We present an algorithm to identify the types of supernova spectra,
determine their redshift, and place constraints on their phase. This
algorithm, based on the correlation techniques of Tonry & Davis
(1979), is implemented in the SuperNova IDentification code (SNID;
Blondin et al., in prep). This code is used by members of the ESSENCE
project to determine whether a noisy spectrum of a high-redshift
supernova is indeed of Type Ia, as opposed to, e.g., Type
Ib/c. Furthermore, by comparing the correlation redshifts obtained
using SNID with those determined from narrow emission or absorption
lines in the supernova host galaxy, we show that accurate redshifts
(with a typical error $\sigma_z = 0.01$) can be determined for SN~Ia
for which a spectrum of the host galaxy is unavailable.