Okay, this is getting out of hand. And by this, I mean everyone crying wolf at Facebook. I'm none to pleased about the way the site is going these days either, but IP addresses = "public information"? Did someone forget how email headers work?
For those who have never read a mail header, I just tested a message sent through GMail's SMTP server from OS X. Not only is my public IP address in there, but my "inside the NAT" address is on it as well.
I suppose we must add Google and Apple to the IP address leaking villain list.
There is no reason any Facebook user (including hackers) would expect this information to be available.
Scenario:
I would like to know your IP.
I sign up for a throwaway FB account (perhaps through a proxy), send you a message designed to evoke a response (even if it’s “Who is this?”).
I have your IP.
Typically websites will only reveal users' IPs to administrators, moderators, owners, and/or peers when direct P2P is involved.
I also highly doubt this practice is described in their privacy policy.
I used to curate a small social site whose topic was somewhat sensitive. IP address exposure became a significant concern with the population. An earlier administration had left it on. Particularly as the site grew and became less "cozy", and as the user base become more informed on the topic of IP addresses, users including old-timers used to the extant exposure became worried. When control started to be passed off to me and a few others, we turned it off.
A significant fraction of the lack of concern in such situations is due to lack of knowledge and understanding. Once users -- users in general, and not just the "technophiles" -- become aware of the situation and potential issues, a significant fraction DO have concerns are very explicitly DO NOT want the "features"/exposure.
A specific example of the latter. Our site was, aside from IP addressing, totally anonymous. Pick a pseudonym, create an account, and you were in. If you caused problems, you would be warned and, if you didn't get a clue, eventually banned.
There was, as there is in most populations, a subset of particularly manipulative users, some with what I guess might be described as "borderline personality issues". One of these went on a kick, including possible multiple identities, causing me to dig in a bit. The combination of IP address -- with attendant locality -- and a reference an unusual hobby -- meant that in under five minutes, I could have driven up to her doorstep. (And, BTW, that would have been in a rural location where help is not a few minutes away.)
Of course, I didn't use the information that way, but it was useful in reminding her, in a suitably oblique and non-threatening but pointed enough message, that her shenanigans were a bit more transparent than she might have thought.
The premise that "privacy is dead" and that there is no real value for users in maintaining it, is completely fallacious. I've personally witnessed, very often, very significant, meaningful exchanges occur in the context of and precisely because of such privacy and anonymity. There are sensitive issues that you may NOT share in common with your neighbors, family, friends. Even when some of those people are supportive, they may not share the actual experience, nor be sources of useful information and counseling on how to address it. And not infrequently, those people are actually the source of problems. And they often know each other, so that if you speak with one, it influences other relationships.
People on the community I curated did expose their real identities to each other. But it was usually in private -- out of band -- via email, IM, and the like. And it was after getting to know one another within the context of the community and coming to trust each other enough to do so.
If a stranger walks up to you and asks you for your phone number, it's unlikely you will provide it -- not solely on that basis. Online social networks need similar contexts. You exchange personally identifiable information if and only if it suits you and AFTER you've established some trust.
ADDENDUM: Facebook, per its terms of service, does insist on a real name. Depending on one's name, that is more or less unique. But it doesn't (so far, at least) insist on your location (or maybe I've forgotten this, having entered a very "vague" big city name, myself) nor necessarily expose it as text. They turned off privacy for profile pictures, but as of this point at least, you don't have to provide one or a picture that is of you.
Their message facility implies privacy. It's a principal means for communicating with another user who is not a friend. Exposing IP addresses can provide fairly specific locality. (Comcast DNS entries are one example I can think of that narrow things down considerably.) The stalker potential is not inconsiderable. (And, to bring the topic of forced public profile pictures up again, combine this with a pretty picture, and you may have one source for a potential problem.)
In the past, social communities have served specifically as a layer of abstraction between users and IP addressing. Administrators had access to addresses, and so one needed to establish trust for a site. But other users did not. Kind of like going to a party. The host may have invited you, and you might wear a name tag, but you don't write your address on the tag.
Have you ever gone to a party or meeting where they surprise you with an attendee address list to be shared publicly? Feel a bit queasy when you learn of this? Start trusting the host less? Just write down your name, or name and a suitably large city name that you live in or near?
Actually, I recall something like that, a few years ago, having both U.S. citizens and Canadians. The U.S. citizens coughed up their full addresses and phone numbers. The Canadians almost all kept it to "Vancouver, BC" or similar, and if they provided any specific means of contact, it was an email address -- mostly of the gmail or hotmail variety.
Interesting, of course your ip address is normally present in the headers of your email anyway, but using a service like facebook (which has a pretty high stalking potential) you might not want to give that much away, especially not when you think that such info is safe.
Many webmail providers put the originating IP address in the headers in order to be able act later if the message turns out to be spam. Hotmail, Gmail and Yahoo all do it - I just checked. As others have said, this is not news.
Correction: Gmail does not do it when submitting using webmail, only when submitting via SMTP.
The difference between sending a message via facebook and via smtp, is that in this case facebook is appending your ip address on purpose. I mean, if I wanted to do the same with my web apps, I'd have to specifically tell my email framework to include the ip of the user.
