|Problems that automatically assist
|topic started 11/7/2006; 5:59:47 PM
last post 11/7/2006; 5:59:47 PM
|jason howald - Problems that automatically assist
11/7/2006; 5:59:47 PM (reads: 116, responses: 0)
Hi -- My name is jason howald. I'm a math faculty member at SUNY Potsdam. I sent a similar message to Mike Gage, and he recommended the development list, and posting here.
I've enjoyed Webwork very much, but I have often wished that it gave students some useful advice appropriate to the problems. I also wish that it would solve the problems *itself* -- I'm too lazy to type the answer as a function of the problem parameters.
I'm interested in enhancements of Webwork which allow for the creation of "automatically assistive" problems, capable of generating their own solutions, explaining those solutions to students, and of course checking student submissions for correctness. I'm focusing on differentiation, since that's a nice simple case. I'm writing to show off a tool, and start a conversation about whether the kind of functionality I'm interested in can be incorporated into Webwork.
In the second case, the student has (ok, I have) opened a bunch of requests for help on subproblems.
I'm particularly interested in defining problems in such a way that the computer will deduce for itself how to solve the problem and what contextual help is appropriate. For example, the problem definitions above look like this: DerivativeProblem("sin(x^2)ln(x)"); The program determines a canonical assisted solution, although of course it checks student answers by evaluation. Because most problems recurse to smaller problems, subproblems appear as the result of requesting help.
Now for the questions! Do you think this kind of student-computer interaction is a good model? Do you think Webwork can accommodate this kind of student-computer interaction? Do you have any general advice?
Have a super day, jason howald