Thread 106091988

Why won't they fucking die?

>>106091988 (OP)
The birthrate is falling anon just wait

>>106091988 (OP)
They’re the closest thing we have to assembly without actually having to write assembly. Performance aside so much critical stuff relies on this it would be hard to break away. There’s not really a good alternative that matches the performance and adoption.

>>106091988 (OP)
Nearly everything that's still worth using today was programmed in these.

>>106092031
>They’re the closest thing we have to assembly without actually having to write assembly
LLVM IR

I've been programming c++ since the mid 90's -approx when templates were a new thing in g++. I occasionally think about spending the time to really dig into another language. What I do instead is double down into c++ itself, and that pays dividends.

Even tasks that I'd think would be faster done in something like python end up being coded as nearly a stream of consciousness. The complex part of programming is the algorithms, datastructures, and APIs - syntax should just be generated somewhere underneath consciousness and your fingers; the same for parsing - somewhere between your eyeballs and consciousness. You don't form out the letters when you speak or read, there is some other circuit that does it except in special cases when you need to really drill down into syntax - and it should be the same for programming.

c++ is really great. Too complicated? sure. But after watching the language evolve, it seems to be going in generally a good direction. Its pretty mature; let the toy languages explore the boundaries of good ideas and then steal the best ones.

I don't think c++ will ever die because it is too useful and powerful and open to its own evolution. Eventually c may steal some of the template stuff from c++, but maybe not. 'c' is good and well baked as it is. People, if there are people in the 22 century, will still be coding in 'c'.

Because all the alternatives are worse

>106092343

The problem with C++ isn't how feature full it is its how shoddily it implements those features

>>106092519
>how shoddily it implements those features

I'm not going to defend c++ for elegance - except maybe in its generated output with 'modern' c++ as input, were the compiler gets great visibility.

But what part is most shoddy to you?

>Why won't they die
Inertia and the enjoyable sensation of not having everything abstracted behind language voodoo.

>>106092044
which basically is assembly

>>106091988 (OP)
Software isn't just magically rewritten in Rust or any other language, anon. It requires effort but is generally just a waste of time.

>>106091988 (OP)
because the competition is shit
fkn duh

>>106091988 (OP)
this looks like AI

>>106092031
>the closest thing we have to assembly without actually having to write assembly
You are naive.

>>106091988 (OP)
Why won't you?

>>106092603
>But what part is most shoddy to you?
Not him but std::regex sucks

>>106093390
you didnt prove him wrong you fucking retard
that would have been mentioning LLVM exists

>>106093607
>you didnt prove him wrong you fucking retard
I don't have to. The pic related is self explanatory.

>>106093607
This board is so bad. it's honestly insane how pathetic it is.

>>106093709
>let me argue about something my interlocutor didnt even say
youre profoundly retarded

>>106093810
I am not arguing for anything. I just posted an image.
Feel free to ignore it if you don't like it.

>>106093801
>sudden explosion of pedo threads, ai shilling, and lgbt coded seetheposts
we are being visited by a bunch of pisscord tranies
thats whats hapenning

janny is in on it
its the same 5 or so fucking retards every time
pedos, lgbtranoids, tranimeposters, and corporate bootlickers
THE cancer that ails this board

>>106093832
>sudden

>>106093841
yeah
sudden
ish

rn its going on since ~12 hours or so

>>106091988 (OP)
They are a lingua franca.

>>106091988 (OP)
Aside from the fact that rewriting a ton of software is a long and tedious process that nobody really wants to do, what else should they use? What other programming language offers the same speed, compatibility and support? What would make switching to this other language even worth it?

>>106091988 (OP)
C is not a programming language anymore. It has become the protocol through which all software components interface with each other. Every other language is a second class citizen having to use some gross FFI. This is why C persists. Using any other language results in friction.

>>106091988 (OP)
>prevents your death

>>106093810
>>let me argue about something my interlocutor didnt even say

Get lost on the way to reddit?

