Whose Problem Is It?

The online discussion over proposed changes to college Computer Science curriculum recommendations has brought out lots of interesting nuggets. Like this one from Mark Guzdial:

I attended an NSF PI's meeting once where I met someone who was one of the "education experts" at the meetings of the AAAS where the "Science for All Americans" science curricular standards were drafted.  He said that the list of attendees were a who's-who of science. These experts thought through what all Americans should know about their fields, what the predeccessor knowledge logically was, and how that knowledge should be distributed across the years of schooling.  As these lists were being formed, the education expert noted that the lists for Third Grade (about 8 years old) and Eighth Grade (about 11 years old) were getting really long.  As he read the items more closely, he began to wonder if the majority of Third Graders could even handle these abstract concepts, in terms of developmental levels.  He raised these questions, and was met with a brick wall. "That's not our problem!"

Well, whose problem is it?  Do we just let the domain experts specify what ought to be learned, apart from the constraints of the amount being crammed into the curriculum and the student's development?  If learning scientists, educators, and psychologists should play a role, when should they play that role?  And if these experts conflict, who should win?

Note that if you're looking to expand your universe of blog reading, this is probably a more fruitful direction to go than, say, marketing blogs, even if you don't have a strong feeling (yet!) about how many hours of instruction should be devoted to functional programming languages.