Observational drivers for a Zeeman-based 3D maser radiation
transfer code are discussed. We present basic information about the code
and its methods of solving for molecular density-matrix elements. We consider
tests of the model using a long cylindrical domain that offers hope of
comparing the model with previous 1D results. Results are generated for
cases of wide and narrow Zeeman splitting, as compared to the Doppler
width. In the wide-splitting case, our model conforms with expectations
for magnetic field orientations parallel and perpendicular to the long
axis of the domain. Results in the narrow-splitting case are plausible, but
need further work for confirmation. We demonstrate the ability to generate
VLBI-style images as well as spectra averaged over a view of the domain.