I've been recording this 10 minutes sessions implementing a very simple dependancy graph in python. It's a bit embarrasing to put yourself out there like this but what the heck: life is too short :)
The motivation behind this has to do with developing a deeper understanding of how graphs work and how to use them to model problems we face everyday as TDs/TAs, and also to help to demythify testing as part of the development process.
I've seen so many talented technical artists starting to implement this kind of projects without much focus/structure and the outcome always lead to not getting much done (I've been there too, working in the GUI side of things before give any thought to the backend and stuff like that). I hope this series help to spread a somehow different approach, where we solve one problem at a time, without going crazy in terms of complexity/overthinking, leaving tests of every feature as we go so the code can actually be refactored when needed.
At the moment there's support for arbitrary nodes, persistent connections, and a push/pull evaluation model. I'm currently starting to work on the serialization of the graph and whatever comes next depends on you :)
So, if any of this seems interesting to you, please feel free to follow along! Any feedback is much appreciated.
Oh! I almost forgot, the code is available at github, but I honestly do not recommend to use it without following the videos as most of the value of this project comes from the process.