Post #4368541

The software practice was once a mathematical activity, à la Hoare, Dijkstra, Bird, et al. But it has not been so, for several decades. If the present #software practice is to mature into a bona fide #engineering practice in the future, #mathematics-based #formal methods, notations, languages, and tools—Z, B, VDM, ACL2, TLA+, Coq, Lean, Agda, Idris, ...—must be adopted. There obviously is no shortage of such tools and techniques. Clearly, many practice areas within IT, including glossy brochure web sites, line-of-business CRUDing apps, and the like, need not be formalised. But certain fields, like operating systems, networking, transportation, utilities, real-time life-critical applications, plant control, voting systems, medical systems, financial systems, defence, and so on, must be mathematical, end-to-end and full-lifecycle—just like traditional engineering fields. Also, there is now an urgent need for a comprehensive framework (statutes, regulations, guidelines, standards, and policies) to normalise the use of #AI in IT, especially in those life-critical practice areas. The time for hoping that the TechBros would self regulate is long gone.