The profs don't have the best record (as in Pascal, Modula-2, Prolog, Lisp, Turing, APL, etc).

Pascal was an extraordinarily good teaching language for an era where computer time was very scarce and expensive, and forced a discipline on the writing of programs that, while many people hated it (including me), was darned important. It also was used very heavily in the DOS days of IBM PCs and in the early days of the Macintosh.

Prolog may well be the best example of a very specialized class of languages, and is still in heavy use in some fields, including AI.

LISP was, and still is (though it suffered from a sort of feature creep that makes perl seem tame and unaltered by time in comparison) a language that embodies a huge number of fundamental concepts in computer languages including the unification of program and data and is, even now, still 10 or more years ahead of its time. (And it's older than you are)

'Turing', as I'm not sure it was ever an actual language as such, forms a good chunk of the theoretical foundation for computers and computer languages.

APL was for the longest time incredibly well suited for what it did--manipulation of vectors and matrices, using terse specialized notation. There was nothing better for its time, and still can't be beat for clean notation. (Its one failing was choosing a character set that made it difficult to use, though there are pure-ASCII versions these days. I should see about getting one ported to Parrot)

Your condemnation of these languages, and no doubt others, is, I think, a bit naive, and reflects a limited view of the field. While they may not be important to you, or be useful in the limited area in which you spend your time, that neither makes them useless nor failures.

Your definition of success and utility is, I think, rather more limited than it might seem at first glance.