This paper firstly provides a brief overview of the research on Statistical Software Testing (SST) associated with the Sizewell B (SXB) Primary Protection System (PPS) Test Guardline. In this context the work has been known as Generic Dynamic Testing (GDT) based on the development of the normal dynamic testing used as an essential element of system validation. The GDT research programme established the feasibility of statistical system testing of plant protection systems containing software. Statistically sound testing has enabled quantitative measurement of software reliability for an industrial system, namely, the SXB PPS. The work was based on enhanced methods of input environment simulation and an established reliability estimation model. In the broader sense the study raised a number of significant questions that need to be addressed to resolve residual uncertainties and develop improved estimation models before the concept can obtain general acceptance and widespread application. Current work supported by the Nuclear Safety Research Programme is addressing these issues and the second part of the paper outlines the ongoing research and describes some of the ideas being developed. These include the development of enhanced reliability estimation models which include knowledge of the software structure with links to coverage testing concepts, the development of cost effective oracles and a consideration of the requirements for continuous control systems. The aim is to develop guidelines which will influence design options and make systems amenable to SST. It is foreseen that the main area of future application will be to widely used components such as smart sensors, ASICs and PLCs.