Prolog; I did a lot of this (well, hundred+ hours, maybe) in college, but that was 1988 - 1990. There were toy versions of this in Scheme as examples in SICP,

Prolog; I did a lot of this (well, hundred+ hours, maybe) in college, but that was 1988 - 1990. There were toy versions of this in Scheme as examples in SICP, Structure and Interpretation of Computer Programs, that I enjoyed, and some AI of the era was based on reasoning from unification of contraints, lots of Lisp. Later in real life, my boss was describing a problem he wanted to address with the scientific computing package that his team (before my time) had created, and I said it sounded like a production-rule system, and that he should try using GNU Make. He got it to work very quickly (!) and we moved on. Semantic Web craze of early 2000s was lots of assertions in RDF, unification with OIL, then I think that was eaten by more mainstream XML SOAP. All of that work seemed to collapse under the weight if its XML infrastructure, but I thought that lots of it was rebooted with much more lightweight "NoSQL" JSON and LINQ F# dotNET. So I feel like a bystander to all this. I've certainly been using Make a long time, but the F# and tuple-store apps are more fun!