https://www.youtube.com/watch?v=er_lLvkklsk is an incredible demo of the barliman tool generating useful programs. (19m in shows an amazing hole based workflow)
miniblog.
What do you find easier to refactor, and why?
Some of our development techniques have research supporting their effectiveness, but others do not (despite advocates). Are they placebos?
Fascinating, terrifying explanation of how numeric computation can vary depending on when data is in a register:
Is bug fixing a skill with an upper limit of what's possible? If not, how we ensure our skills keep improving?
A really helpful analogy: dynamic programming is bottom-up memoisation:
TIL the Ruby interpreter exposes ObjectSpace to normal Ruby code, so you can do things like iterate over all classes
Which would you rather refactor: a codebase with immutable types but dynamic types, or pervasive mutability with static types, and why?
SHRDLU was this incredible NLP system built in the 60s that hasn't really been surpassed since:
Swift Playgrounds is an incredible live programming experience. It's also one of the few apps exploring a tablet based IDE.
I used to think it odd that JITs rarely beat AOT because a JIT has more info. Turns out JITs can't afford powerful but slow optimisations.
Fascinating post on how diff algorithms actually work: https://blog.jcoglan.com/2017/02/12/the-myers-diff-algorithm-part-1/ (multiple possible implementations!)
"Clang can codegen a switch statement with lots of disjoint cases as a binary search" https://zneak.github.io/fcd/2017/02/19/divisions.html (impressive!)
Showing 3,811-3,825 of 7,585 posts