My talk will draw on 27 years of experience at Cytel in developing simulation software for statistical analysis and clinical trial design and implementation. I will discuss development and testing processes that have been important in helping us provide reliable computational results in Monte Carlo and Markov Chain Monte Carlo simulations. I will highlight the role of documentation, openness of design and customization in facilitating regulatory verification and sensitivity analysis. I would like to suggest that FDA and Industry collaborate to develop suitably disguised case studies based on actual clinical trials where simulation models were a key part of the trial design. These cases would serve as vehicles for communicating good simulation practice in submissions to the FDA.