Yes, that makes sense. The hidden layer is not so quick to realize as we then have to implement a neural net library instead of a simple formula. But an analogue output of the perceptron should be relatively easy.