The Open Source Computer Science Degree
github.com
github.com
I think "degree" should mean a degree - if they want to say "computer science curriculum" that's fine, or "degree requirements" or something
Now, don't get me wrong: I actually would really enjoy if there was such a thing as "degree as a service," so one self-studies via Khan Academy or MIT Open Courseware or whatever and then shows up to (e.g.) MIT and sits for the exams to satisfy mastery over the subjects
I know there's politics around "brand dilution" or whatever, but I don't see how it's any different from sitting for the AWS Certifications in that way. AWS doesn't have their brand diluted if you self-study and the pass the exam
I think Peter Thiel's answer is correct - The value of the education is similar to a night club's. It's about how long the line is outside of the club that determines how desirable it is.
It's not only the knowledge that people are paying for. It's the branding and the filtering that provides most of the value. So Stanford/MIT granting degrees to anyone that studies open source material will never happen because that dillution destroys most of the value proposition
Edit: Another point: what other business would not seek to dramatically increase supply of their product if they could only sell their extremely expensive product to <5% of the willing buyers? Any other business would invest significantly in increasing their production capacity. But with universities if they increase their capacity then the actual value they provide diminishes. Stanford is no longer Stanford if they have 900k students
Edit 2: My hope is that a university without enormous branding/filtering risk (Like WGU) could implement a model like this. Or a tech co could spin up a small attached accredited university that exclusively focuses on granting degrees to async learners. Like Amazon expanding their certifications they provide to granting an actual BS in CS if the student passes a bunch of exams.
I applaud my alma mater for their MSCS effort, which in my mind elevates their prestige not dilutes it: https://www.gatech.edu/academics/degrees/masters/computer-sc...
The degree doesn't matter in most of tech but if I ever wanted to work for the government or military I'd be automatically disqualified for not having a degree despite having the ability. The tech interview process that many people deride actually provides the opportunity to get a great tech job if you just simply know the material. It's been a huge advantage for me
You do not need a degree to work for the government or military, at least not as a contractor, and almost all of the interesting work is done by contractors.
Is any specific thing worth it once that many people can easily recieve it?
I remember how well-respected certain technical certifications were a long time ago, and then you had braindumps, 3-day cram courses... and they basically became useless because most of the people that had them couldn't even answer basic questions.
Maybe I'm not following something, but I doubt very, very seriously that a college degree could be termed "easily" for any accredited institution. I'm open to those high value places getting all the best teachers, and unquestionably the networking opportunities are outrageous, but I still wouldn't expect someone to sleepwalk through a no-cost community college course, for example, just because it is free and not Stanford-Priced
Do they though, or do a lot of people apply to big name schools hoping to have that acceptance for cachet and not actually care about attending or even have the ability to attend for financial or other reasons?
>So Stanford/MIT granting degrees to anyone that studies open source material will never happen because that dillution destroys most of the value proposition
Presumably it wouldn't happen regardless because not enough people would actually pass the exams anyway.
I agree, I often think the same about 'college' or 'university', your company's internal training catalog shouldn't be called a university.
And notably (and related to my Java observation) there are two OOP classes here.
I like the concept but it is some key things a 4 year degree's sometimes (not saying always!) have:
- Compiler Design
- Architecture (though some of the "systems" courses get into this)
- Databases (there is a DB course in there but it says "Essentials" and it is debatable if it actually even covers that)
- AI / Machine Learning
- Networks / Distributed Systems
- Embedded Systems (less common in CS degrees but almost every CS student I know has messed with Arduinos so it's not a stretch to have a basic embedded systems course)
Also, one of the most fun parts of an undergrad can be the electives, it would be nice to see some electives added here. I took 3D Graphics Programming and Bioinformatics, which I enjoyed at lot.
- a lot more maths - really a lot - what’s here doesn’t come close to covering what we did in the first year alone. For a CS degree, I would at least expect far more algebra, some graph theory, a course on optimisation and dynamic programming, some PDE and a course on numerical methods and algorithms and some formal logic (but from what I saw when I did my post-grad the amount of maths done in my country for a CS engineering degree goes further than a maths bachelor in the US so I’m extremely biased).
- Compiler and some language theory
- a course on logic programming, another on functional programming
- a course on assembly and low level programming
- a course on Networks and systems
- a course on databases architecture
- on top of a ton of physics and courses not directly oriented towards CS but required to be an engineer here.
This varies widely by University. The Git repo matches my experience pretty well except I also had Calc II and Discrete Mathematics.
> a course on logic programming, another on functional programming
This was already in the Git repo when I looked. I think the "Programming Languages" from University of Washington Courses cover that. Specifically "Programming Languages, Part C"... though it doesn't mention Common Lisp or Scheme and uses a lot of Ruby, which is an odd choice (IMHO).
Graph theory wasn’t covered. We had Calc 2, Statistics, Linear Algebra, Differential Equations and Logic. Assembly, Fortran, Java, JavaScript, Standard ML, Ada and C++ were covered.
- Digital Logic Design (basic CMOS based logic, HDLs, and basic VLSI to add on top of computer architecture)
- Real time systems (WCET analysis, scheduling, etc)
Also, some schools blur CS and Electrical Engineering and for those you may take a whole other set of courses neither of us mentioned.
Stuff like basic circuits, signal analysis, semiconductors (doping and transistor design/analysis), and for some people optionally also RF and network analysis (not that network, the one with smith charts).
Games like Rust have crafting recipes that downright make fun of reality. You produce complicated fire arms using the most primitive tooling imaginable. "video game logic" is kind of tiring.
Nand2tetris, for instance, is a course I like and recommend to the curious. But it is at best equivalent to a survey course in computer engineering. It’s a motivator to go deeper, but not where a college student should stop. If it were given for credit at a university it would be intended for first year students.
To me, the necessary elements of a CS degree are the following:
* Discrete math in a fashion (as much a refresher in symbolic manipulation from high school algebra as anything else)
* Introduction to programming
* Introduction to computer architecture (basically nand2tetris in terms of scope)
* Data structures and algorithms
* Software design [this is usually flavored as intro to OOP, but that it be OOP isn't necessary]
After that, it's really down to a depth + breadth requirement: concentrate a certain amount of learning in a narrowish track, and also take some basic courses from other tracks. But I don't think there's any other particular course that I'd consider a graduate deficient if they didn't take that course.
You need to have familiarity with both a high level and a low level language.
Creating a lexer to match tokens uses finite state automata / regular languages.
Creating a parser to match a syntax uses push down automata.
There are some complex topics built on top of simple data structures. For example, symbol table lookups with scope.
Each one of those is a semester long class by itself (well, the DFA and PDA were both in the theory of computation class for me).
My complier class tied together many of the previous classes that I had and gave me a practical understanding of it. With that, I was then able to properly understand the tools of lex and yacc, the effort they saved me and the type of errors that I would encounter.
I remember that it took me quite some time to go through it coming from a math background, but it definitely improved my understanding of how stuff worked, and I haven't really done more learning since then.
Once you understand how memory actually works, how processors execute conditionals, etc it helps massively understand what your high level language might be doing in the background.
Is a CS degree complete without the general education?
So you want to learn physics… https://www.susanjfowler.com/blog/2016/8/13/so-you-want-to-l...
"How to Become a Pure Mathematician (or Statistician)" [1]
"Chicago undergraduate mathematics bibliography" [2]
Python or Rust seem much friendlier.