Thread 106208779

literally all good code is object based.

>>106208779 (OP)
1. it come with wrong memory usage and spamming new()/free() , it make it more slow than python and make it crash often
2. it come with STL/boost shit with is no any faster than python
3. 99% cases namespaces is enough

>>106208807
OP never mentioned C++

>>106208807
>make it crash often
maybe if you're retarded
>it make it more slow than python
lol

Literally all good code is written in C, which is objectively not object based you retard.

>>106208816
3. 99% cases namespaces is enough

>>106208807
>3. 99% cases namespaces is enough
i use real OOP only once in my live, it was UI for microcontroller
it was written on clear C

>>106208779 (OP)
OOP makes it easy to hide incompetence with formalisms.

>>106208779 (OP)
Its main strength is that it's an intuitive way to define and describe things. Its main issue is unoptimized CPU cache usage, poor memory locality and slower code (vtables, etc), all of which DOD handles better.

>>106208841
>which is objectively not object based
yet everyone just treats it like cpp anyway by putting everything in structs and writing their own constructors and destructors and even using pointers to mimic virtual functions. you're fucking retarded if you think C has any advantages over cpp whatsoever, the only reason to use it is to be a contrarian.

>>106208779 (OP)
oop kills performance in my experience by doing too much unneeded work

They never had a single lesson for PLT.
They don't know what is an abstraction technique.
They can't implemented a single design pattern.
They never did anything relevant and consistent enough that the adoption of an abstraction technique over another would matter.

>>106208943
fuck off, teacher who cant code

>>106208841
> Literally all good code is written in C
You don’t need the “good” in there.
Everything is, or was originally written in C.
C++. Java. Rust. Go.
C++ itself was a transpiler to C, which was then compiled (cfront).
It’s C all the way down.

>>106208807
>>106208886
Blame the language designers for popularizing the heap based approach with garbage collection for OOP. OOP doesn't require heap based allocation and isn't much slower, it all depends on how you draw encapsulation boundaries for objects. If you have an object encapsulate a large array that it works on, it isn't going to be much slower for that one vtable call to work on it.

Don't do vtable lookups / dynamic dispatch in tight loops and have an objects deal with a large amount of data vs dealing with each item separately.

Sometimes a vtable is what you want, but the question is more how often and where and why are you using it.

>>106208886
>>106208909
here are the inevitable LARPers who have to pretend performance matters in 2025 because it's the only possible defense for their contrarianism

>>106208901
No dumbass, I treat CPP like C.

>>106209024
So uh, do you have any argument? Or do you just pretend criticism is invalid because you make up imaginary scenarios about the people who are dishing out the criticism?

>>106209024
actually I work on a platform getting millions of hits a day and we waste 250ms+ per request by faggots coding $a = MyReallyImportantLibrary->new( $b->id );

>>106209038
It's a non-issue. Your boss or QA never once said to you "this runs too slow" or "this uses too much memory" anytime in the past 5 years except for database issues. Maybe if you work with 3d graphics or something but the vast majority of devs none of this shit fucking matters

>>106209009
it's still C
C is the only language I fully know, or C++/11
I write PS and Rust now, people ask how I know what methods are doing

>>106209021
just use clean C

you can't implement ffmpeg without OOP

>>106209086
>OOP
(You) mean pointers on functions?

>>106209086
ffmpeg turns into machine code at the end, so by definition it's possible to write that machine code

>>106209065
So it's relevant for billion dollar industries like video games & CGI and for essential tools like compilers & operating systems?

>>106209082
I use Forth.

>>106209065
what about automotive, aerospace, military, medical, or financial where there's legally required worst-case operating times on every single instruction?

>>106209142
Nobody’s done that in years.
There’s developers that are *able* to do it.
Take tesla, for example. The real-time code for the braking system was written ages ago and uses 68000 processor cores to execute that ancient code.

>>106209106
no compiler in history was written without OOP

>>106208807
>it come with
Anti-OOP types are ESL retards. Is anyone surprised?