>>106094203
>reddispace
no, retard
reddit is for mentally disabled people like you
you know
thats where you come from

>>106094070
*blocks you path*

>>106094244
Are you the darkplaces source port schizo? The way you shit up a thread is similar to his.

There are no better alternatives. They'll probably live for longer than most people on this board.

>>106091988 (OP)
No real competitors. They're all bloated horseshit

>>106094280
wow
lurk more, retard

>>106095019
ah and youre that annoying nocoder
im not an indian though

>>106091988 (OP)
>C
just the right abstraction and the right set of features, can built a compiler under a week-end, would be invent again nearly as-if it was invent by someone else because it just makes sense.
>inb4 hurr CVE
fuck off with your scam, fuck security retards, barely any CVE is actually exploitable anyways, who the fuck cares (no one as C is growing again now that people start to realize the scam of shitty language like rust, odin, zig or whatever else slop)
>C++
because it started as C-with-classes and does not require a C ffi to work with C, basically drop-in your C code (use extern as of gcc 10 or it will cry)

we don't have time to deal with braindeadry from research, C is very good as it is right now.

>>106091988 (OP)
the operating systems are written in them so they're first party native inherently.

>>106095028
>im not an indian though
Right.

>>106095065
and what makes you think i would be one?

>>106095130
>projection turbo
you know it.
you just called me an indian. thinking its gonna insult me lamao

ive been to india btw (south. 40 days). it was beyond excellent

>>106094266
>You should drop the entire Arduino ecosystem to use my special goodboy autist language
No thank you, Arduino IDE with C++ is enough for me.

>>106095223
>rust
the crabs are desperate
they think changing the narrative is gonna yield anything

they already had the narrative under control tho
~2 yeas ago we had multiple rust threads on the c atty at all times

>>106093882
don't you mean years

>>106095243
nah
there is variation throughout days
and you cannot compare the current situation to the absolute shithole this place was c. 2 years ago

>>106091988 (OP)
Why do you want them to die? They seem just fine.

>>106091988 (OP)
Too many libraries, the syntax style won and still, even to this day, there is nothing quicker than highly optimized c/c++. Zig is the only language that stands a chance of dethroning them but still, I doubt it given how far it is off 1.0 (I doubt it reaches that state). Rust was a nice idea but it's too far out there. All I see is rewrites in Rust because doing anything novel in it makes solving the problem much harder because you have to prove to some static analyser that your program is correct when you have no idea what shape it's even going to be yet. That makes Rust a no go in exploratory work and C/C++ can be used for that because it is permissive instead of restrictive..

>>106095276
i intensely distrust the direction where both these languages are going

Nobody has the money and time to replace all that is legacy.

>>106095276
>>106095313
dont want them to die tho
just i disagree theyre "doing fine"

>>106095223
There is no need to drop anything though.

>>106091988 (OP)
>Why won't they fucking die?
Because they are the foundation of all software today. Cross platform and fast as fuck.

>>106095313
C barely moves to begin with

>>106095353
Ok, I'll bite. I have used these libraries over the past year to build custom Arduino-based instruments and if you claim that I don't have to drop anything then you can probably tell me how to use them from Rust without tinkering. On Arduino IDE, I just click "Library Manager" and install them, I write my code and then I click "Upload" to execute it on my Arduino, no terminal required.
https://docs.arduino.cc/libraries/midi-library/
https://docs.arduino.cc/libraries/adafruit-ssd1306/
https://docs.arduino.cc/libraries/inifile/

>>106095663
cant entirely disagree but, uh
were at c23
and we still dont have a force inline keyword, we need to rely on the compiler to make the calls itself

ite pretty fucking important because you can use force inlines to subdivide your code into its components at 0 cost which negates the need for generics or templates
like in picrel
everything gets force inlined including the function itself bc its all building blocks

>>106095739
Dunno how it exactly works on Arduino.
When using esp, I would just pull a library into my project and add code in build.rs to make it compile along with my code. Then I just press Run on my IDE.

