When have you ever personally come upon the halting problem in the field? This can be when a co-worker / boss suggested a solution which would violate the fundamental limits of computation, or when you realized yourself that a problem you were trying to solve was, in fact, impossible to solve.
The most recent time I came up with it was when studying type checkers. Our class realized that it would be impossible to write a perfect type checker (one that would accept all programs that would run without type errors, and reject all programs that would run with type errors) because this would, in fact, solve the halting problem. Another was when we realized, in the same class, that it would be impossible to determine whether a division would ever occur by zero, in the type-checking stage, because checking whether a number, at run-time, is zero, is also a version of the halting problem.
I literally got assigned the halting problem, as in ‘write a monitor plugin to determine whether a host is permanently down’. Seriously? OK, so I’ll just give it a threshold. ‘No, because it might come back up afterward.’
Much theoretical exposition ensued.