On Wil Wheaton, twitter links, and Shiny Object Syndrome
According to CNN, only about 27% of twitter users are 'active' users, meaning they follow more than 10 people, have more than 10 followers, and have tweeted more than 10 times.
Wil Wheaton has ~1.6 million followers. Let's assume that 27% are active.
Now let's pull a random number from nowhere and say that only 10% of those users are 'hyper active,' i.e., likely to follow a link within the first three minutes of a post, because face it Wil Wheaton does draw a lot of nerdy, technology-addicted followers.
This means that if Wil Wheaton posts a link, that website must be able to sustain 240 hits per second on top of its normal traffic, or it will crash. Some smaller sites simply can't float that.
Let us now assume that of the remaining 90% of active @wilw followers, only half will follow the link within 5 hours, and that rate will be relatively constant compared to the initial surge. The result is a baseline traffic on that site of 11 hits per second above their normal traffic flow.
If we create a rule not to follow a link within the first three minutes, we must deal with a phenomenon that I will call 'Shiny object syndrome.' The mandate on a hyper-active tweeter to wait 3 minutes before following a link increases the possibility that a subsequent tweet will distract the tweeter, thus delaying their return time to the website in question.
Since we're astronomers, let us assume a Gaussian smearing. For integrating simplicity, we will assume that some hyper-active tweeters will violate the mandate, so integration time starts at 0 and not 3 minutes. Assuming that all of the hyperactive users return within 5 hours, we find that only 4% of the initial hyperactive users follow the link within the first 3 minutes, thus increasing traffic flow by about 10 hits per second. Adding this to the casual users, that puts a maximum strain of only 21 hits/second on a website, which is much more manageable.
Thus, I should make a rule never to follow a link posted by @wilw within the first 3 minutes of the tweet, lest we crash/flood the website!
