Tove suggested to me that we allow users to effectively
'take notes' on their analyses by providing
user-editable text nodes on the result tree. This
would make it easier for users to keep track of why
they did the analyses they did, using their own
terminology. These nodes should be saved when MeV's
state is saved.