>>106095663
or this
this is more generic/template-like
i have two nigh identical functions
i use a force inline to keep the code thats common to both in one exemplary
it gets force inlined so theres no function call
everything gets meshed toegether as if it were all contiguous code

the other example is just using FI to contain the spaghetti into individual units

>>106095792
And of course you would have to write bindings. But these can often be generated automatically with existing tools.

Obviously this is not going to be as streamlined as Arduino IDE, but it's not that different from just a embedded setup with generic IDE. The point is, you do not really have to give up existing libraries and tools.

>>106091988 (OP)
who dis

>>106095787
computer is cheaper than memory right now, sneed

>>106093390
>o my...
>is O0 using the stack for everything
>is the compiler avoiding division
>this has completely shattered my cs101 pythonista mental model
>help me niggerman

>>106095029
funny thing is, if rust didn't have the borrow checker and unsafe blocks and instead just let you manage your own memory with pointers and allocators all C++ people would have jumped ship gladly since it's just a better C++ otherwise. the borrow checker killed rust. the borrow checker trannified rust.

>>106097524
I'm not sure about that, the syntax of rust is insane whereas most of C++ isn't templated shit and looks just like C

>>106097524
Na, the syntax is still fucking ugly

>>106095787
>dont have a force inline
#define

>>106096914
Aiko-chan.

>dumb nigger hours
Every fucking day.
How is this thread even up?

>>106097562
>>106097567
the syntax would become a slightly simpler C++ since most of the retardedness comes from lifetime annotations

>>106092343
>People, if there are people in the 22 century, will still be coding in 'c'.
C and C++ are holding back computer design. Hardware can have garbage collection, dynamic typing, lazy evaluation, and all the other advanced features of modern languages. Lisp machines were from the 80s and there were computers made for Java and Haskell. None of that requires any changes to the overall design of the computer. It's still von Neumann with RAM and instructions, just a different CPU architecture. In the 22nd century, computers probably wouldn't have von Neumann style instructions at all, but might be some kind of neural network or something we can't even imagine.

>>106093937
>What other programming language offers the same speed, compatibility and support?
Lisp, Fortran, Pascal, Ada. A lot of languages do.
>What would make switching to this other language even worth it?
It's already worth it. Lisp machines are like having a JavaScript or Python level language for OS development, with support from the hardware.

>>106094031
>C is not a programming language anymore. It has become the protocol through which all software components interface with each other.
C is the worst language for that because it's the weakest language. What you want is the most powerful language with the most powerful type system. This was the Multics designer's idea with using PL/I for Multics. They wanted the most powerful language for the time, because that would reduce redundancies between languages and increase compatibility.
>Every other language is a second class citizen having to use some gross FFI. This is why C persists. Using any other language results in friction.
If you use C, everything, even C itself, is a second class citizen. C just gives you null-terminated strings and for everything else, you get nothing.

>>106091988 (OP)
>need a solution
>check GitHub
>99% code base in C
maybe if other langs made useful projects c and cpp would die.

>>106097590
too messy
force inlines are clean and ethnically homogenous
by that i mean no messy variables
and easier compiler error output

>>106097590
>>106097704
and macros dont resolve recursively
unless i missed something

>>106097652
>C and C++ are holding back computer design.
they aren't holding anything back, people are creating new languages all the time, it's not C and C++ fault that no one uses these allegedly better tools.
>Hardware can have garbage collection, dynamic typing, lazy evaluation, and all the other advanced features of modern languages
wrong, these are very high level features, assembly and below don't have any of these things
>Lisp machines were from the 80s and there were computers made for Java and Haskell. None of that requires any changes to the overall design of the computer. It's still von Neumann with RAM and instructions, just a different CPU architecture.
how would you represent lisp feature in verilog or whatever HAL language you would use to buld that thing?
>Lisp, Fortran, Pascal, Ada. A lot of languages do.
>support
none of these languages have the support of C and C++
>It's already worth it.
and yet it does not exist in real-life outside computer museums...
>C is the worst language for that because it's the weakest language.
define weakest
>What you want is the most powerful language with the most powerful type system.
types don't exist at the hardware level. show me papers on how "more powerful" types (whatever that means) allow compilers to generate better asm.
>because that would reduce redundancies between languages and increase compatibility.
then C happened and the rest is history.
>If you use C, everything, even C itself, is a second class citizenIf you use C, everything, even C itself, is a second class citizen
the fuck does that even mean
>C just gives you null-terminated strings and for everything else, you get nothing.
it's a 5 minute exercise to implement C++-style string header in C... I have my own implementation in my personal git that I can use if needed (never as string.h is already enough)

