1981 through ~1996, I used Forth to build expert systems and, eventually, a discovery system modeled after Doug Lenat's Eurisko coupled with Ken Forbus' QP Theo

1981 through ~1996, I used Forth to build expert systems and, eventually, a discovery system modeled after Doug Lenat's Eurisko coupled with Ken Forbus' QP Theory. That platform used a Lisp-like codebase supplied by Bill Dress. The system was used to research and defend a dissertation on polymer curing - process control; a taxonomic knowledge base could be built with a simple constrained natural language interface - that project found its way into an honors biology class for exploring immune response to various insults. It was written on an ancient Mac with UI powered from Pascal. What do I think about Forth? I will agree with anyone who tells you that Forth is more a way of thinking, far less a programming language. Charles Moore was chock full of wisdom he would toss around; things like "if you can, let the compiler do the work". That was one of the neat features: you could jump out of code and directly into the compiler to compute things. I have moments when, now that we are not saddled with the limitations of 8-bit CPUs, I think it would be fun to dust off my program and bring it up on a large memory system. Alas, that codebase is locked up on a truly ancient cartridge disk for which no drivers exist anymore.