Will Crichton

Maybe this analogy is helpfulhttps://twitter.com/samth/status/1187906887003582465?s=19 …

Maybe the confusion here is what constitutes "far future" and "current practice". There are some functional ideas in many of the most popular languages, and those functional features are increasingly popular. Maybe that's what you're referring to?

I think most of the popular new programming languages today are functional languages, and the definition of functional language from 20 years ago would now encompass virtually all languages in use. I think newer FP ideas are now being adopted in the most popular frameworks.

But the point I keep trying to make is that your description of what PL research is doing, even in the realm of functional programming, is very different from the self conception of the people doing that research -- that's why those people are now disagreeing with you on Twitter.

Indeed, there’s quite a mismatch! I get some of my impressions from direct conversations with PL researchers, but most from reading the work. Maybe that’s where the discrepancy lies.

You might want to talk to people who use FP as well! As someone who helps run an organization that benefits from one of those space-age FP languages, I deeply appreciate the human-centered benefits of PL research.

Also, the idea that PL plays only in the land of the knowable seems somewhat backward. PL is a weird field in part because it lacks a scientific mechanism for deciding what's right. It's a little like architecture, a mix of science and art.

Sure, a lot of math goes into it, but figuring out the right balance of features to fit into a language is a subtle thing. Too much, and it becomes too messy and complicated; but too little, and you don't provide features that would give people real leverage.