I'm much more familiar with the Exeter problem sets than I am with IMP. I did cut my teeth on the CPM curriculum, though. I assume there will be some similarities between CPM and IMP. With that, I'd say I prefer Exeter. I'm not sure I can put my finger on it other than the fact that the Exeter problems don't do as much of the lifting for the students allowing more individualized conversations between teacher/student. I'm always for that.

I'm in favor of context, but I think we sometimes disregard math as its own context. I think the Exeter problems do a pretty good job of self-scaffolding in that many later problems build on previous problems and they capitalize on math being a context in and of itself. Again, much of what I do in class is based on my years of teaching CPM. I don't directly answer many questions, answer questions with questions, etc. However, I felt constrained by the curriculum. I don't feel that way when I work with my students through some of the Exeter problems.

I think I felt constrained by the scripted nature of the lessons. Questions were pretty much asked and there was a definite direction of each lesson that wasn't necessarily clear to me. I remember there being times that I would not be clear about what students really needed to take away from the day's work. Perhaps that had more to do with where I was in my career than it did with the curriculum, though.

I know IMP well. I think of it as a rather profoundly coherent and meaningful curriculum thats greatest strength is as a lever to re-engage teachers in thinking carefully about their work as teachers, about children's ways of thinking and knowing, and about mathematics. It is published in a textbook. It works quite well as a sequence of mathematical activities. I remain amazed that the initial author team could construe a sequence of approximately 1200 activities that comprise a rah development of the School Mathematics (content, and a very particular content) appropriate for *all* children grades 9-12.

Is it the endgame for textbook? I most sincerely think not. Simply put, because it is scripted prior to the learners involved--which includes the teacher.

I know Exeter's program some. It has an impressive load of rich problems. And, probably mostly because I only know it at a distance--that is studied it but never taught it--I have 2 basic questions/concerns. First is that while acknowledging it works exceptionally well, but that it works exceptionally well for Exeter students/classrooms/teachers. I have concerns about how it might transfer to other educational contexts.

Second, my larger concern, is that I wonder about it's underlying curricular theory. It seems to greatly value both the Content Standards and Standards for Math'l Practices (to use some of the simple language from CCSS) -- maybe devaluing children's mathematics (see, for example, Les Steffe's work, or Dewey's Child and the Curriculum). Maybe at least equally troublesome is that it seems to take the value of student interactions out, and create a more individual exercise from mathematics. See David's languaging above: "Exeter problems don't do as much of the lifting for the students allowing more individualized conversations between teacher/student." And in doing so, leaves the authority external to the child--the need for interaction with the teacher seems evident.

IMP does have an explicit curriculum theory. (There is also an explicit educational theory - that all are smart in different wys, that tracking is inequitable, and likely some others. And an implicit teaching theory--harder to say, easier to see/feel.) It is written about in the 1997 NCTM Yearbook. They state 4 principles for curriculum design:
1. Students must feel at home in the curriculum.
2. Students must feel personally validated as they learn.
3. Students must be actively involved in their learning.
4. Students need a reason for doing problems.