>>106097524
No they wouldn't. The only reason anyone uses C++ is because its the only language that can natively use C headers. People suffer C++ to avoid FFI.

>>106097761
Just read this.
http://lispm.de/genera-concepts
>All Genera activities are themselves Lisp functions, based on a pool of thousands of Lisp functions, all implemented in Lisp. Below those, it's just the hardware - you never need to deal with a lower-level internal implementation language.
>Generic operations. When you need to add two numbers, you do not need to care about whether one number is an integer and another is a floating-point number. Just use the generic operation "+" all the time, and Genera figures out how to do "+" correctly. There is no performance penalty for this flexibility and freedom from detail - it is handled by special hardware.
>Automatic, dynamic memory allocation is one of the things that makes it easy to write large Lisp programs. The coupling of special hardware and ingenious software on our machine makes garbage collection very practical and inexpensive; it is neither of those things on other machines.
>Each data object contains knowledge of its own type. For simple objects, some of this type knowledge is captured in hardware tag bits. Data objects in the object-oriented Flavors and CLOS languages also know their own types and the generic operations that apply to them.
>The hardware provides run-time array bounds checking, data-type checking, and automatic reclamation of unused storage. In traditional systems, programs need to be protected against other programs' "wild" stores due to out-of-bounds array references, uninitialized pointers, data structures not in the assumed format, or releasing storage that is actually still in use. The hardware checking in a Symbolics machine prevents these conventional sources of damage from buggy programs.

>>106098054
It's not like C++ ABI is compatible with C by default, nor can you always compile C code using C++ compiler. Both Rust and C++ need extern "C" to interface with C which is pretty much FFI.

>>106098189
extern "C" is not even remotely comparable to FFI. Also, most C library headers are designed to be usable from C++. So there is no barrier. You just include the C header directly in C++ and it works.

>>106098274
>extern "C" is not even remotely comparable to FFI.
That's how you do FFI in Rust too...

>>106098338
Retard. C++ can directly load C include files. all extern "C" does is tell the compiler to use undecorated names. You can't include a C header in rust, you have to convert the entire thing to native rust definitions.

>>106091988 (OP)
shut up, jeet

>>106098441
>all extern "C" does is tell the compiler to use undecorated names
>What is ABI

>>106098441
>You can't include a C header in rust,
There exist macros for that.

>>106094266
>reopens the path

>>106092343
You can be really good at something as it goes obsolete, and still live a good life. It’s happened many many times through history

>>106091988 (OP)
I hate these AI threads

>>106091988 (OP)
I would run through that AI generated pussy til she call me Stuxnet

>>106099459
Comfy

>>106091988 (OP)
Because there are people that have to write real software and not some jeet web slop.

>>106099560
you'd unironically fuck an AI hag?

>>106091988 (OP)
I love sepples and I'm tired of pretending otherwise.

>>106099744
>>106094266
MCU bros... We're being replaced

C++ is already dead. C is making a comeback (in jeet countries)

>>106102373
>google trends as proof of a language's usage
Retarded nigger.

>>106101359
>heh I have to write garbage no one cares about
Starting to see how coal miners felt now. It's funny how a lot of "real software" is going web because people are tired of dealing with your shit low IQ segfaulty code, plus no installation helps.

>>106102492
It's going web because it's targeting the lowest common denominator and the industry got filled with kiddies that shit out "features" fast. But it's all a house of cards. That's why everything is crap.

