The mental model doesn't seem consistent to me. Within a function diagram the boxes (which seem to just be functions) have input and output nodes, but the func

The mental model doesn't seem consistent to me. Within a function diagram the boxes (which seem to just be functions) have input and output nodes, but the function that you are working on contains it's arguments. Taking a look at the first example from the tutorial, it would make more sense to remove the salesA and salesB boxes and instead have nodes on the outside of the function. Then you could have external nodes (one for salesA and one for salesB) feed into the grossProfit function. This would solve the problem of the fairly inelegant "Input Values" dialog. Ultimately I agree with gruseom on this one, any real world program would be sufficiently complex to make this not very useful. But as a teaching tool I think it could be wonderful to take abstract concepts and turn them into concrete objects that a student could play with and inspect. It's also nice to see all the intermediate values.