Refactor to separate uses of IP #38
Conversation
|
I don't think the geolocation use case really makes sense as either tracking or anti-abuse. |
| determining "reputation" {{!RFC5782}} in conjunction with other signals to | ||
| Tracking of IP addresses is common place on the Internet today, and falls | ||
| roughly into two broad categories. The first is profiling, the creation of | ||
| interest profiles for a given user to serve them relevant advertisements. The |
There was a problem hiding this comment.
If we define profiling just in terms of advertising then I tend to agree with @bakkot that geolocation doesn't fit well in either bucket. May I suggest that the category be personalization, of which profiling for ads is one example of personalization. Other examples would be tailoring content to your location (such as local news, or local stores).
|
Good point, I've replaced profiling with personalization. Let me know what you think |
| ([ODoH]({{?I-D.pauly-dprive-oblivious-doh}})) and Oblivious HTTP | ||
| ([OHTTP]({{?I-D.thomson-ohai-ohttp}})). While they both prevent tracking by | ||
| individual parties, they are not intended for the general-purpose web | ||
| browsing use case. |
There was a problem hiding this comment.
Should probably expand on this to explain why they're not a general purpose solution. This can be a follow up bug though.
There was a problem hiding this comment.
Yeah that text predates this refactor so let's address that separately.
As proposed in #36, this PR attempts to differentiate uses of IP into two broad categories: personalization and anti-abuse. It then refactors some sections to leverage that separation. Some further tweaks will be needed after but I think this improves how the document reads.
Here is a rendered view of this PR.
Fixes #36