I don't know how hard it would be for text adventure authors (including myself) to get used to how Prolog works. It seems a little strange to me and also might be harder to transpile to other languages. I don't remember it being very popular on home machines of the 80s.
I quite like the Amzi demo, but it does rely heavily on rules being asserted and retracted dynamically.
At the same time, I am convinced that a lot of Prolog fans try to shoehorn applications into it which would be better implemented in some other language... just as a lot of Pascal fans do the same. And something like natural language processing might be better handled by some different language entirely.
Of course, there are hybrid Prologs and hybrid Pascals which are both capable of tackling problems unenvisaged by the original language designers, and which depart a long way in philosophy from the designers' core principles.
I obviously agree that Prolog was unheard of on "home" computers in the 80s, but in fairness neither Pascal nor C (or FORTRAN or any of the other stalwarts) was common on the 8-bit systems which were dominant in that environment. But I think I have already mentioned at least one Prolog implemented in Turbo Pascal which in principle would have run on a CP/M system, although I think I also emphasised that it didn't implement dynamic rule assertion and retraction.
MarkMLl