Automatic correction of complex technical exercises

We present a novel technique for the automatic correction of open exercises in the area of technical science. The task is not solvable in general, and thus we restrict ourselves to exercises whose solutions are represented through data structures that can be automatically analyzed. Example of such data structures are: undirected and directed graphs; finite state automata; truth tables, programmable logic arrays (PLAs), and Boolean functions; and stacks, trees and lists. Restricting the task to such data-structures has the added advantage that the answers can be encoded through web-forms or through cheaper paper-forms to be read through optical mark readers. This allows the delivery of such exercises both as a self-assessment tool through the web and as formal exams.