I took an online Python course from MIT edX. Partly to expand my basic Python knowledge. Mostly to be in "student mode." Here's a thread about the structure of the course and what I learned to help inform my remote teaching for the new school year. Skip to the end for TL;DR 1/13

The course was asynchronous. Work was released in weekly chunks. Each week followed the same format: video lectures interspersed with exercises to check for understanding, followed by a problem set. The consistency was super helpful! No surprises. 2/13

The videos were fairly short, from 5 to 15 min, average about 10 min. I watched at 1.5x speed. The videos showed the professor talking to camera with his slides visible. Occasionally he'd annotate a slide or screenshare a running program. About 12 videos per week on average. 3/13

I know this sounds silly, but I liked seeing the professor and hearing him talk. He occasionally made typos or misspoke and had to start a sentence over. It helped me make a "connection" and made watching videos feel like less of a chore. 4/13

Sprinkled between the videos were check for understanding exercises. These were low stakes. They counted towards the grade, but not much. Right/wrong feedback was immediate and I had many attempts to get each question right for full credit. 5/13

Here's what surprised me about the exercises. At any time I could show the right answer and explanation *and still get full credit* when I corrected my answer afterwards. Could I have simply clicked "show answer" for every exercise and be done in 5 min? Yes. But why? 6/13

I did get stumped on a few exercises and I was really glad I had the ability to see the answer, correct my mistakes, and get full credit. I knew that if I didn't make an honest attempt, I would be in poor shape for the week's more challenging problem set. 7/13

Each set of exercises also had its own discussion forum. This was useful when I was still confused after seeing the solution. Usually several students would chime in with their own ways of looking at it, which I found helpful. 8/13

The weekly problems sets were worth more points. Like the exercises, I had multiple attempts and immediate feedback if I was right/wrong However, I could not see the solutions until after the due date. Discussion board usage was encouraged, but no spoilers allowed. 9/13

In most cases I knew whether my problem set program was going to be right/wrong before turning it in because I was given test cases: "If the input is X, then a correct program will output Y." In this sense, the goal of physics problems is very different from CS problems. 10/13

Physics problems give you the input X and the goal is to get the output Y. CS problems give you both the input X and the output Y with the goal being to show the process which makes that happen. How can I make my physics problems more like CS problems? 11/13

The midterm was a mix of small exercises and larger problems. However, there was an 8 hour time limit. Multiple attempts were given, but fewer than regular HW. Immediate right/wrong feedback, but no public solutions even after deadline. Upcoming final exam is the same. 12/13

TL;DR what worked takeaways:
* asych w/ deadlines
* short videos w/ teacher's face/voice
* small teacher errors OK if corrected
* immediate feedback and solutions for low stakes exercises
* give input and output for higher stakes problems so that showing process is the goal
13/13