While i was away, someone pointed me to a paper called Warhol Worms: The Potential for Very Fast Internet Plagues. The paper documents how a worm gets distributed through the Internet and how its exponential growth caps out at around 1,000,000 people. I was encouraged to read this because one question that’s been on my mind is how long Friendster growth will continue and whether or not it will just cap out. For a while, it was growing at 10% per week, but this hasn’t been maintained all summer.
While i don’t have the data about signup dates, i did trapse back and look at join dates for a variety of numbers, just to get a sense of when certain numbers were hit. Unfortunately, this isn’t as valuable as the actual dates, but it’s interesting none-the-less:
User 101 (Jonathan) signed on in April 2002.
User 175 signed on in May 2002.
User 250 signed on in June 2002.
User 375 signed on in July 2002.
User 500 signed on in August 2002.
User 1000 signed on in September 2002.
User 1250 signed on in October 2002.
User 1500 signed on in November 2002.
User 2500 signed on in December 2002.
User 5000 signed on in January 2003.
User 10,000 signed on in February 2003.
User 25,000 signed on in February 2003.
User 50,000 signed on in March 2003.
User 100,000 signed on in April 2003.
User 250,000 signed on in May 2003.
User 500,000 signed on in June 2003.
User 750,000 signed on in June 2003.
User 1,000,000 signed on in July 2003.
User 1,250,000 signed on in July 2003.
User 1,500,000 signed on in August 2003.
User 1,750,000 signed on in August 2003.
[As of August 31, there are 1,853,799 User accounts]
Needless to say, Friendster is well beyond the 1M marker, but it’s hard to tell how much beyond. While there are close to 2M accounts, i have no sense of how many of them are fake or invalid. Plus, i keep running into people who have multiple accounts (same information, same name, all “real”). I can’t tell if this is because they forgot their login or because they want to separate different friend groups.
Also, so many of the account seem to be inactive. I still hypothesize that people have an interest span of around 2 months from when the last group of friends joined. ::sigh:: Of course, without the data, it’s hard to test my hypotheses.