they can to prevent user abuse." Add, I don't know, User-Agent and Referer to the mix to more reliably track different sessions on the same address. My address doesn't change frequently. I'm on a temporary lease but when it expires I get the same IP again or at least have since I started my account. This is not uncommon for a broadband connection in my whole country. If you consider tracking individual users' behavior for ads user abuse, let me remind you that one of the largest perpetrators of that are working on their own browser and regularly get a say in how the web works.

olliej 7y ago

Safari already segments cookies, storage, etc such that the ability to use those to track people is limited. I see no reason to believe it’s not going to continue making tracking harder. The user agent string does not provide as much information as people seem to believe. Whitelists aren’t usable, and don’t work at scale. The whole point of breaking tracking in general is so that the “unwashed masses” get privacy. Every study and survey has found that users value their privacy, but have been told (by advertisers and I assume you) that privacy is dead, or that they cannot have the internet without it. It is absolutely not something that just “power users” care about. It’s just power users are the only ones who currently get any. Finally, the attribution isn’t attributing to a specific user - the entropy is heavily limited, the reporting is time delayed by a random amount, and is made with a stateless session so there isn’t any way to directly tie it to a single specific user, leaving just the IP address, which is now typically shared, and frequently changes.

boomlinde 7y ago
> Safari already segments cookies You said browsers in plural, though. * >The user agent string does not provide as much information as people seem to believe.* How much information do people seem to believe that it provides? > Whitelists aren’t usable, and don’t work at scale. I don't need it to "work at scale". I need to be able to turn blocking off occasionally. Also, "at scale" and "usable" are vague to the point of meaninglessness. Why aren't they usable? What scale does a whitelist need to work at that makes you say that it can't? > Every study and survey has found that users value their privacy, but have been told (by advertisers and I assume you) that privacy is dead, or that they cannot have the internet without it. It is absolutely not something that just “power users” care about. It’s just power users are the only ones who currently get any. That's my point exactly, because privacy isn't the browser vendors' top priority, nor are they generally doing "everything they can" to improve it. The standards don't even allow it. > Finally, the attribution isn’t attributing to a specific user - the entropy is heavily limited, the reporting is time delayed by a random amount, and is made with a stateless session so there isn’t any way to directly tie it to a single specific user leaving just the IP address, which is now typically shared, and frequently changes. "It's not so bad, really, for most users" is not the same thing as "browsers are doing everything they can to prevent user abuse." Add, I don't know, User-Agent and Referer to the mix to more reliably track different sessions on the same address. My address doesn't change frequently. I'm on a temporary lease but when it expires I get the same IP again or at least have since I started my account. This is not uncommon for a broadband connection in my whole country. If you consider tracking individual users' behavior for ads user abuse, let me remind you that one of the largest perpetrators of that are working on their own browser and regularly get a say in how the web works.