Language Design with Leo White | Hacker News Reader

Language Design with Leo White

signalsandthreads.com

103 pointsbshanks8 comments

I'm looking forward to listening to this one. My favourite so far as been the one on Clock Synchronization (https://signalsandthreads.com/clock-synchronization/)

> Clock synchronization, keeping all of the clocks on your network set to the “correct” time, sounds straightforward: our smartphones sure don’t seem to have trouble with it. Next, keep them all accurate to within 100 microseconds, and prove that you did -- now things start to get tricky. In this episode, Ron talks with Chris Perl, a systems engineer at Jane Street about the fundamental difficulty of solving this problem at scale and how we solved it.

It was so enthralling that I ordered a GPS receiver for my Raspberry Pi to bring high accuracy time to my flat!

> If I just make a [new language designed from scratch], the odds of anyone using it other than me are almost 0. The mechanics of how languages become popular are confusing and unknown and you are very unlikely to produce the next big language, so, to a certain extent, you don’t really have a choice, if you want to actually solve real problems and actually help real people with their programming problems.

Supposing your design is good, wouldn't this just be an issue of promotion? How is a language (and the surrounding toolchain) different from other products?

> wouldn't this just be an issue of promotion? How is a language different from other products?

Network effects. Languages benefit (and become more useful) the more people use them:

- You are less likely to run into compiler bugs

- You are more likely to find a high quality library that solves your problem

- You are more likely to be able to get help with your problems

Building up the community that drives that network is a huge amount of work.

The addition of a glossary for people new to these types of conversations is so amazing. I don't think I've ever seen that before and wanted to highlight it. Really cool.

I was surprised they didn't discuss a recent major change to the language, effect handlers.

But we did, though we used the term "algebraic effects". And while it's being actively worked on, it hasn't landed yet, so I don't know that I'd call it recent...

Indeeds, sorry. I searched for "handler" and saw no hit. It's a very interesting topic for me. I wouldn't call them 'algebraic' anymore because the way they are done in OCaml the algebraic aspects are fairly obscured. It's the handling semantics which is critical. Anyway, this is besides the point.

btw, on a tangential note, in this episode of Python, OCaml, timestamp seems to be off.

Reply on news.ycombinator.com