Yes, maybe it wasn't clear, my comment was meant as a reinforcement of your statement. The einstein-sum example is in line with what some C++ libraries can do,

Yes, maybe it wasn't clear, my comment was meant as a reinforcement of your statement. The einstein-sum example is in line with what some C++ libraries can do, but I expect the common lisp implementation to be much simpler. Even more complex libraries like deal.II for pde-solving would probably benefit immensily from a domain specific language instead of a C++ interface.