>>106102492
>the industry going web
lolle
what a fucktard

>>106102492
>software going web
horrible

>>106102757
It's going web because software distribution on windows is still unsolved because people refuse to accept the windows store. It's a lot easier to go to website than deal with the humiliation ritual of getting admin to run some shitware that installs 40GiBs of crap and also wipes out the configuration for your older environments. Either way, the software being "native" is irrelevant if everything has to be transacted with some three tier architected software.

>>106098166
this does not explain at all how the hardware works lmao

I like having finer control of my software

>>106102958
oneAPI and SYCL is better than ROCm but damn I don't know if Intel will have any input in GPGPU by this time next year.

>>106105990
It blows me away. Intel is sitting on incredible stuff, and they are too dysfunctional to even sell products they have already designed (ex. b770). What a management shitshow.

>>106105990
>>106106309
i guess kikes and being good at money is just "another racial stereotype"
kek

(if i had to upgrade my current capabilities i would go with a cluster of 4 b770 btw. you cant do better in price/compute/bandwidth)

>>106106779
tfu; i meant a770

They're fun hacker languages.
White people like Brian Lunduke seethe about them and say all pedos should die.
They hate Richard Matthew Stallman.
Same reason they hate C.
>>"The nominee is quoted as saying that if the choice of a sexual partner were protected by the Constitution, "prostitution, adultery, necrophilia, bestiality, possession of child pornography, and even incest and pedophilia" also would be. He is probably mistaken, legally--but that is unfortunate. All of these acts should be legal as long as no one is coerced. They are illegal only because of prejudice and narrowmindedness."
>>
>>RMS on June 28th, 2003 https://stallman.org/archives/2003-mar-jun.html
>>
>>--------------------------
>>
>>"I am skeptical of the claim that voluntarily pedophilia harms children. The arguments that it causes harm seem to be based on cases which aren't voluntary, which are then stretched by parents who are horrified by the idea that their little baby is maturing. "
>>
>>RMS on June 5th, 2006) https://stallman.org/archives/2006-mar-jun.html#05%20June%202006%20(Dutch%20paedophiles%20form%20political%20party
>>
>>--------------------------
>>
>>" There is little evidence to justify the widespread assumption that willing participation in pedophilia hurts children.
>>
>>Granted, children may not dare say no to an older relative, or may not realize they could say no; in that case, even if they do not overtly object, the relationship may still feel imposed to them. That's not willing participation, it's imposed participation, a different issue. "
>>
>>RMS on Jan 4th, 2013) https://stallman.org/archives/2013-jan-apr.html#04_January_2013_(Pedophilia

>>106106779
Same here.

>>106106835
if one can subdivide their workloads, its the rational solution.
1300 eur for 2TB/s badwidth and 64 gigs in 4 banks of 16
its better than sex
30 years ago this was an entire server floor

>>106106890
The problem is the cooling, they need 3x pcie slots because of their fan cooling. I have 3x across 2 PCs rn - and need much more. I have considered designing a waterblock that will let me pack them together. The only reason I have not yet done this is I am focused on pushing software as far as I can before making any more hardware moves.

>>106107421
how about riser cables?

>>106106309
Intel has willingly hobbled themselves into being a pony who's only trick is x86.

>>106107421
and concerniong the mobo i think you need an e-atx format workstation mobo with 4 pcie slots for the setup to be practical
its gonna be a mess otherwise if you add networking to it
but you could make it work in an elegant manner with opencl i think
i never touched upon the "distributed" part of opencl but it says on the tin thats one of its use cases

>>106107421
>>106107658
ah, scratch that
opencl doesnt support that
seems regardless what you will be working with youre gonna have to write or find a transfer/management layer to cluster your machines
sounds like a very interesting thing to do btw

>>106091988 (OP)
C doesn't die because it used to write useful software.
Not something you know anything about.

why would any language die?
if you tried hard enough you could write quake in basic

>>106091988 (OP)
C is practically dead though.