I think implementing some data structures like hash tables, lists and trees is the best way to get a feel for when one is suitable and what their traits actuall

I think implementing some data structures like hash tables, lists and trees is the best way to get a feel for when one is suitable and what their traits actually are apart from just “has O(1) insert”. You can accomplish the same by lots of use, benchmarking, studying their source code etc too. But implementing at least to me is the quick and fun way of developing this gut feeling. It’s also not that rare to suddenly have the need to implement something you’d usually find in a library because you need some unique variation of an algorithm/data structure. Whether this is truly fun I guess depends on where you are on the developer spectrum. I’d much rather develop low level libraries than app code so any chance I get is just a great opportunity to dig into the low level stuff.