In practice, I've found signalling NaNs to be completely unworkable and gave up on them. The trouble is they eagerly convert to quiet NaNs, too eagerly.
I am firmly in the belief that sNaNs were a mistake in IEEE 754, and all they really serve to do is to create hard trivia questions for compiler writers.
