Open ubesluttsom opened 2 months ago
If you want to go bananas you can create a vector embedding database with entries of all IFI course descriptions, and check the cosine distance between the new issue and the description vector embeddings. This way you can get a nice automatic classifier.
IFI Course Space™.
This could be a lovely quality of life improvement for lowering the entry point for new IT volunteers. However, the knowledge/time required to create the "IFI Course Space" vs. the benefits of the finished product make it difficult to assign this issue to anyone currently.
Anyone who wishes to create a proof of concept are welcome to do so
Even though I also would appreciate such a change, there would need to be some clear guidelines in place for what to tag the issues as. For instance, virtually all issues could be tagged as needing IN5320: We use React, and that isn't taught on anything lower than MSc level at Ifi. I am willing to bet most of us haven't had 5320, but issues are still being fixed. As I said, I do not disagree with the proposal, but there is a risk of new contributors being discouraged by issues being tagged with subjects they may have some knowledge on (or can gain knowledge on through YouTube and Google), but that are many semesters away.
Issues here should have tags based on what IFI courses are relevant to implement them. Especially for enhancements and new features.
Example: “Search index based on text embeddings” could be tagged with
IN3120 – Search Technology
. Or “Some database feature” could beIN2090 – Databases and Data Modellering
.