Finished OSSU, ask me anything

[spamegg1]

(I'll update this as I go along)

Proof of completion (you have to scroll sideways a bit)

Originally written: Tuesday, June 9, 2020

Course-by-course review (FINISHED!)

Took 2 years (late May 2018 - early June 2020)

My significant advantages:

• used to teach college math. So I skipped math classes
  (there used to be A LOT MORE math in the curriculum, some of you might remember!)
• worked on OSSU full time

Disadvantages:

• easily demotivated, difficulty feeling rewarded

I am Strongly biased towards math and functional programming, biased against object oriented programming. I learn by doing, I dislike long lectures/concepts/info dumps.

What happened in 2 years?

• two deaths in the family
• got demotivated and "quit" twice: once for 3 weeks, once for around 2-3 months
• lots of stress from economic/political/war stuff that happened where I live. Had a few panic attacks! Financially OK (quit stock market early enough to avoid losses)
• hardware problems:
  GPU broke right when I was taking CUDA programming. The irony!
  Burnt motherboard's CPU socket from too long parallel execution. CPU is fine though!
• COVID-19 (right at the end, during my Spec. Quarantine took a psychological toll on me. Also good luck finding hardware in a pandemic!)
• Lost much weight, built a little muscle

Courses I took, roughly in this order:

• CS50
• How to Code 1,2
• PLABC
• Learn Prolog Now!
• Haskell from First Principles
• Nand2Tetris 1,2
• Intro to Networking
• Hack the Kernel
• Intro to CS and Programming using Python
• Core Theory
• Databases (Stanford)
• Computer Graphics (skipped last assignment)
• Machine Learning
• Compilers (from Udacity)
• Software Debugging
• Software Testing
• Software Architecture & Design
• LAFF - On programming for correctness
• Intro to Parallel Programming (very hard!)
• Functional Programming in Scala (5 courses, free to audit the whole thing!)

Courses I took that are NO LONGER on the curriculum:

• Software Construction 1,2
• Cryptography 1 (very hard MATH course)

Courses I took that are NOT on the curriculum:

• CS50's second half and its final project
• Software Processes (from Coursera)
• Software Architecture (from Coursera)
• Functional Programming in Haskell (from FutureLearn)
• Design of Computer Programs (Udacity CS212)

Courses I did NOT take that ARE on the curriculum:

• Software Engineering Intro/Capstone
• The Security courses that were added recently

Things I wish I knew / I wish somebody told me before:

• difference between Computer Science and Programming. Apparently it's huge! I was more interested in Programming since I have theory from my math background. Ironically it turns out I don't like Programming that much. Much prefer CS, especially algorithms.
• How to create a "learning lifestyle" and stay motivated, feel rewarded.
  Those "Mindshift" classes buried at the bottom of EXTRAS should be the first thing we are required to take!

The "shining core" of OSSU

• CS50, PLABC, N2T, Core Theory and Machine Learning.

Best Courses:

• Core Theory (Kruskal's Minimum Spanning Tree algorithm with Union by Rank and Path Compression and its running time analysis with the Inverse Ackermann function are so beautiful, I had tears in my eyes!)
• Nand 2 Tetris (makes you feel like you can do ANYTHING)

Most useful courses: Core Theory and PLABC (by far)

Takeaways:

• Be very clear about your goals and expectations from the beginning. OSSU might not give you what you expect!
• If you only "kinda" like CS/programming, it looks "cool", but you don't "love" it, it's gonna be hard. Maybe reconsider? Or do only as much as interests you? Put things into a long term perspective.
• Specializations are no big deal. Like "normal" courses. By far the most useful to prepare me for the Spec were PLABC and the Haskell book. The rest were not relevant. I could have EASILY taken the Spec after those (I had not even taken Core Theory at that point).
• Don't pay! Not even for Specializations. They offer very little support and some stuff is outdated. Not worth the money.
• Sometimes I focused way too much on finishing a course as quickly as possible and getting it out of the way, moving on to the next. I did not learn too well. When confronted with courses that were not very good / that I did not enjoy, my motivation would switch from learning to "getting it over with", forcing my way through. Don't do this! Better to skip such courses entirely.
• Physical exercise is extremely important to stay consistent and motivated.
• I pushed my "challenge yourself" thing a bit too far I think. At the same time I started OSSU I also started lifting weights and intermittent fasting. You should adjust things so that what you're doing feels A LITTLE challenging but does not overwhelm you.

What now?

