months later in a classroom could be a natural incentive. Likewise, viewing a real classroom on tape could be motivational and useful as a way to think about issues of practice and their relation to the content knowledge needed to effectively function as the teacher in the videotaped classroom.

Another factor is that sites of practice are instances of actual application. I think this connects back to something Hy Bass said yesterday, the notion that we teach a calculus course and include in that course applications that an engineer might find realistic. In the same way, sites of practice such as we have discussed offer the possibility for devising applications and problems of teaching that teachers might find productive as places to think more about mathematical questions.

Sites of practice might also be thought of as generative. As an individual, they cause me to think about, or to focus on, areas of mathematics that might not have seemed as productive or important in other settings. The complexity of a set of student answers can draw you into other kinds of mathematics that you might not have been choosing as a focus at the beginning.

Finally, sites of practice serve in an activating role. They cause you to think about mathematics you might not have thought of before and to draw on mathematics that you might have known but that might not have come to the surface without this sort of spark.

These reasons sites of practice might be promising are just a beginning that poses a tentative set of suggestions. It would seem feasible that the next steps are to add to, refine, and explicate the list, making the conjectures more robust. As next steps then, in an interactive mode, we should begin to propose arguments for why these might be reasonable conjectures, and we should begin to actually design experiences to help further the argument. One way to go forward is to actually try out certain kinds of ideas to get a better understanding about which elements of this list might be most useful in terms of making the case. We need to begin in a disciplined way, using what we know about doing research, the task of studying outcome as we build coherence and language, and grounded arguments about using sites of practice to help preservice students come to know the mathematics they will need to teach well.