I had been thinking about writing something about Facebook for a quite a while. Cory Doctorow's recent article prodded me just enough to actually get around to it. His basic premise is that Facebook, like all social networking sites, is doomed by its very nature to fail. I disagree. This is in part a response to that article, and in part an exploration of my own thoughts.
email notifications and message editing
Doctorow takes issue with Facebook's "steady-stream" of content-free email notifications. I will acknowledge up front that this is a tactic that I often find annoying with other services. Sure, they're primarily there to draw people back to the site and drive up the hit count. But I already refresh Facebook regularly in my browser or on my iPhone, and I don't think this is an atypical usage pattern. With the high frequency and typically diminutive amount of content in most Facebook updates I really don't think email is an ideal delivery mechanism anyway. Viewing the updates in context on the site is a much better experience. I've turned off email notifications for almost everything. But I do like that there are a decent amount of orthogonal email notification options. Perhaps the primary issue is that so many of them are on by default.
He also criticizes reading and writing of messages via Facebook. My god, how can this man not be aware of just how many hordes of people use a web-based email system as their primary user agent! I'm not saying it's pretty, and I'm not saying that it doesn't fill me with nausea at the very thought of it, but the masses of people that will collectively decide the fate of social networking sites like Facebook seem to be quite content with—or at lease acquiescent to—this means of composing messages.
Doctorow's dig at Eudora as a comparably poor messaging environment is flatly ridiculous. I have been trying for ages to find a better email client. The list of things that Doctorow slams Eudora for—composing, reading, filtering, archiving, and searching—are some of the very features that have made it impossible for me to give it up when I fail to find adequate implementations of one or more of them in other user agents. Sure, Eudora has it's share of bugs, quirks, and idiosyncrasies, but for being an end-of-lifed app, it still kicks the butt of everything else I've seen, on any platform. I invite anyone to please, please prove me wrong about this!
This is always a tricky one, but as far as I know there currently isn't anything approaching a feasible business model for sustaining a no-pay, large-scale, ad-free, information-based service. At least not one that would fly here in the States. So far Facebook has done its advertising tastefully and hasn't yet done anything overtly loathsome. Yes, Facebook provides a useful service, and yes ideally it would always be completely free in all ways. But developing new features, providing support, and maintaining an increasing infrastructure sure as hell ain't free for them, so there's got to be some quid pro quo. And I'd be surprised if anyone thought that making it a for-pay service would do anything but immediately undermine its popularity and thus utility.
As conflicted as I am about the motivations of the company that pays me to write code, I've got to say that if I've got to look at advertising I would much prefer well-done targeted marketing to an indiscriminate barrage of repulsive spam. (BTW, I clicked on the link he provided in his article to a Times Online article, "Facebook shrugs off privacy fears with plan for targeted advertising" and got a both a pop-up ad window and a floating css ad window in the middle of the page that I had to dismiss in order to read the article it was obscuring. Sure makes the sidebar banner ads and occasional, innocuous news feed ads on Facebook seem a lot less objectionable to me.)
laws and cluesticks
Doctorow alludes to the possibility that social networking sites may be more subject to Brook's Law than Metcalfe's Law, but he makes no further effort to explain or substantiate this. Red herring or laziness? How does adding people to your social network impede the intent behind having that social network? Hell, as I think about it, even the invocation of Metcalfe's Law is suspect. Certainly it applies to a social network as a whole, as the greater number of subscribers there are the greater the pool of people you can potentially add to your list. But once in the milieu of individual friend lists, the mechanics become radically different. You're no longer talking about point-to-point networks, you're talking about broadcast networks. Each node is not just a potential edge, it is an edge. And this situation starts to introduce problems, as Doctorow points out with with his "boyd's law" examples. But this is not the problem itself, it's a symptom of it.
People have been changing the way they interact with social networking sites. We've been adapting our habits and sites in turn have been adapting their features to accommodate the varying levels of comfort we have with sharing information about ourselves. I think Doctorow really needs to more deeply explore why people do use social networking sites.
the rejection problem
How do you effectively deal with that perennial social networking problem wherein you receive a request to add someone to your network whom you know in passing but don't really want all up in your business? You don't want to be rude, but you also want to be able to hang out candidly with your tight peeps? The crux of the rejection problem is that most social networking sites do not offer sufficient granularity to reflect the various levels of intimacy we have with different people in our lives. There are things you tell your lover that you don't tell your family, things you tell your family that you don't tell to certain friends, things you tell to some friends that you don't mention to your coworkers, and things you say to your coworkers that you don't bring up in a conversation with an old acquaintance you happen to run into on the street whom you like well enough but who is not actually likely to catch up with you sometime nebulously later for drinks somewhere. And it works the other way, too. You might mention something to someone that you haven't seen in a while that would just never occur to you to verbalize to those in your more intimate circle because, for instance, by virtue of being close to you they have already effortlessly shared the very experience you must now labor to articulate.
