before the AI winter on the end of the 80s there was a short hype on 5th Generation Languages (mostly around prolog) that didn't live up to the promise, and like AI after them they were ignored for a long time: https://en.wikipedia.org/wiki/Fifth-generation_programming_language

When I got introduced in Prolog I found it really interesting and Prolog led me to write a post to HN which led me here. However, I think Prolog and “classic logic programming” are not as useful for general purpose programming as advanced types systems and Dependent Typed systems like Idris.

Having seen a bunch of people (linguistics students) introduced to Prolog, it's just way to easy to get into infinite lookup loops.

I think there's a lot we can learn from Prolog. If you skip over the problems of recursive rules, it offers a pretty damn decent data storage & query model. It's non-hierarchical like an SQL database, yet doesn't force you to contort your data into normalized tables.

Oh certainly! Prolog's worst feature is its strict semantics. And some implementations surely eschew this. More broadly, tooling and feedback really matter. Learning a language is tough. It's weird to see professionals sit down at say, MATLAB imagining that they'll be able to do something without reading a manual!

There was a big battle in the early 80's inside the USA for grants from the DOD to do what they called "automatic programming" which was supposed to dramatically simplify and make cheaper programming. There were 2 big academic/industry groups competing, one was the Project Mac which used LISP, and the other was a french group that used PROLOG. For obvious political reasons, the US team won the contracts. These were big contracts. A few years later, when zero results were shown, it poisoned the well for the term "AI" as it had been hyped, but total failure resulted. The Japanese govt saw the missed alternative and started a multi-hundred-million dollar project based on PROLOG. It also was a dismal failure. Neither LISP nor PROLOG was a suitable foundation for this type of project. Very dangerous to promise big things, unless like the Hot fusion people you promise it in 30 years, which they have been doing for 50 years, because after 30 years you are retired and can't be punished for failure!