Hacker News new | ask | show | jobs
by kmavm 5875 days ago
You're asserting what you seek to prove: that some large fraction, perhaps a majority, of Facebook users do not understand who can see their status updates. It is absolutely true, of course, that some users do not understand who can see their status updates; the law of large numbers assures us of that. So, we actually agree on the fundamentals of the situation: some users don't understand who can see their status updates. The question is, how many? What is your actual guess? 0.1%? 1%, 10%, 50%? What percentage would be acceptable, given that 0% is not achievable no matter what?

There are some complex corners of FB's privacy system; e.g., some of the ins and outs with photos are pretty subtle. However, status updates are one of the more clear areas: Q: Who can see this? A: Everyone.

Finally, with respect to the "trolling" charge, I am only human. Those HN readers who are also Facebook users, thanks for putting food on my family's table, and I'm truly sorry, and professionally humbled, if our product has let you down. However, this is not a Facebook customer support forum; it is a community of technologists, and it would be condescending not to speak to HN as if I'm speaking to my peers. I think the silence of my fellow Facebook engineers on threads like this one, while showing admirable restraint, has left the bogus impression that we do not talk, think, or care about the implications of the products we build. We do, and a searching, frank dialog with our more technically minded users, which will necessarily include argument, can help us figure out how to make things better.

1 comments

Very well put with respect to trolling. I, for one, appreciate the public dialogue. One nitpick: the law of large numbers doesn't mean what you intended here.
You're right: http://en.wikipedia.org/wiki/Law_of_large_numbers

I hope the intention was clear: with 400 million users, it is a certainty that a feature will be incomprehensible, unusable, etc., to some non-zero number of users.

Yes the intention was clear, and I agree.

I'm now wondering what the rigorous way to describe the phenomenon you describe is. Something like, as the number of realizations of a random variable increases, the probability of seeing a realization below a given threshold approaches one. This sounds a lot like some of the theory related to hypothesis testing but it's been too long.