Massive Facebook outage caused by improper error handling, not improper Facebook flirting as previously feared

This has not been a good week for social networking. First there was that security hole on Twitter's site, and yesterday time stood still (and unsightly error messages popped up all over the web) when Facebook experienced its worst outage in four years. According to the company, the problem stemmed from an automated system that checks for invalid configuration values and replaces them with correct values from a database. A great system, to be sure -- as long as the database itself is error-free. If it isn't? Incorrect values beget incorrect values, until the database cluster enters a feedback cycle that can only be remedied by shutting down the entire site. Which is exactly what happened -- for something like two and a half hours (or the total running time of The Social Network, plus an additional 30 minutes to get some popcorn and find a seat). On the other hand, The Atlasphere -- the world's first social networking and dating site for "admirers of The Fountainhead and Atlas Shrugged" reported zero downtime whatsoever yesterday.