>Object-oriented programming is eliminated entirely from the introductory curriculum, because it is both anti-modular and anti-parallel by its very nature.
>>54842306
CMU has a HUGE push towards functional programming and parallel programming.
The guy who invented the parallel scan algorithm among others (Blelloch https://www.cs.cmu.edu/~guyb/papers/Ble93.pdf) teaches a required class 15-210, which, on top of the survey course in discrete math, is considered the pinnacle of the school's CS degree.
Similarly, the dude who made SML (a functional language) was a professor at CMU (not sure if he still is).
15-210 is a great class and the website is phenomenal if anyone wants to learn parallel algorithms:
http://www.cs.cmu.edu/~15210/
another great parallelism class is 15-418 (parallel programming rather than algorithms):
http://15418.courses.cs.cmu.edu/
>>54842306
>both anti-modular
LOL WUT?
>and anti-parallel by its very nature.
Nope. Look up an old language called Prograph. Fully OOP and inherently parallel because it was a dataflow language. The company folded before multi-processor / multi-core systems became common. But any operation, at any level of granularity, could be run on multiple cores once all the data was available to that op. This would have been trivial for the compiler+runtime to do automatically for the programmer.
OOP has nothing what so ever to do with whether or not code can be made to run in a parallel fashion.
It would be ideal to have both, lisps balance that well but I guess being as strongly parallel as possible has huge benefits.
I don't understand why everyone hates OOP. It makes modeling data in different ways very straightforward.
so true. nobody has successfully written anything truly parallel or modular in any OOP language.
oh wait.
>"I learned the language. And the secret language of science is mathematics."
Michael Minovich - solved the gravity assist conjecture at age 25
wtf are you doing, /g/?
>>54844565
gravity assist, which made the Voyager 1 and 2 missions possible.. Up to that point, no one had been able to solve it.
He used the fastest supercomputer in the world of the day, located at UCLA, at a cost of $1000 per hour to operate, to run the equations. He gave equations for a complete tour of the solar system by means of gravity assist alone, giving NASA plenty of options.
NASA then verified the equations, and when they were found to be correct, the mission was planned.
>>54844565
I'm writing websites for small companies. I'm a humble man, I might not be as smart as them. There's nothing wrong with admitting that.
>>54844461
>I don't understand why everyone hates OOP.
You're on /g/. You're listening to NEETs who can't into OOP so they hate it.
>>54844565
At least I can get laid lol
itt butthurt pajeets
>>54842652
lel I just pirated the text book for that class from their google drive ^_^
God I hate my school where the professors are too kiky to give out there text books for free >_<
To be fair the Mellon Universty will prolly pull the book from google drive once its out of beta and make there students buy it for $250 like all the other kikes.
>>54850503
>supposedly in college
>"there students"
kill yourself
>>54847944
I can and have into OOP but its just so meh. I mean how often do I really even want to use inheritance and the other shit associated with it? Out side of class I have only ever used objects when I wanted to return more than one variable in C++ and was too lazy to do it properly.
Personally I even hate typing and want a language with no set types like assembly. Variables would only have sizes the basic datatypes would be bool, 8 bit, 16 bit, 32 bit, 64 bit, 128 bit, pointer, range (pointer to the start and end of a block of memory // start pointer and length), list and range list.
>>54850503
Neither class had a required textbook when I took them. The 210 one in development seems to be a website so although they may sell prints I doubt they'll pull the website
>OOP is good at preventing code duplication and boilerplate
Inheritance is broken, anything can do data composition, and it doesn't do function composition.
>OOP is good at controlling complexity
Only if you structure your program in a tree where there is exactly one reference to any object, which is a pain in the ass. Use a proper linear type system and you can get the same guarantees without the dumb restriction.
>OOP is good at ensuring entangled data remains in a valid configuration
Dependent types allow you to do this without making a mess of getters and setters.
>OOP is good at modelling the real world
Domain modelling never leads to good code. Code is about telling a computer what to do, not painting a picture.
>>54844565
Working at my night shift caregiving job until my one year here so the company will shell out money for CNA training.
I'm only here in /g/ because Clover uses it as its default board, and I thought this was /v/.
>shitsburg
>>54850503
>pirates the textbook
>puts it on his hdd in the folder E:\Downloads\Text Books\Comp Sci
>never looks at it again
>>54842652
Wonder if they use F# for the parrallel programming features (and being similar to ML)
>>54850757
>i've never worked on a large project
>the post
>>54850983
>>Inheritance is broken, anything can do data composition, and it doesn't do function composition.
>what is calling the super method?
>>Only if you structure your program in a tree
>what are interfaces?
>>Dependent types allow you to do this without making a mess of getters and setters.
>nope.gif
>>Domain modelling never leads to good code. Code is about telling a computer what to do, not painting a picture.
>i've never worked on a project larger then fizzbuzz: the post
>>54842306
Good. OOP is pretty much dead in the real world anyway.
>inb4 code monkeys defending OOP and trying to explain that UML is still relevant too
>>54850757
>with no set types like assembly
Enjoy your non-debuggable code.
>>54852282
When it comes to large projects, software engineering has its merit.
