Do programmers really love writting LOTTA code in most beloved low level unsafe Rust with manual memory management, pointers, and constantly maintaning in mind memory model overhead of a code?
e.g. spending more resources, calories & time on typing more, longer code?
e.g. spending more resources, calories & time on reading longer code?
what would Grugbrain dev say?
is offloading of complexity to compiler, typechecker, static analyzer by making more verbose code (& less human & less accessible) code -- good?
or
is less verbose, less safe, prone to errors & more expressive, more accessible, more natural lang code -- better?
>>2
"Make it correct, then make it efficient, then make it beautiful."
According to the saying above, correctness is the most important. Therefore, one should offload complexity to the compiler to reduce errors.
At some point processor architecture will evolve from babbage's analytic engine to better reflect what we actually use them for today
>>3
one part of my brain agrees, but ...
other part of my brain thinks :
do you really like to write correct proof code yourself (& enforce it on everyone else) (& how big is portion of correct software overall being written & used) where:
1+1=2
requires a proof that 1 & 1 are Natural numbers & positive, non-negative Integers & not Floats or Doubles or chars (but could be bits). A bits that are opposite of (!=) 0 (e.g. not equals (which is revert condition)).
...
etc
There is whole 3 volume book Principia Mathematica from 1912 that spend 360 pages proving 1+1=2.
or, life is short, you write MVP in any practical language of you choice, maybe even ASM, maybe with few small bugs like everywhere, that will be payed to be fixed not by you, you sell your MVP Startup to Google, for 100M. You spend money to meaningful things like donating to FLOSS, offering sposorhip grants for some ones Homotopy type theory research, paying for healthcare of you fam.
e.g. before writing a correct code, heck, even before learning how to write safe correct code, you have to answer one single most important question in Computer Science formulated by Dr.Terry Davis.
most important question in Computer Science formulated by Dr.Terry Davis.
What is that question?
if StackOverflow wouldn't lie about Rust being #1 most beloved than:
- haskel must have been at #2 place.
- haskel should have been at #1 place, way before Rust became popular
but it's at 21 place in Loved category, 29 place in Popular category of '22 report. And in later '23, '24 they made it hard to sort & view, to become more vague & opaque.
Programers are humans, humans not used to think in types & don't like to, neither they adore thinking about low level mem management.
Bugs are not always guaranteed to happen. Tests are not always even being written in first place for many software programs, code.
And even if bug happen in C, today with plenty of enemployed Bachelors, junior devs, in 3rd world, who struggle find a job at market and GPT availability - it's easier to have some one fix bug in easy lang (such as C), than pay x3 to smartass rewrite from scratch in Rust/Haskell.
I wrote a piece of static functional code yesterday, it compiled, typechecker said it is correct. It even looked correct at first glance. Only later I've noticed, the logic & variables names were wrong.
You will not have safe correct guarantees of quantum fluctuations around, but people been dying & starving on a Hill of 'Strong Static Typesystems & Safe Memory management'. Are we willing to die there?
string strong man static type systems are just strong & static as Cuba.
An idea to statically analyze, typecheck everything is insane ... seriously how you going to statically typecheck (automatically evaluate in your compiler) if 100mb ai generated videofile output is of High Quality?
You know, I'm not a Lisper/Schemer by no means, but at this minute here I compliment Scheme/Lisps dynamic type system & macro - because it makes easer to think & reason about code.
SchemeBBS should consider implementing #tag based posts, cuz this post could be considered reply to both threads, this 776 & 771.
---
ps
I noticed lesser amount of 403 for me, so issue seem fixing & we kind of nicely geting along? I have not expect to embrace Lisp, because friend of mine compared it mostly vs Dynamic langs in 777 thread, but in comparison vs Rust/Haskell as we see here ITT it does have very nice attractive advantages
(. ) ( .)
Those are Marjorie Simpson irises