Will Crichton

. @yminsky can confirm, but pretty sure Jane Street *chose* to rebuild. (And your first rebuild causality is pointing the wrong way.) Whereas Target (say) uses Haskell but they haven't had to "reshape the world" to do it. You can actually interop over ASCII, TCP, etc. <-;

Jane Street adopted OCaml incrementally, building more infrastructure as we went. It was practical from the day I started using it there. Over time, we've built a ton of infrastructure, and now are helping to solidify and extend that world.

But this is all something of a distraction. FP is practical, and is used in many forms. We're one case; React is a different kind of adoption; influence on mainstream languages is another.

The way I read Amy's point it was that FP is a technique that is in some sense radical, a fundamental break from how things are currently done. That mostly doesn't line up with the world I see, both at Jane Street, and beyond.

@amyjko, I take your point that FP has perhaps a larger radical element than the hci world. And certainly, conferences like icfp can often feel like they're about the space age future.

But I think your article hit a nerve by making it seem like PL is defined by this radical approach, and that description doesn't feel true to life, and gives PL let credit than I think it deserves.

This is a great summary of this thread! Thanks for synthesizing.

This very distributed convo raises a good research question: with all this experience from FP in the real world, can we finally distill what aspects matter most in practice? I've never seen a holistic, grounded, data-backed review of FP.