top | item 12254765

(no title)

rmi_ | 9 years ago

I am really interested in the technical details. I think we have yet to see a working (and practical) approach to this.

discuss

order

Cthulhu_|9 years ago

I can think of a few simple tricks already. One: Host from the main facebook.com domain, either inserting it in the basic html page response from the server, or from an API / content location that cannot be predicted. (ad calls should be indistinguishable from 'normal' content calls). Two, do the same with content locations, so no 'div id="ad"' or anything like that. Should be easy enough.

topspin|9 years ago

"Host from the main facebook.com domain..."

If true I believe that this would actually be an important improvement. This removes some plausible deniability regarding malware and other abuse; Facebook (or whomever else adopts this scheme) must guard more carefully against abuse when the content is coming from their own domains, as opposed to some third party.

waterphone|9 years ago

The efforts I've seen so far from other websites include randomized IDs and class names, and base64 encoded images inlined so the file path/hostname can't be used as a parameter for blocking.

juliand|9 years ago

Me too. This is clearly a cat vs mouse game or more like a lion vs mouse and that's what makes it interesting to me.