While you can change the server every second, it still has to look roughly the same to the user. If they try changing the name of the HTML elements on every page load, Rapportive just needs to detect based on position on the screen.
If they start looking at the DOM or making behavioural tests to see if Rapportive is loaded, then Rapportive just needs to make sure the browser returns the same results as when it's not loaded.
They could even run it in another process, and just position a window over the right area.
User agents (tend to) run on general purpose computers with no lock down. If you control the user agent, you win the detection war.
Lawyers are Google's own means of effectively stopping Rapportive, but the PR fallout might be interesting to watch.
If they start looking at the DOM or making behavioural tests to see if Rapportive is loaded, then Rapportive just needs to make sure the browser returns the same results as when it's not loaded.
They could even run it in another process, and just position a window over the right area.
User agents (tend to) run on general purpose computers with no lock down. If you control the user agent, you win the detection war.
Lawyers are Google's own means of effectively stopping Rapportive, but the PR fallout might be interesting to watch.