Recovering Pentecostal

Archive for November, 2009

You used to be considered “cutting edge” if you had an email account and checked it.

But now, even though services like Facebook and Twitter require all users to have a valid email address, email itself is falling out of use and social networking sites are becoming a primary mode of communication.

On one level, there’s absolutely nothing wrong with this: social networking sites are far more flexible and graphically interesting than “old-fashioned” email. And sites like Facebook and Twitter have allowed many, myself included, to maintain relationships and reconnect with folks from my (often distant) past.

But there is also an inherent problem. Social networking sites are increasingly more important as storehouses of contact information and platforms of communication – yet they are single points of failure. The strength of email (at least back in the ‘good old days’) was that it was a distributed service. There was no single email server out there processing the world’s email — there were many servers, each handling email for their little piece of the ‘net world. Every domain has to make provisions to handle its own email, and many otherwise incompatible systems can exchange email through a set of very basic shared protocols. The distributed nature of email meant that even if one server crashed, 99.99% of the other servers would probably still be up and running. In other words, the failure of one member of a distributed system was not enough to bring down the system as a whole. Recently, Google’s gmail service experienced some well-publicized difficulties… yet non-Google email users were still able to contact other non-Google email accounts with no problems. (How mega-services like Google are un-distributing previously distributed services is a topic for another day!)

Social networking sites are not distributed systems — they’re centralized. When twitter goes down, it’s down for everyone. When Facebook goes down or (perhaps worse) implements a sweeping change (that will almost never be received well), all users feel the impact. We are empowered or crippled at the whim of developers, hackers, and hardware failures. Just ask any T-Mobile Sidekick customer if they ever stopped to consider the danger of entrusting too much data to a centralized system before the now-infamous crash.

I’m the proud owner of a Palm Pre. One of the Pre’s strengths is its contact aggregation from sites like Facebook. In the dark ages before smart phones, if I needed to contact someone by phone, I might have looked up their published info on Facebook or some other ‘net resource, then manually created a contact for them in my cell phone. And then it would just sit there — never to be updated again unless I tried the number later on and found it to be disconnected. But my Pre is smart. It grabs all of my Facebook friends, and if they have phone numbers listed in their profile, those numbers are automatically part of my contact list. When they change their phone number on Facebook, my phone is also updated. It’s a wonderful feature…

Or maybe it’s a nightmare waiting to happen. Let’s imagine that something catastrophic happened at Facebook, a glitch that caused all phone numbers to be changed to, say, 555-1212. My Pre would dutifully update my contacts with this bogus info. We’d have to wait either for a) Facebook to fix the glitch, or b) individuals to realize that something was wrong and correct their own data individually. (Incidentally, the flood of folks trying to connect to change their info would probably bring Facebook down, and the resulting Tweetstorm would likely render Twitter hopelessly lagged and unresponsive…)

Facebook doesn’t allow you to export contact info, so there is no easy way to keep a backup copy of your friends’ information. (There are some hacks out there, but they seem to be shady at best.)

So what would we do if we woke up tomorrow and Facebook was down, perhaps indefinitely… perhaps forever?