• I feel very burnt out. Don't want to see a single line of code for a while! Not even clever functional one-liners that magically process gigabytes of data.
• I want to do some fun stuff later. I'll take CS50's intro to Game Development, then maybe audit the Game Specialization.
• My goal was not to get a job (although a low-key part-time remote job would be nice), but to challenge myself. MISSION ACCOMPLISHED! By a lot.
• My original goal was to take all of Advanced Programming. I'll do that at some point. Advanced Systems and Theory are left. Should take another 3 months or so.
• Eventually I should bite the bullet and stop avoiding the topics I disliked: web programming and Javascript.
• Looked at Scala jobs, Functional jobs, Triplebyte, and some local jobs. All want "senior" devs with 5+ years experience and lots of other tech I don't know yet. I can only work remotely. So I gotta keep at it! In addition to Spark, I will look into learning other tools in the Scala sphere: Akka, Play, Scala.JS

The Most Difficult Thing:

• Reading an existing code base. Reading is extremely hard. This is probably why so much software goes unmaintained, rewritten and into the garbage.

The Most Important Thing:

• Everything is at least 20x easier than Hack the Kernel.

[martinrg]

Thank you for this summary! You mention that you didn't take any math course... If you had to cove / review math for CS, what courses / books would you recommend? Moreover, regarding CS50 vs MIT 6.0001, what's your take? Another question that I wanted to ask you is whether you took CS50's Web Programming with Python and JavaScript ? Thanks again for your help! Best, Martin.

…

On Tue, Jun 9, 2020 at 9:45 AM spamegg ***@***.***> wrote: (I'll update this as I go along) Tuesday, June 9, 2020 Course-by-course review (coming soon!) Took 2 years (late May 2018 - early June 2020) My significant advantages: - used to teach college math. So I skipped math classes (there used to be A LOT MORE math in the curriculum, some of you might remember!) - worked on OSSU full time Disadvantages: - easily demotivated, difficulty feeling rewarded I am Strongly biased towards math and functional programming, biased against object oriented programming. What happened in 2 years? - two deaths in the family - got demotivated and "quit" twice: once for 3 weeks, once for around 2-3 months - lots of stress from economic/political/war stuff that happened where I live. Had a few panic attacks! Financially OK (quit stock market early enough to avoid losses) - hardware problems: GPU broke right when I was taking CUDA programming. The irony! Burnt motherboard's CPU socket from too long parallel execution. CPU is fine though! - COVID-19 (right at the end, during my Spec. Quarantine took a psychological toll on me. Also good luck finding hardware in a pandemic!) - Lost much weight, built a little muscle Courses I took, roughly in this order: - CS50 - How to Code 1,2 - PLABC - Learn Prolog Now! - Haskell from First Principles - Nand2Tetris 1,2 - Intro to Networking - Hack the Kernel <#690> - Intro to CS and Programming using Python - Core Theory - Databases (Stanford) - Computer Graphics (skipped last assignment) - Compilers (from Udacity) - Software Debugging - Software Testing - Software Architecture & Design - LAFF - On programming for correctness - Intro to Parallel Programming (very hard!) - Functional Programming in Scala (5 courses) Courses I took that are NO LONGER on the curriculum: - Software Construction 1,2 - Cryptography 1 (very hard MATH course) Courses I took that are NOT on the curriculum: - CS50's second half and its final project - Software Processes (from Coursera) - Software Architecture (from Coursera) - Functional Programming in Haskell (from FutureLearn) Courses I did NOT take that ARE on the curriculum: - Software Engineering Intro/Capstone - The Security courses that were added recently Things I wish I knew / I wish somebody told me before: - difference between Computer Science and Programming. Apparently it's huge! I was more interested in Programming since I have theory from my math background. - How to create a "learning lifestyle" and stay motivated, feel rewarded. Those "Mindshift" classes buried at the bottom of EXTRAS should be the first thing we are required to take! The "shining core" of OSSU - CS50, PLABC, N2T, Core Theory and Machine Learning. - I felt like the other classes did not add much, or were mid/low quality, or too unenjoyable, even though I enjoyed some of them A LOT, such as How to Code. Best Courses: - Core Theory (Kruskal's Minimun Spanning Tree algorithm with Union by Rank and Path Compression and its running time analysis with the Inverse Ackermann function are so beautiful, I had tears in my eyes!) - Nand 2 Tetris (makes you feel like you can do ANYTHING) Most useful courses: Core Theory and PLABC (by far) Takeaways: - Be very clear about your goals and expectations from the beginning. OSSU might not give you what you expect! - Specializations are no big deal. Like "normal" courses. By far the most useful to prepare me for the Spec were PLABC and the Haskell book. The rest were not relevant. I could have EASILY taken the Spec after those (I had not even taken Core Theory at that point). - Don't pay! Not even for Specializations. They offer very little support and some stuff is outdated. Not worth the money. - I would not recommend doing OSSU as it's listed. Pick and choose your own courses. If a course feels boring, irrelevant, or not very useful, skip it. Don't get hung up for months like I did! - Physical exercise is extremely important to stay consistent and motivated. - I pushed my "challenge yourself" thing a bit too far I think. At the same time I started OSSU I also started lifting weights and intermittent fasting. You should adjust things so that what you're doing feels A LITTLE challenging but does not overwhelm you. What now? - I feel very burnt out. Don't want to see a single line of code for a while! Not even clever functional one-liners that magically process gigabytes of data. - I want to do some fun stuff later. I'll take CS50's intro to Game Development, then maybe audit the Game Specialization. - My goal was not to get a job (although a low-key part-time remote job would be nice), but to challenge myself. MISSION ACCOMPLISHED! By a lot. - My original goal was to take all of Advanced Programming. I'll do that at some point. Advanced Systems and Theory are left. Should take another 3 months or so. - Eventually I should bite the bullet and stop avoiding the topics I disliked: web programming and Javascript. - Looked at Scala jobs, Functional jobs, Triplebyte, and some local jobs. All want "senior" devs with 5+ years experience and lots of other tech I don't know yet. I can only work remotely. So I gotta keep at it! The Most Important Thing: - Everything is at least 20x easier than Hack the Kernel. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#727>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAPFBPSJ36MJHBRNBMJPZ4DRVYVH5ANCNFSM4NZL2YAA> .

