You aren't at all interested in knowing how common the Earth is? Whether the process that lead to the Earth orbiting the sun happens only rarely? What other planets are like? Do many (any?) of them have life, or something like it? If they do, what form does it take, is it like us? While Kepler won't answer all these questions, it's a small, but significant step in a long-term plan to address all of these questions. Even if we never are capable of ever going to any of these planets, I'd still like to know the answers to the above questions. And, as another poster mentioned, what better way to convince people to work on technology to travel to other stars than to find a specific destination to go to?

You're right regarding the importance of the atmosphere. In principle you could measure the atmospheric composition through transmission/emission spectroscopy at primary/secondary transit, however this is not feasible in the short term (maybe with 30m class telescopes it could be done). They'll have a hard enough time measuring the masses and securely confirming that any particular one of these things is planet (note the trouble with accurately determining the mass of Corot-Exo-7b for which the velocity semi-amplitude of the star is ~10 meters per second, whereas for the habitable Earth-size planets Kepler is looking for it's ~10 centimeters per second). But Kepler really isn't designed to tell us about particular habitable planets, rather it's designed as a census to figure out what fraction of stars have planets comparable to the size of the Earth orbiting at distances of an AU or so. Knowing that number will have a big influence on how we design future missions to find/study other planets that are really like the Earth (Earth-like climates).

Maybe I'm missing something, but why are they comparing the total number of convictions in two years (2008 and 2007) to the total number of convictions in 1 year (2005)? Isn't this more like a 145/40 = 3.6-fold increase in the conviction rate between 2008 and 2005 not a (145 + 100)/40 = 6.1-fold increase?