I'm not sure if everyone noticed, but the primary interesting thing here is NOT that it blocks off domains at certain times (everyone and their dog wrote such scripts for themselves). The nice thing is that each rule is defined as a function, which means that you can put a bit more complex rules inside, or just add some logging and track every name resolution against given domain.
Which actually brings me to question: aren't domain-IP pairings cached by the browser?
Yes, but it modifies the response to have a very low TTL so that it's cached only for a show while (in as far as browser actually get the query response).
Which actually brings me to question: aren't domain-IP pairings cached by the browser?