Is there a reason why the protocol should handle inserting the result of a repl evaluation with dedicated code vs just letting the lisp side print it? I ask because now that there are nested presentations, there are advantages to having the lisp side do the printing. In particular a list result which has presented objects in could have the elements of the list work as individual presentations in addition to the whole list being available for re-use, whereas in the current situation you can only get hold of the complete result.
-Alan