A well designed online course needs online tested milestones, segment by segment. Student get an online automated test. What he gets wrong creates a fork to an

A well designed online course needs online tested milestones, segment by segment. Student get an online automated test. What he gets wrong creates a fork to an online module to teach that item. Fail again = re-teach in a different way. Repeated failmode might attract an online human to step that person along. Then back to the lessons. This is just smart teaching programs do. Labs can be done youtube style, but need careful design to deal with each possible fork in the road where a student can get off track. Some people might not be smart enough for complex abstract stuff - I have met and failed many 'memory machines' - students who can read and repeat everything they see, but are unable to use the knowledge in a novel way that has not been worked out in examples before.I used to tutor goofs like that in Engineering Physics at U of T. They had near perfect recall, but off the beaten track they could not solve new problems. Weekly problem sets were made to make them fail. They fought back with study groups - I was always in demand to help. In the end they failed the problems exams or at year end in the exams because they were basically dummies with near total recall.