Phil,

As you probably know, Kepler works by measuring a short term variation in a star's magnitude (light intensity) due to a planet transiting the star's disk. Planetary transits are fairly rare events. All of the spectral class F-G-K stars in Kepler's field of view could have "Earths" orbiting them but Kepler might only detect a tiny fraction of them. Part of the data reduction for Kepler would be a statistical analysis to estimate the probable number of stars with planets based upon the number of actual observations by Kepler.

The Terrestrial Planet Finder (TPF) uses a completely different approach from Kepler. If there is a planet orbiting a star within 50 light years then TPF will likely detect it (no catch-as-catch-can like Kepler). However the "brass ring" concerning TPF is its spectroscopic capability. If TPF detected a planet orbiting a Class G2v star at 1 AU radial distance, that would be very interesting but would not demonstrate a viable world (it could be a Venus). However if the spectroscopic data indicated the planet's atmosphere had free oxygen (ozone), water vapor and methane then it would be a clear indication the planet harbored life. This capability is why TPF is one of the most exciting astronomy concepts currently proposed.

I should add that Kepler is a neat idea and I'm very please that it has been funded. In terms of scientific bang/buck, Kepler is one of the best astronomical satellites ever proposed. Quite frankly, I'm mildly annoyed that it took NASA so long to fund Kepler since it's such a great idea.

Gary