An alternative workaround to writing `float f = 0` in languages without NaN: float f; bool thingIsFoo = condition1; // store the result… if (thingI

An alternative workaround to writing `float f = 0` in languages without NaN: float f; bool thingIsFoo = condition1; // store the result… if (thingIsFoo) f = 7; // ... code ... if (thingIsFoo && condition2) // and explicitly depend on it later ++f; But this causes an extra `&&` to be computed at runtime, so it seems NaNs are still better for this case.