New numerically robust algorithms are presented for converting linear continuous-time constant-parameter state models into equivalent discrete-time state models (discretization) as well as the reverse problem of determining continuous-time models to represent given discrete-time models (continualization). Two methods of discretizing linear uniformly sampled systems have been considered for their utility in computer-aided design. These methods are the standard zero-order hold method which assumes that inputs are held constant at their previous sample value for the duration of the sample interval, and a method which assumes that the inputs are linearly interpolated between samples.