My thesis is that Dr. Jones should play the role of a semi-intelligent diagramming partner in software design exploration.  Dr. Jones knows about software structures, how to diagram them, and some of the ways they might be changed, but not enough to make design decisions – that’s up to the programmer.  Dr. Jones automates the drawing of the diagrams, and, like hand-drawn ones, they contain task-specific information and remind the programmer of the changes he intends to make.
	But, unlike hand-drawn diagrams, they are dynamic: they can show redesigns and record multiple design alternatives.  I propose a three-phase model of interaction between Dr. Jones and the programmer, which I believe will accomplish this goal.