This paper reviews the historical and contemporary precedence regarding the development of knowledge, its reformulation in computer codes, and subsequent application in decision making. It highlights the practical challenges of this process as it applies to the investigation of engineered systems to deliver on both promised benefits and protection from postulated failures. A model for demonstrating model content, completeness, and consistency is described, invoking and extending a knowledge/content model attributed to Popper. While the specific example examining the evolution of the thermal-hydraulic knowledge base applied for nuclear power plant safety analysis and its capture in the RELAP series of computer analysis codes is presented, the framework is general, true to the scientific method, and thus broadly applicable. It concludes that while content of our knowledge base is perpetually increasing, completeness and consistency are fundamentally unattainable; however, within a well-designed evaluation methodology, measurable proof, sufficient for regulatory deliberation, is possible.