For me it's a bit of both: experimental code or one-off scripts can be a total mess while at the same time I care a lot about readability in other projects. For

For me it's a bit of both: experimental code or one-off scripts can be a total mess while at the same time I care a lot about readability in other projects. For one it's still helpful when I'm tired, but also while I may be able to read my own sloppy code a few days after writing it that won't be true forever. Another reason is that I have trouble with very abstract code/concepts, the more I can reason about what's actually happening with data and memory the easier and quicker I understand the code. That's why I'm terrible with Haskell etc, it requires a kind of thinking that feels like a mental foreign language.