[disgruntledphd2]

This is an option in many tools, but it doesn't tend to work that well in practice.

I do agree that this is the ideal though (As an example if Pluto is always reactive, then this workflow becomes much more difficult as when you change a downstream datapoint, the model will be re-run).