Conversation

Notices

1. Embed this notice
   Denis Defreyne (denis@ruby.social)'s status on Friday, 26-Jul-2024 01:20:22 JST Denis Defreyne

   • Matthew Lyon

   @mattly I believe that’s a tuple (fixed-size) rather than an array, so it wouldn’t have push().

   • Embed this notice
     Jon (Snarf) Mason (snarfmason@hachyderm.io)'s status on Friday, 26-Jul-2024 02:31:51 JST Jon (Snarf) Mason

     • Matthew Lyon

     @mattly ugh TyoeScript.

     Your former employer has asked me about converting our little app over to TypeScript and just ugh.

     I don't know if it makes more sense in a fresh code base (although I doubt it) but trying to convert and having bits of both is definitely worse than just JavaScript.

     Matthew Lyon repeated this.
   • Embed this notice
     Jon (Snarf) Mason (snarfmason@hachyderm.io)'s status on Friday, 26-Jul-2024 02:46:14 JST Jon (Snarf) Mason

     • Matthew Lyon

     @mattly all good points. But especially the last one.

   • Embed this notice
     zero (zero@woof.group)'s status on Friday, 26-Jul-2024 03:02:38 JST zero

     • Matthew Lyon
     • Jon (Snarf) Mason

     @mattly @snarfmason As someone who for years was totally in on dynamic typing (I was on the Smalltalk team at Apple, after all!) - I now see deep value in type systems for practical engineering projects.

     BUT - they have to be real, strong type systems. Not "gradual" types, nor bolted on to dynamic languages.

     With a strong type system (like Haskell's) - I've found real engineering projects benefit greatly from initial deployment, to staying bug free over their lifetime of enhancement.