Comments by "LoneTech" (@0LoneTech) on "ThePrimeTime" channel.

  1. Well, one of these points is mostly accurate, that Haskell tends to use heap memory (though more so its GC heap than the C heap, and both are available). Fun fact: Haskell is horrible at strings by default, because the standard implementation is a singly linked list of (wide) characters. The only way it can do it half decently is because the compiler is great at optimizing away lists entirely. If you want to do serious text processing, you import Data.Text instead. We all know you can write code in C. That's not an argument for or against it. But the "as long as you /everything/ manually" list rather is.
    8
  2.  @BrunodeSouzaLino  Yes you can. There are values including functions, and we can thread things together using monads, which is useful to interact. It's people stuck in a rut (e.g. that insist class must mean struct, or code must be sequential) that decide it's hard.
    2
  3. Assembly is more of a category than a language. For instance, PIC14 assembly has more in common with 6502 assembly than PIC32 assembly, and pretty much all they share with GA144 assembly is that they have steps. As for APL, I've looked at it some but I think I'm fairly satisfied learning BQN instead. Same family, but more regular syntax and different glyph selection. Is the base assumption here that everyone starts with an Algol family language? LISP is an excellent suggestion, though you can make do with Scheme. I'd also throw in Forth, for how it relates directly to what computation is possible and how simple a compiler can be, while still allowing to grow your language to any task.
    1
  4.  @BrunodeSouzaLino  And here we see the trick to the self-fulfilling prophecy of "noone can give me a satisfactory explanation": you're the sole arbiter of satisfaction and decided, sight unseen, not to be satisfied. It's made easier by not explaining your expectations, so you can stake out new goal posts every time any attempt occurs. Monad comes from greek, same root as mono. In APLs monadic functions take one argument. It's the same in Haskell, used to thread one singular state through an ordered chain of functions and/or actions. We can use them to make decisions about computation, like how Maybe or Either can shortcut (replacing exceptions), STM can retry, and IO can interact and delay. IO does that by leaving the confines of pure computation. Lifting out only the parts that need it is one of the benefits of the monad, and another is to indicate where these actions are possible. Personally I enjoy the fact that Haskell teaches me terminology for concepts that were harder to consider without names.
    1
  5. So, I see notifications for this thread suggesting I probably replied to it at some time, but whatever that was is missing. Most of the other replies appear to be in response to other vanished posts. Youtube being youtube, clearly.
    1
  6. Might want to skip the FUD crowd and look at actual use instead. Are you taking a programming paradigms course?
    1
  7. That's a pretty fitting catchphrase for STM, which has the best implementation in Haskell.
    1