So, currently, rejecting someone from your social network is something akin to refusing to even acknowledge that remote acquaintance if you happen to run into them on the street. Or like snubbing a coworker who invites himself to come along to a movie that you're going to see with a bunch of other folks from the office. For most people these types of responses would be indefensibly rude or at least impolite. On the other hand, it would take an unusually audacious coworker to invite himself into your hot bi-curious orgy. Or a shockingly presumptuous acquaintance to tag along while you get yourself tested for STDs. I think here most people would understand if you were not so decorous in your responses.
If a reasonably well-adjusted person in meatspace doesn't interact with you very often or very deeply, they generally do not presume to be privy to your daily agenda much less your innermost secrets. Why should it be significantly different online? I contend that it's because the social networking technology currently available for communicating those nuances is far too coarse. A person is either in or out. Binary simply doesn't work as a social model.
Now, Facebook has already made some progress on providing a little more granularity here. First, you can outright block particular people from searching for you, seeing your profile, or contacting you. The next level is that you can designate some people as having access only to a limited version of your profile. Further, you can adjust to a fairly good degree which aspects of your profile this limited group will be able to see. Nevertheless, this limited profile itself suffers from false dichotomy. You can't create multiple limited profiles each with different settings. But this does seem like the beginnings of a tactful way to start to model typical meatspace social conventions online. An observant casual acquaintance may notice they're not seeing your "currently online" status in their feed, but they're not likely to be too upset by that as they would be have to be a bit dim not to acknowledge that they are not a part of your most intimate circle—and besides maybe you have your online status indicator turned off globally. Moreover, and this is critical, this restriction is not visible to anyone else in your list. Apart from what you explicitly reveal in communication outside of details the site itself presents, each of your friends appears to the others equally, simply as being associated with you. Yet you retain full discretion about what details and updates you reveal to whom.
But this sort of fine-grained system of disclosure can't be too difficult or unwieldy to manage, or potential users will be put off by it. I'm sure there are some fastidious folks who will absolutely delight in endlessly tweaking a massive array of cool new digital knobs that correspond roughly to the social conventions we regularly improvise effortlessly in meatspace. But I suspect that for a far greater number, a more likely scenario is closer to wanting to have full confidence that the regular updates they are posting about their wild nightlife activities are not going out to that creepy ex-boss they felt compelled to add to their list.
Here's an idea. Many social networking sites are already starting to flirt with features very similar to the now-standard ecommerce feature of "other people who have bought this item have also bought these". Why not borrow patterns from other webapp archetypes as well? Apply to social networking sites the same basic principles behind the reputation systems developed for auction sites and user-moderated discussion boards. Track the number of direct interactions between each of the members in the social network, then give users the ability to use this data to moderate what is available to each person in their list. If somebody has sent you five messages, and you have sent them only one, maybe they don't see updates about your relationship status. On the other hand, let's say you've gotten 129 messages from another friend and sent them 143; maybe each of you can chose to get an alert while you're browsing your profile and the other logs in. Getting something like this working well would certainly take some work, but I can imagine it eventually being not unlike training your email spam filter, adjusting your Amazon recommendation preferences, or your Last.fm music tastes.
sometimes it's healthy to forget
Just as in meatspace, inevitably there will be ample opportunities online for gaffes, regrettable indiscretions, awkward situations, and unwelcome intrusions. Welcome to life. Social networking sites will not somehow provide exemption from basic human ineptitude. Yet stuff online does tend to linger a little too morbidly. Those idiotic comments you posted on a major technical help site in desperation way back when you didn't have any idea how the hell to quit vi now show up as one of the top hits when your prospective dot com employer googles you. That extraordinarily cheesy poem you injudiciously plastered all dewy-eyed on your homepage in college is now archived forever in that vast series of tubes for all otherwise prospective love interests to discover and use as grounds for pursuing no further contact with you.
And you know what, I think Facebook even has a start on ameliorating this problem. Whether this is by deliberate policy or by fortuitous technical (or aesthetic) limitations, I can't say for certain, but it used to greatly annoy the completist in me that older items appeared to just fall off your profile into oblivion. But I am starting to really appreciate this and now even consider it a to be a feature. If you slip up and say something off-color at a company party, that little impropriety will likely only last as long as coworkers are interested in retelling the story. Generally after a few days, maybe weeks, it will dissipate into the countless eddies formed by the perpetual precipitation of new events. Yeah, maybe at subsequent parties at the same time of year someone will recollect that you embarrassed yourself and imperfectly describe the incident more sparsely and less precisely each time, but for the most part it's not even worth a footnote in an already unwarranted biography. A digital version of this event horizon is not unwelcome.
Facebook could certainly screw up the relatively good thing that's its got going right now. They could make poor decisions that alienate their user base. They could fail to make enough money to sustain the service without irrevocably compromising its quality. So I'm not saying that Facebook is the be-all end-all of social networking sites. I just don't think much of what Doctorow is claiming as reasons for its immanent demise are fatal characteristics necessarily inherent in all social networks. The greatest danger to Facebook is that someone might build a better one.