[thsealienbstrds]

When it comes to Haskell, I want to add that especially when you start doing parallel programming, it becomes even harder to reason about. You might expect that doing that extra data type conversion in parallel rather than in sequence is going to make things faster, but you could find that the GC needs to do more work for whatever reason, and it ends up actually becoming slower. It's very counter-intuitive. It's a shame because writing parallel code in Haskell can be very elegant. If only the behavior was more obvious it would be so great.