[spamegg1]

@martinrg 1. I actually kinda went through the Math for CS class. I spent 2-3 days on it. It was a decent, OK course. I did not like its textbook. It was super wordy and thick and confusing. I'll have to think of alternative books.

2. I took both CS50 and MIT 6.00.1x. Personally I prefer MIT. But actually I took CS50 as my "first beginner" class (that's how the curriculum used to be). I think they are difficult to compare because CS50 uses C and goes into low-level matters (memory, pointers etc.) To a beginner I would definitely recommend MIT 6.00.1x.

3. OK my story here is a little weird. I did take CS50's second half, and completed the Javascript Homeworks too. But I didn't really understand/learn it. For my web project I bent over backwards to avoid Javascript, so I found something called Brython that lets me use Python directly inside HTML. It's horrendously slow, but it works. My site looks very crappy but it has a working interactive Python console and an Ace code editor!

[WildRyc]

I'm in the midst of MIT 6.00.1x right now, so that's a relief. Thank you both for having this discussion!

…

On Tue, Jun 9, 2020 at 9:51 AM spamegg ***@***.***> wrote: @martinrg <https://github.com/martinrg> 1. I actually kinda went through the Math for CS class. I spent 2-3 days on it. It was a decent, OK course. I did not like its textbook. It was super wordy and thick and confusing. I'll have to think of alternative books. 1. I took both CS50 and MIT 6.00.1x. Personally I prefer MIT. But actually I took CS50 as my "first beginner" class (that's how the curriculum used to be). I think they are difficult to compare because CS50 uses C and goes into low-level matters (memory, pointers etc.) To a beginner I would definitely recommend MIT 6.00.1x. 2. OK my story here is a little weird. I did take CS50's second half, and completed the Javascript Homeworks too. But I didn't really understand/learn it. For my web project <https://number-python.herokuapp.com/> I bent over backwards to avoid Javascript, so I found something called Brython <https://brython.info/> that lets me use Python directly inside HTML. It's horrendously slow, but it works. My site looks very crappy but it has a working interactive Python console and an Ace code editor! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#727 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABDLAES3H7Y7JBEGOT2QEADRVY453ANCNFSM4NZL2YAA> .

[spamegg1]

@WildRyc It's a great course, I really like the instructor Eric Grimson too. It packs a lot into one short course (running time analysis and big-O notation, object oriented programming, data visualization etc.) and has just the right level of challenge in programming assignments (they are not so easy!)

Killer beard by the way!

