SIGCSE 2017 - a path to github part 2
Note:
Just for some quick context. The process I started to describe in the last post and continue here chronicles what I did at Stuy with high schoolers. Now that I'm at Hunter college, I'm adapting the process but the overarching themes remain the same.
Post starts here:
Now that your students have been pulling the classcode repo regularly it's time to move to the next step - adding and pushing files to a repo.
I don't claim to do this the "right way" or the best way but it's worked for me and it's gotten my students up to speed.
I ease my students into adding and pushing files using git through homework submission. I create a single repo, usually named submissions and make a folder for each student underneath looking something like this:
. └── submissions ├── abram_sarah ├── chen_stuart └── thompson_steve
Before going any further, let me clear up one thing. The repo is public so all the students have access to each others work. I'm OK with that. We could have each student make their own repo and use git submodules to make a submissions repo but administration is much trickier and we lose the chance to build a culture of cooperation and trust. We can lock down our submission systems as much as we want. If students want to share assignments there's nothing we can do about it.
My preference is to try to build a culture that minimizes bad behavior and rewards good behavior. I'd rather see an incorrect homework so that I can tell what we need to work on than a copied perfect assignment. I'd rather see students cite each other, share properly, and support each other than copy in the shadows.
I'm not naive enough to believe that copying doesn't happen in my classes but I still prefer erring on the side of openness.
Once the repo is set up, I give all the students access. This is a little onerous since I have to type each of their GitHub usernames under the collaborators section of the submissions repo but I only have to do this once a semester.
Then I teach the workflow. Assignments go into a file or folder under their directory named ##_name where ## is the assignment number (01,02,03, etc) and name is the assignment name. The student workflow is:
- git pull
- git add (sometimes)
- git commit
- git push
Since every student is working in their own directory, things basically work.
The first issue that students come up with is when another student pushes before another student. The second student merely has to do another git pull before proceeding. This provides an easy introduction to dealing with the distributed nature of git.
There will occasionally be other issues when students don't follow procedure exactly so I'll have to fix the repo every now and then but nothing too bad and each time is a opportunity to show the class:
- you can pretty much always restore and fix things once you commit
them.
- the specifics to fix things.
There will be a few bumps in the road but after a few assignments the sailing will be pretty smooth.
At the end of the day, the students are a little closer to being fully fluent using git, they're hopefully using each others code as resources but in a proper way and since the students are all using a consistent assignment naming convention I can use simple scripts and shell commands to do all sorts of processing on assignments.
That's the end of phase two. The next step brings us to small projects using git.