Programming practices that increase total code volume lead to more bugs and less performance. It's not a tradeoff, it's a lose-lose. When you use significantly more code than is necessary to implement a feature, you provide an order of magnitude more code path combinations for bugs.
People think "delivering features" is somehow at odds with delivering reasonable performance. The opposite is true. Reliable code often tends to perform very well on modern CPUs. It's unreliable code that manages to cripple a modern CPU, because it is built out of massive stacks of unnecessary layers whose interactions have never been thought through by anyone.
#Sylves, a c# library for grid geometry, is finally released!
Check it out https://t.co/GrOE8I9ms1
I'll be posting more about it and what it can do in the coming days. It has powered all my recent demos, and much of Tessera, so you've already seen it in action.