What in programming are you anticipating most. For me, my heart skips a beat every time Serokell has a new blogpost on Dependent Haskell.
What in programming are you anticipating most. For me, my heart skips a beat every time Serokell has a new blogpost on Dependent Haskell.
I'm excited for this to be fully fleshed out (almost have concurrency working fully) so that I can move to more important work with it
Is this some kind of Go-based shell language?
it is a Lisp without variadic arguments (hence no parentheses are needed). a persisted side-effect log guarantees the ability to resume a program to its most recent state upon restart.
Oh, that sounds interesting. Especially the resuming.
Neat
I personally follow D and Beef very closely because I'm fed up with C++'s limitations, its idiotic STL and header source duality.
>D
I don't have much hope for that looks like the devs lost their motivation
> Beef
cool but building this was a pain in the ass on linux.
Why Beef instead of Zig or Odin?
because he is moronic
Because nice syntax and OOP, basically C# native with amazing comptime features (which are on par with zig)
I'm not saying because I don't want subhuman geets touching it
Understandable, have a nice day.
>geets
that's a new one
I'm positively extatic to see the new Python standard lose it's global interpreter lock in the coming update
I am shivering with anticipation of the C23 standard.
>Dependent Haskell
lmfao
>Dude you know this extremely bloated language?
>What if we tacked on one of the most complicated and mind-twisting cutting edge concepts in computer science?
Just use Idris
i think it's obvious that what haskell needs most is even more complicated type system features
the problem with idris is that it actually has many of the flaws haskell is claimed to have. haskell is claimed to be slow, idris actually is slow. haksell is claimed to have no libraries, idris actually has no libraries. and so on...
>Slow
Idris is fast enough for anything that's not real time. Modern Haskell can be fast enough, but requires writing highly unidiomatic code where you work exclusively with GHC primitives and do a lot of boxes and unsafe operations.
>Libraries
Idris has a fully featured FFI that interfaces with C (and any other language that interfaces with C, like Scheme), Javascript, and a few others. By extension, you can use any library from these ecosystems that you wish. The whole libraries argument is a non-issue. Always has been, in pretty much any language it gets brought up in.
>performance
idiomatic haskell code is faster than idiomatic idris 2 code by a considerable marign. (at least idris 2 isn't idris 1)
>no libs? just use ffi
this works but it throws a lot of the benefit of using idris or haskell down the toilet and also means you "get" to deal with a multi-langauge codebase, adapter functions and all that fun stuff. not an ideal experience
this is mainly true in the sense that idris has some better defaults and avoids having a bunch of type system extensions by just using dependent types. otherwise, using haskell is generally pleasant provided that you don't go balls deep into type trickery
Forgot to add, meanwhile the actual problem with Haskell is that it's a language with way too much bullshit and cruft. It's the C++ of the pure functional languages. Idris doesn't have this problem and never will.
>What in programming are you anticipating most.
True kotlin multiplatform.
Imagine one codebase for native desktop, iOS, Android and Web. Millions of people out of a job except for kotlinchads
A language better than C. At this rate, probably won't happen in my lifetime