The cosmic microwave background (CMB) gives us a glimpse of the Universe as it was only a few hundred thousand years old. The tiny fluctuations — one part in 100,000 — that we observe in the CMB trace out the fluctuations that would eventually become the galaxies and clusters that we see today. Our theories, on the other hand, predict only the statistical properties of those initial conditions. Measuring cosmological parameters therefore requires disentangling those statistical properties from our finite, noisy observations. Today, those observations (from instruments like the COBE and WMAP satellites) tell us that the universe is geometrically flat, and consistent with the predictions of an early period of cosmological inflation. However, they do not yet tell us whether the Universe may be, say, topologically connected. Moreover, current data don’t yet help us choose between the theories that underlie the observed phenomenology: why does this particular set of physical laws obtain? Can we hope that future data from experiments like ESA’s Planck Surveyor will help us physicists answer these questions?