I remember living through the AI Winter and Japan's Fifth Generation Computer Project.
Those of us who were interested in AI at SAIC, at least the people I talked with the most, were skeptical of the approach of using Prolog. This is also a personal bias since I have logged about ten years of commercial Lisp development and only perhaps six or eight months of using Prolog professionally.
That said, there are languages like Prolog, Lisp, Haskell, etc. that really help get different perspectives on developing software.
Yeah...we talked a lot about this in the mid-80s with equal parts reverence for the audacity of the goals and Cold War-ish "Mr. President, we must not allow an AI gap!". This was when the Japanese "economic miracle" hadn't had it's bubble burst and we thought they were going to steamroll the US. I vividly recall one of the elder AI types finally stepping up and saying (paraphrased) "Guys...don't loose sleep. These guys are talking about things that won't be practical for 20-30 years, if ever. Get back to work.". Considering they were talking about commodity MPP computers, dataflow architectures, logic programming, deep machine learning, etc., he was pretty much dead on.
P.S.: Lest you think he was a true visionary, the AI kids went back to work on Symbolics Lisp machines, because we all knew that was the future of AI.
Those of us who were interested in AI at SAIC, at least the people I talked with the most, were skeptical of the approach of using Prolog. This is also a personal bias since I have logged about ten years of commercial Lisp development and only perhaps six or eight months of using Prolog professionally.
That said, there are languages like Prolog, Lisp, Haskell, etc. that really help get different perspectives on developing software.