>>106209161
>nobody's done that in years
what's the alternative?

>>106209142
sounds like LARP to me

>>106209187
AHГЛИЙCКИЯ ЯЗЫК ЭTO AДAПTAЦИЯ ЛATЫHИ ДЛЯ OCTPOBHЫХ OБEЗЬЯH

OOP with inheritteance only where truely needed is good. Mixins fix a lot of the issues with OOP

>>106208965
Read uncle bob. Your llm agents will not rescue you from your mediocrity.

>>106209192
Hire some old white or japanese guys.
Such as myself.

>>106209277
reading billions of shitty "programming" books for retards is waste of time
professionals reads only source codes

>>106209277
> Read uncle bob
Never heard of this guy until recently.
Snake-oil salesman.

>>106209297
no i mean explicitly the alternative approach you dumbass

>real-time code
>worst-case operating times on every single instruction
demo/game retro computer coders can do it

>>106208779 (OP)
>why do retards hate OOP so much?
Because it's what made them retarded.

>>106209053
That syntax makes me think that the real problem is that you subhumans spawn a new PHP interpreter for every connection and new object allocation is a red herring.

>>106208779 (OP)
Alright anon all say it with me:

STUPID
FROG
POSTER
>:(

>>106210608
Hurtful.

>>106208841
Bro doesn't know about GObject

>>106208779 (OP)
I write object-oriented python

>>106208779 (OP)
stop editing pepes to avoid my frogposter filters. or ill scream.

>>106208779 (OP)
Its abstractions are inaccurate to how machines work.

>>106210971
that's literally the point of abstraction from hardware

>>106211792
The point of abstraction is to make everything inefficient with 0 benefit?

come home white man

>>106211895
>0 benefit
It makes it easier for humans to read, write, understand, and debug. That's the whole point of high level programming languages in general. Your argument is an argument against C itself. You might as well go write in Assembly or pure binary if hardware adjacency is your only concern.

>>106212166
Real humans got replaced by H1B subhumans, you can't really argue that abstractions benefit humans.

>>106208779 (OP)

there is a learning curve to OOP. getting chatgpt to generate me procedural php which I don't even read is easier.

t. chatgpt vide coder

>>106209092
ffmpeg solves a problem that requires OOP solution. it operates on classes of objects. it doesn't matter how you implement it, but by using OOP language you can make solution more elegant.

>>106212520
ffmpeg solves a problem that doesn't exist

>>106208886
>unoptimized CPU cache usage, poor memory locality and slower code (vtables, etc)
Right, the code you write is slow because of THAT, not your shitty programming skills

>>106208779 (OP)
watch this OP, hopefully you'll understand
https://www.youtube.com/watch?v=SxdOUGdseq4

Just use CLOS for good OOP

i only program in bare metal. code is for faggots i just burn a new FPGA every time i need a new feature

>>106208868
>OOP makes it easy to hide incompetence with formalisms.

Based chad.

>>106209175
>no compiler in history was written without OOP

Lol, you're a kid, right?

>>106209277
>Read book of boomer that has never written real mission critical code for money

Uncle bob should be shot.

>>106212166
>That's the whole point of high level programming languages in general

Thinking "OOP is the highest level" only makes sense if this is year 1980.

>>106213092
>Just use CLOS for good OOP

Chad

But CLOS' OOP is diametrally different from C++/Java and far more powerful. Like the difference between a car dashboard and a plane cockpit.

>>106212972
It's literally an objective fact about OOP my shitskin friend.

Try to guess why objects containing irrelevant details during certain tasks leads to nonoptimal CPU cache utilization.

>>106213650
I'm not saying you're wrong, I'm saying that suboptimal CPU cache utilization does not have a significant impact in 99% of software.

>>106213650
>thinking objects are stored in memory as a whole
>it's another sub90 twp haver from north europe making the site worse for everyone else
every time

>>106213528
>Having reading comprehension this shitty
I pity you.

