Abstract
Inflation produces nearly scale-invariant scalar and tensor perturbation spectra which lead to anisotropy in the cosmic microwave background (CMB). The amplitudes and shapes of these spectra can be parametrized by QS2, rQT2/QS2, nS, and nT where QS2 and QT2 are the scalar and tensor contributions to the square of the CMB quadrupole and nS and nT are the power-law spectral indices. Even if we restrict ourselves to information from angles greater than one-third of a degree, three of these observables can be measured with some precision. The combination 105S1n QS2 can be known to better than ±0.3%. The scalar index nS can be determined to better than ±0.02. The ratio r can be known to about ±0.1 for nS≃1 and slightly better for smaller nS. The precision with which nT can be measured depends weakly on nS and strongly on r. For nS≃1, nT can be determined with a precision of about ±0.056(1.5+r)/r. A full-sky experiment with a 20 arc min beam using technology available today, similar to those being planned by several groups, can achieve the above precision. Good angular resolution is more important than high signal-to-noise ratio; for a given detector sensitivity and observing time a smaller beam provides more information than a larger beam. The uncertainties in nS and r are roughly proportional to the beam size. We briefly discuss the effects of uncertainty in the Hubble constant, baryon density, cosmological constant, and ionization history.