In a new attempt to do a live coding series, I've been recording these sessions implementing a very simple naming convention manager/thingy in python. Just like dgpy, the idea is to keep it short and straight forward (hopefully each series gets better, although I like to keep them kinda raw so we can share how things evolved and not an artifial session where everything works first try).
What are we trying to solve here?
This comes back from when I was developing riglab a few years ago, one of the things I wanted to do there was to make it "pipeline friendly", so people can integrate it seamlessly without introducing any innecessary friction into the pipeline. This meant that any tool using metadata stored on names cannot assume anything about it and a configuration layer is needed. This also applies to any sort of distribuitable tools.
Anyway, there are 11 videos at the time of this writting and the project provides enough functionality to be useful in a real scenario.
One thing I particullarly like about this series is how we moved from some not-so-great code (in order to keep things moving) to a more desirable end result with total confidence thanks to the development process, it wasn't planned that way but it kinda happened.
The source code is on github (MIT license), but the value of the project is more on the educational side of things (imho).
Let me know what you think :)