>>106208901
>everyone just treats it like cpp anyway by putting everything in structs and writing their own constructors and destructors and even using pointers to mimic virtual functions
>OOP code, in a language not primarily designed for OOP, is more comprehesoble and performs better

>>106213993
that's the problem of the language if anything
>primarily designed for OOP
>C and classes superglued together is primarily designed for OOP

>>106214019
That's the thing, C++ and it's sister languages like Java tend to borrow too much of syntax from C, that's why they fail at being good OOP tools
While I agree with OP, that good code is incompatible with OOP paradigm, it definitely has it's strengths
Python and Delphi are the perfect examples, they are easy to write working code with and are incredibly easy to debug

>>106214078
Nobody cares about OOP in C++, OOP is absolute slop.

you faggots need to define what you mean. you are talking past each other because you are using words that different people have different definitions of. probably because of jeets.

i'm not a career coooooder but i dont get the hate for OO. i use powershell not infrequently and vastly prefer it to com/bash/etc because it returning objects I can pick attributes from (or being able to run commands against only certain attributes of objects i've picked) is incredibly handy

>>106208841
>look at the code of large C projects
>filled with macros to replicate OOPisms

>>106214213
People are too retarded to realize they can pick and choose which OOP features are useful for them and instead think some one-time-use component needs to be written like it's going to be part of the standard library

>>106214414
>to replicate OOPisms
True, but that seems reasonable, as standard C is lacking functionality that would enable it to be OO-capable straight away, yet it greatly benefits in terms of coding speed (its major drawback) from emulating OO concepts

>>106214414
This is literally every non-OOP language btw

>>106214456
fair. i know a lot of the PS scripts I write are satanic, but I also know they're one-use and not meant to be extensible. I can see how the >NEED to maintain the OO charade in actual production code could annoy people when they just want to return one perishable interger

>>106208779 (OP)
its trash

>>106208779 (OP)
OOP is fucking annoying to deal with. I don't need another reason.

>>106218028
Not an issue with CLOS.

>>106218028
OOP users wouldn't claim that abomination on the left as OOP

>>106218028
The right looks like javascript

>>106214456
the standard library is pretty shit in most languages

>>106208779 (OP)
Alright that's half decent bait.

>>106209024
>she doesn't know L4 cache misses are by orders of magnitude the largest performance bottleneck in modern computing systems

>>106209053
don't fight with microsoft managers, you wont win

FP > OO.
It's that simple.

>>106218028
>Leggable
Gets me every single time

>>106218028
they will never admit that the logging framework is a dependency.
oh it's dependency injected so it doesn't hard link.
oh it's got if statements around every log statement which means that we don't actually need the logger to be there or not.
just because our code execution changes depending on if a dependency is there or not doesn't hecking mean that we have a dependency.

>>106218028
>subhuman
>fleas: 14

>>106214414
Don't know what shitty codebases you've been looking at.

>>106218028
This. I ain't gonna traverse 10 layers of inheritance to find how a member is used.

>>106221124
CLOS doesn't have this problem

>>106208779 (OP)
Programs are best understood as logic flows, not a bunch of objects shitting out objects or objects shitting between other objects or shitting into other objects. With OOP instead of thinking more about how to efficiently control the logic of the program you think more about how to fit the problem you're trying to solve into the OOP framework.

>>106208779 (OP)
Object orientation is too much wankery. The idea is to make programming retard-proof so the Indians can do the needful for pennies on the dollar while the glorious White Engineer from the North sends them the UML from his chaise lounge.

The only thing where it makes sense is Data, because it does make sense that a Data point is like an object, but then the gay and retarded hierarchy idea comes in with inheritance and it all just goes to shit. Structs are just better.

"Gee I gotta create the RequestHandler so it handles my request" statements by utterly deranged idiots. You are delusional if you think programming is about "defining the correct hierarchy of classes", which it's not.

