document current ethics guidelines in stats collection
Now that we've looked at various statistics we might collect, and we deployed a pair of them (total number of onion addresses, fraction of total Tor traffic that has to do with onion services), and we've chosen not to deploy many others, we're in a good position to revisit these tradeoffs we decided about.
What guidelines can we produce for future decisions based on these past choices?
This is a subset of the generic research ethics guidelines, but let's try to answer this one rather than trying to turn this ticket into the bigger topic.