Sunday, July 4, 2004

Random Weblog

An interesting talk about weblog research last night led me to Lillia Efimova's research notes. I agree that counting comments is a terrible way to measure interactivity.

But Herring's idea of studying an array of randomly-chosen weblogs assumes that you have access to the entire universe of weblogs (or a randomly-selected subset of them) from which to choose. That's difficult, at best: how do you find the boundaries of the population? Do you count test blogs? Abandoned blogs? Someday, the majority of all blogs will of necessity be silent, because all their authors died centuries ago; does that make weblogs less interactive?

If you defined the universe as, say, everything on blogspot, you might get one answer. If the universe were "the 250 MT weblogs that were mentioned in slashdot", you might get another. If you look at LiveJournal weblogs, you'd get another. If you spider blogrolls, you'll get another answer -- and that one depends a lot on where you start your spider.

It's an interesting question: if you want a representative collection of "random" but arbitrarily-chosen weblogs, how do you find it?