>>106219975
>they will never admit that the logging framework is a dependency
it's not a hard dependency because you depend on interfaces, not classes
dunno what you're trying to prove

>>106220829
linux kernel?
essentially every C codebase since header/source files (classes) declare structs and methods that work on that structs. basic level OOP is just syntactic sugar around what you already do on C (structs, function pointers, etc.)
>>106221279
nocoder detected

>>106221736
>essentially every C codebase since header/source files (classes) declare structs and methods that work on that structs
Data structures are not OOP. Shit turns into OOP when you start using retarded ideas like inheritance and encapsulation.

>>106221781
>Data structures are not OOP
this board gets more retarded every day

>>106221876
k

>>106219975
>they will never admit that the logging framework is a dependency.
why not?
>it's dependency injected
>it's got if statements around every log statement which means that we don't actually need the logger to be there or not
that's not how log4j or other java logging frameworks are conventionally used

>>106208779 (OP)
ffmpeg, sqlite
>>106208901
choosing C over C++ is no brainer if you want to make a library you want to be consumed from any language. You can write whole application in C++ and it might be less tedious and better for you, but if you write a library in C++ you are screwing up anyone else that doesn't use C++. While it's possible to use C++ library it's fucking annoying.

>>106208779 (OP)
>why do retards hate OOP so much?
because it's far too easy to turn your clever idea for an Abstract Factory Mediator Memento into an untestable, unmaintainable pile of absolute garbage, when a simple module with a few basic functions would suffice.

>>106222494
This.

>>106208779 (OP)
Because you should favour composition over inheritance

>>106221876
>>106222165
stay away from computers

>>106222165
just kill yourself nocoder

After reading this thread, I am now certain that OOP is just an IQ filter.
t. OOP masterrace

>>106218028
>// !!!!Before Adding a new Animal, read the Readme.txt file in this Folder!!!!
>readme.txt
>To add a new Animal, you have to change the following Functions and add the new Animal to the If-Else-Statement in each Function:
>SetAnimalLegs
>SetAnimalFlees
>SetAnimalSound
>....
>HandleSpecialCaseAnimal
>>106221124
Let me guess, you non-no-coders are pro, and use Switch-Case instead of If-Else?

>>106208779 (OP)
you can't leverage it to make invalid program states unrepresentable.

>>106224155
>OOPshitters still haven't come up with a new example after 40 years
Oh yes I model different kinds of animals so often thank you object jesus I think I'll subclass rodent to implement a squirrel now

>>106208779 (OP)
function pointers inside structs is superior in every way to the engineering clusterfuck that is OOP

>>106224336
nah. it's a neat trick but it stops compiler optimizations.

>>106225143
If your compiler sucks

>>106208807
Alexander Stepanov is the largest contributor to the C++ STL by far and he hates OOP though.

>>106224322
nocoder came up with the animal example. I jsut abstracted. Something your kin is unable to do.

>>106225309
scott meyers has discussed this at length in his books and talks. the problem is the language wants a strict logical inheritance but programmers try to impose "colloquial" inheritance, like penguin deriving from a bird parent that has a fly method, or payment card number deriving from a number parent. madness that way goes

>>106224155
>Let me guess, you non-no-coders are pro
Yes.
>and use Switch-Case instead of If-Else?
If appropriate.

>>106212077
It has some good features.

>>106226538
>the problem is
what's the solution

>>106227235
inheritance is gay and wrong in like 99.9999% of cases. if you think you need it, you most likely don't. if there is one thing that the last 30 years of software dev has taught us, its "jeets, not even once" and very close behind that its "programmers cannot be trusted with the responsibility of a certain few language features"

>>106227296
so your solution is to not use inheritance? lol

>>106227549
yes. not out of blind obediance to some rule but because if you think through your problem you will realize that inheritance is almost never what you actually want

>>106208779 (OP)
Because they've never worked with a large team on a large codebase.
>literally all good code is object based.
That's literally wrong though. It's one of the effective ways of arranging things when doing programming in the large, but not the only one. For some classes of problem, functional programming is a better option. For some big and actually important programs, none of that is needed and instead you just need utter numeric horsepower.
OOP love vs hate is just a dumb ignorant nerd slapfight. Smarter or wiser nerds avoid that idiocy. I guess you're still dumb and ignorant.