Prolog excels at solving NP complete problems (e.g. the SAT solver) or combinatorial problems in general that would normally require polynomial-vs-exponential t

Prolog excels at solving NP complete problems (e.g. the SAT solver) or combinatorial problems in general that would normally require polynomial-vs-exponential time to find a solution whereas the use of Prolog allows to arrive at the same solution in the linear-vs-quadratic time. It is not really possible to accomplish it with a DSL. Another amusing property of Prolog is that a typical Prolog program can be run forward... and backward, i.e. it is possible to «ask» a question and receive an «answer», or if the «answer» is known, the program can arrive at a set of all possible «questions». Hence heavy Prolog use to build expert systems and alike. Back in the days, we used to giggle that if the 42 answer was fed into a Prolog problem solver to arrive at a full and exhaustive set of questions about the meaning of life, the universe and stuff, that would instantly create a singularity at the point in time and space where the solver was run and henceforth result in a collapse of the universe unto itself.