Thank you so much for this insightful comment Ethar Alali. Unfortunately, log(x) vs ln(x) is only a tip of an iceberg. I have seen other issues with the use of complementary error function, which also tends to have an ambiguous naming conventions.
Unlike, the good ol’ times, we start relying heavily on the methods of computation that offer no analytical solutions whatsoever. The only feasible solution seems to be extensive unit testing. Something, that goes very often far beyond budgets and time-frame of most scientific computing ventures.