The light curve of supernova (SN) 1993J is calculated using two approa
ches to radiation transport as exemplified by the two computer codes,
STELLA and EDDINGTON. Particular attention is paid to shock breakout a
nd the photometry in the U, B, and V bands during the first 120 days.
The hydrodynamical model, the explosion of a 13 M-. star that has lost
most of its hydrogenic envelope to a companion, is the same in each c
alculation. The comparison elucidates differences between the approach
es and also serves to validate the results of both. STELLA includes im
plicit hydrodynamics and is able to model supernova evolution at early
times, before the expansion is homologous. STELLA also employs multig
roup photonics and is able to follow the radiation as it decouples fro
m the matter. EDDINGTON uses a different algorithm for integrating the
transport equation, assumes homologous expansion, and uses a finer fr
equency resolution. Good agreement is achieved between the two codes o
nly when compatible physical assumptions are made about the opacity. I
n particular, the line opacity near the principal (second) peak of the
light curve must be treated primarily as absorptive, even though the
electron density is too small for collisional deexcitation to be a dom
inant photon destruction mechanism. Justification is given for this as
sumption and involves the degradation of photon energy by ''line split
ting,'' i.e., fluorescence. The fact that absorption versus scattering
matters to the light curve is indicative of the fact that departures
from equilibrium radiative diffusion are important. A new result for S
N 1993J is a prediction of the continuum spectrum near the shock break
out (calculated by STELLA), which is superior to the results of other
standard single energy group hydrocodes such as VISPHOT or TITAN. Base
d on the results of our independent codes, we discuss the uncertaintie
s involved in the current time-dependent models of supernova light cur
ves.