I agree this is a nice property. But I don’t see how the new API is relevant to this. If at any point during execution I can declare a new parameter name, how is it different from making the name visible when first getting the parameter value? You can say the best practice is to declare all the names when the process starts but then you might as well say the best practice is to get all the values when it starts. And it’s almost certainly not going to be followed anyway because there will be libraries in other packages that will want to access parameters from the current node and will end up just declaring right before they access.