If you were monitoring a network of water pipes supplying water to various areas, and are worried about an outbreak of water-borne disease, you can monitor the spread of this disease by placing sensors at important junctions and pipes (instead of monitoring every single pipe).

Researchers at Carnegie Mellon University used this approach to answer the question “Which blogs should I read to be most up-to-date with important stories?”

They took data from 45,000 blogs and 10 million posts over one year (2006) and tracked 1 million links from blogs to blogs to see how information spread, and to determine which blogs were at the most important junctures, therefore worth reading.

They came up with a list of 100 blogs – some of the blogs listed make sense (like Boing Boing, metafilter, TUAW, and so on) but others really don’t: donsurber.blogspot.com? He’s not even in the top 100,000 of Technorati but is listed as #2 on the list. Anglican.tk? That’s just a spam blog, guys!

Some that should be on the list weren’t: Engadget, the world’s top ranked blog by technorati isn’t there, neither is gizmodo, the second ranked blog. Where’s Huffington Post? If you include reddit as a blog to watch, then where is digg?

A big flaw in the paper (which I didn’t read carefully) is the presumption that there is a single blogosphere where in fact there are networks of blogs that don’t actually link to one another. You’d expect political blogs, which dominate this list, to link to other political blogs, but not to technology blogs. Which “blogosphere” you end up monitoring depends on which blog you pick first.