I have not read all the comments in this thread, but I have read the comments accompanying An approach to fair ad blocking - Feedback summary
I have noticed that many people (like me) only object to intrusive advertising. The problem is that intrusive advertising is difficult to detect automatically: how do you know if that chunk of JS code is for advertising or content like Google Maps? One of the commenters linked above pointed out that even static text-based Ads or images can be intrusive if they are misleading.
Another thing to consider is bandwidth usage: I realize a lot of people don't bring it up, but it is a legitimate concern. When you are listening to the radio or watching TV, the Ads use about the same amount of bandwidth (boosting the volume during commercials uses a little
more bandwidth) as the content. If I am reading a text article that is 10KiB, but also end up downloading a 1MiB video Ad, the Ad is using 100x the bandwidth as the content. If the article is 250kiB of text and images, 100kiB of banner Ads are not so out of place.
The point of that anecdotal story is that intrusive advertising often has side effects: side effects that can be measured.Proposal:Add an option to automatically block Ads if more than 25% of the CPU(s) is/are in use or if the size of the Ad content is likely to be disproportionate to the size of the non-Ad content.
Such functionality will likely require a "grey list" of known Ad sites or simply assuming third-party sites are Ad hosts. I realize the filesize is not always known before download with HTTP, but you can probably guess based on content type (assuming the server doesn't lie about that too). If a server lies about content type, it should be automatically black-listed.