On a more touching note, a student crocheted this adorable keyring for me last night, to thank me for the semester. I had to fight against the tears. It even has a little scarf!!! Too cute.

Anyways, I was really happy to give this class. Maybe I'll give it again next year. For now I'm happy the semester is over so that I can finally enjoy my Sundays again. But I'm also happy I managed to do it on top of my already full time job.

Trusting the students is also key. When working on practice sets, the students would naturally exchange together, ask questions, read documentation. Most of them had written notes of code, just like many of us had in our first years of coding!

Exercises ranged from reading and explaining code, to predicting outputs, explaining core concepts (vocabulary is key!), and of course write short code.
I even managed to add a bit of Christmassy theme to the final, which was a big win for me since I can't complete AOC properly.

But how do we grade the students now? There's been quite enough conversations around here, all which have helped me think about evaluation thoroughly. I know @mariaa.bsky.social has been collecting ideas and resources on the matter.
Personally, I went a little old school: all exams were on paper!

You still need to be able to review your peers' code, you still have to code live during interviews. And aside from that, vibe coding takes all the fun from programming.

This post by @sramsay2.bsky.social articulated more of my thoughts on the subject perfectly!
stephenramsay.net/posts/vibe-c...

As models improve in engineering and coding tasks, we can only wonder what our courses on programming mean today. The way I approach it is the following: no matter how good the models are, at the end of the day, you still need to understand what's going on, you still need to be able to design code.

thank you so much for this post!! You've perfectly articulated some thoughts I've had on vibe coding for a while now. loved reading this. 🫶🏻

big FOMO incoming.......

ah mince à ce point ? il fallait que je regarde, ça fait un bout de temps que c'est sur ma liste de lecture...

the midterm was on paper: exercises ranged from explaining code snippets to predicting code output, debugging code etc...

I'm teaching an introduction to programming in Python this semester and no one uses ChatGPT in class. I still get emails from students who didn't fully understand how loops worked who tried their best to explain what they didn't understand. maybe there's still hope?

Ideas (as someone who graduated not so long ago):
- if evaluation is based on a project of some kind, have the students present their project through an oral presentation + q&a.
- create more exhaustive assignments (? if that makes sense)

thank you so much for this!! I definitely need to review my website someday soon... (for now I'm using a Jekyll static site, which works out well!!) those you saved look amazing! great inspos :)