An opiniated piece on the “two-language problem” in machine learning and my first adventures in APL/J.
Thoughts on the do-notation in Haskell and why I don’t like it. As an alternative, some new combinators are proposed which admit an applicative style of programming with monads.
A piece connecting global variables – a common code smell – with generic programming – the hallmark of extensibility. In the end, both draw their power (for use and abuse) from the ability to reconfigure the meaning of existing code for novel contexts.