Not a bad list. Off the top of my head, I would say it is missing two things:
- Discrete Math (formal logic, sets, probability, etc)
- Theory of Computing (not just algorithms, but things like Turing machines, NFAs, DFAs, etc.). These may not be strictly the most practical courses, but I think a Computer Science degree would be incomplete without these.
The “Introduction to Operating Systems” link no longer works (redirects to “Autonomous Systems” courses). Instead, I would recommend using Operating Systems: Three Easy Pieces, which is the textbook I use in my OS course.
Finally, something like The Missing Semester of Your CS Education would also be a nice extra.
As others have said, discrete math is one of the obvious missing pieces. My uni also has C as the first language students learn as a part of their degree, and follows up with Java and Haskell to teach students about OOP and FP as paradigms. It’s useful to have something like C so students can learn about memory management. I’m also not seeing anything on Networking and Cyber Security (aside from Cryptography), which my university also taught.
Why is it important in this day and age to learn about memory management? That’s like saying it’s important to learn cursive, when it really isn’t.
Embedded. I’m currently writing software with 96 bytes of RAM. My next project I get to splurge and have 8kB of RAM and 32k of Flash.
I’m more scared with how badly I’ll handle/manage the 8k of RAM.
That’s a very specific usecase though that the majority of programmers likely will never have to face.
I think it’s more important than ever. Software is getting slower and slower, and the bloat is ridiculous. Imho this is because we just work with abstractions over abstracions ignorant to how it will be computed on a real machine. I think a more appropriate methapor would be, that you can speak and understand language (program) while being illiterate at the same time (not having a fundamental understanding on how a computer works). Of course this is a exaggeration, you don’t need to know about these stuff to be a adequate programmer I think.
@Suppoze @dubyakay one thing I liked about programming on Atari 8 bit machines was that your code could and was expected to hit the hardware directly. It was assumed the programmer understood the nature of the hardware and would directly “talk” to it to get it to perform their task. This made coding very efficient. Not a single CPU cycle of byte of RAM was wasted. A program that analysed data from multiple environmental sensors, tabulated, averaged and plotted the results and sent then to a charter plotter would run comfortably on 16kb of RAM.
My phone take a thousand times that to fail to open my emails.
They sure love them some Java. It’d be nice if they focused more on C/C++/Rust, you know, actual bare metal system languages that make you think about memory management.
Edit: I used to have a roommate who was studying compsci and they were making him program a PIC18F microcontroller on a development board in assembly. It was kinda fun because while he wasn’t using it, I’d have fun with it just programming normal C and making all the blinkenlights and gizmo peripherals on his board do shit, while he was struggling to even read a sensor.
In my CS degree I would have only learned and used java if not for my optional data science courses, a single class on machine language, a single SQL course, and a c++ course at community college before going to uni.
My data science courses introduced me to matlab, bash, r, Julia, python, machine learning, docker, Linux, and aws. My uni didn’t even have a data science degree, those courses primarily counted towards my math minor since they were under statistics.
The one piece of advice I still give to every CS student I meet is to diversify your classes whenever possible, don’t just stick to the core comp sci classes and take throwaway electives
Absolutely.
I don’t have a CS degree, I have a Cybersecurity and Forensics one. But, I love programming, and between the overlap of the two degrees and and my advanced designation I ended up taking about 3/4ths of the classes needed to get a CS degree.
Diversifying helped so much with me becoming a well rounded developer. My assembly programming class, while optional for CS, was mandatory for me, made me a significantly better dev. That assembly knowledge got me to become a skilled debugger, which made my C++ classes 10x easier, and it helped me understand memory at a lower level, making the memory problems easier to diagnose and fix.
I convinced a CS friend to take one of my cyber classes, Reverse Engineering, and he found te components of the class where we analyzed a vulnerable program to find and exploit the vuln, or the bit where we tried and determined the bug based on malware that exploited it is insightful to learning to program securely.
Learning about the infrastructure used in enterprise during a Windows admin or Linux admin class will make it easier to write code for those systems.
From the cybersecurity perspective, many of my CS classes carry me hard. Knowing how programs are written, how APIs are developed, and how to design complex software lets me make more educated recommendations based on what little information I’m given by the limited logs I am given to investigate. Writing code that interfaces with linux primitives makes it easier to conceptualize what’s going on when I am debugging a broken linux system.
Interesting, in my degree we had one lesson in Java for OOP (the rest of the course was C++), Java for android programming, Python in another course, and everything else from year 1 to year 4 (that had programming) was in C/C++. Except for assembly in computer architecture.
To be fair, this is my experience in academic computer science also
Edit: didn’t click the link before commenting, this is referencing what universities already teach
To be fair, this is my experience in environmental sciences too.
But okay, it was pretty broad. Like, pretty much everything scientific, data.
But no language courses, not much programming, missing stuff here and there.
But my Uni also encouraged (actually enforced) to choose random extra courses.
If it’s compsci, then it doesn’t need to be bare metal. It should be a language that’s good at demonstrating abstractions. Java wouldn’t be my choice, here. Elixir would be a good one.
You might want bare metal as a prereq to an operating system course.
If it’s software engineering, OTOH, then yes, a bare metal language has a bigger place.
If the Computer Science Degree is open source that means I can redistribute it.
It reminded me how much of a time waste formal education can feel. How come we can learn things on our own 10 times faster IF the motivation clicks in
It’s so weird tbh that at some unis they learn things in a year what you can get to know in a week if you don’t follow the slug pace plan and adhd hyperfocus kicks in
Please, teach me in a week how to write my own compiler and under what conditions re-compilation converges.
Sorry, I don’t know what’s a for loop or what’s a set, I only know how to do 2+2 in excel.
Go.
You misunderstood me. All the can’s and some’s aren’t purposeless in a sentence you know. Besides uni gives you other things like friends and connections that are invaluable and motivates you for plethora of subjects you don’t want to learn.
All I said is that unis can feel super slow compared to on your own rate of learning assuming you could find motivation to learn it all on your own.
I once met someone from 3d art program that struggled to make a chess piece in blender. Something that took me what 3-4 days to learn from scratch?
Or also that you could be dropped into the middle of Germany with a dictionary in hand and learn more Deutsch in two months than in 5 years of formal education. (God that sounds like some ww2 operation stuff)
I guess the point is that you learn things you like super fast compared to the average assumption of pace by the course/degree makers and thank god because how else would League of Legends tournaments fit into the schedule?
The pace is relaxing and that’s absolutely fine by me and when you go to a job market you still have pretty big upper hand and use 1/50 of stuff you learned.
But sometimes, sometimes you feel like fuck maybe I should be doing something harder. And „Is that all?”. Cracking decompiled programs in assembly as a kid was harder and much more fun than your run of the mill backend dot net coding that just doesn’t hit dopamine receptors anymore.
Trust me, if you wanna learn German by being airdropped into the country with a dictionary, you’ll have a very very bad time compared to what Goethe Institute teaches.
Because in order to properly speak the language you need to know the grammar. And you will almost certainly not ever grasp it properly on your own without guidance by an experienced educator. Germany is full of those people who after decades of living there keep making awful grammatical mistakes. While people who went to Goethe Institute usually don’t.
Same for people who learn JavaScript from a tutorial and suddenly they’re a web dev, but understand nothing about algorithms complexity and so the whole fuckin internet is so slow it hurts.
Precisely because of that attitude of yours.
It takes a day to learn how to train an AI model in python. It takes a PhD to understand what you’re doing.
I don’t know what formal education you got but after 15 years in the industry, having learned, worked and taught colleagues in most major areas, I have yet to observe anything as efficient at learning as my university’s CS program.
