Conversations about COBOL, RPG, SAS, and the systems the world actually runs on — and what it really takes to understand, translate, and modernise them.
The case for making COBOL understandable rather than replacing it — why we started Language Foundry, and what we're actually trying to solve.
220 billion transactions a day. Three trillion dollars in daily financial flows. A language most developers under 40 have never touched — and the retirement cliff that's making it a boardroom problem.
SAS has been the lingua franca of quantitative risk, insurance pricing, and clinical analytics for 50 years. It's inside every major bank's VaR engine, every insurer's reserving model, and more clinical trials than anyone wants to admit. Nobody is replacing it — so we built a transpiler to move it to Python without rewriting the logic by hand. In this episode: how SAS actually works, why its macro preprocessor is the hardest part, what 96.9% test coverage looks like in practice, and the three classes of program that trip every automated translator.
We work with financial institutions, insurers, and government agencies across Europe and beyond.