[epikkoder]

I'm curious if you eventually developed the proper lifestyle you needed to keep at it. I, myself, want to really get into computer science to level up in my career, but the hardest part is to have the consistency and discipline to do it everyday or almost everyday WHILE not neglecting my other responsibilities in life. As it is, it already kinda feels like there's so many "adulting" things I need to do and learn (taxes, financial literacy, everyday chores, keep the wife happy, etc.), and it's like I'm being pulled from all different directions.

So if you can share any advice regarding that, I'd really appreciate it. 😊

[spamegg1]

@epikkoder Yeah I eventually did, but I must say that I have a very simple, minimalist lifestyle. My life was pretty much eat-sleep-exercise-code-run some errands. I feel you, sometimes aside from the adulting, even taking a shower or brushing teeth feels like a chore. It might be difficult with family, spouse, kids and other obligations which I don't have (except my mother). Not these last 2 years, but long ago I was also swamped. I had to make some sacrifices and cut out some things.

I eliminated most things that are pulling me in directions. Greatly simplified my finances, got out of stock market/investments into no-hassle, safer options, automated all the payments, taxes etc. I don't have a car so I don't have to worry about insurance, gas, check-ups etc. Got rid of my cell phone years ago. I gave up on some hobbies and interests, gave away pets and plants, got rid of furniture etc. to reduce cleaning, and significantly simplified/shortened my exercise (which actually improved my health).

I also have to run errands but I squeeze them into my "break from coding" times. Many of these are physical in my case so I treat them as exercise. It helps that I need frequent breaks anyway because that's how my head works.

I suppose in your case you have to separate some "me-time" everyday, turn everything and everyone off for a while, and stick to it, also make this clear to your family members. Gotta be selfish a little bit, and say no to others every now and then. That's the only way to do it.

This is actually similar to the "writing problem" that many academics are/were having back when I was teaching. Among all the other obligations, teaching, office hours, friends/acquaintances etc. we had no time to write our dissertations/research. We were trying to do all the other things perfectly, and feeling guilty if we didn't "do it properly".

TL;DR: You must simplify as much as possible, eliminate some things and make some sacrifices, respectfully tell others to leave you alone for a while, and learn to be a little selfish and stop feeling guilty or inadequate.

[josh-keller]

Can you talk a little bit more about the How to Code classes? I'm just starting the first one and thinking about how to approach it. Did you use the online book, "How to Design Programs" in conjunction with the courses? What do you feel is the best way to approach these two courses?

[spamegg1]

@wboard82 I did not use the book at all. I actually did not know it existed until much later. DO NOT feel like you are missing out or incomplete if you don't use the book. The courses/videos are complete on their own. The best way to approach is to have Dr. Racket open, and follow through the videos and type along with Prof Kiczales.

(As an aside, I'd like to say to everyone that doubts/feelings of incompleteness/unreadiness will keep haunting you forever if you don't tackle them; you gotta start combating them early. The trick is to move on to the next course without exhausting all the relevant resources. When I started my Specialization I had never seen a single line of Scala code, but I came out dominant, confident and on top. I didn't read any books first.)

Consciously stick to the discipline they teach you in that class: creating stubs for functions, writing the signature (input-output type), writing the tests first, writing a template for the function body from the data type you are using, and writing the "solution" ONLY at the very end (resisting the urge to guess/write the solution immediately). Just this one thing will carry you through both courses.

These courses are functional so they are all about "functional decomposition" and rigidly sticking to this discipline will force you to break down problems correctly. The data types will get more complicated but always remember, they are just made up of smaller parts! You'll see what I mean at the final project of the second course. It's super hard! But it works out as long as you keep breaking it down to smaller problems.

Also, think recursively!

[RmdanJr]

Keep going, dude! Thanks for this beautiful review.

[SarCoptU-zz]

Wow, congratulations! That I can only imagine was a massive undertaking.
The idea of reducing all the clutter in your life in order to finish a difficult task could not ring more true for me. I used to have 5-6 hobbies that I tried to squeeze in every day. I does lead only to frustration and burnout. Also trying to complete every task and really get any perfect result at anything gets you nowhere, a lesson I learned the hard way.
I will take my chances and ask this, although it is ok if you don't think is of interest to the topic and ignore this part of the comment.
I wanted to ask an opinion on the little web app I created as a final project for CS50x - Intro to CS. It was inspired by OSSU and it is basically a database with basic info on the courses taken/in progress so far with a web interface and Flask on the server side. It is done in CS50 ide with the SQL imported from CS50 module.
Youtube video https://www.youtube.com/watch?v=zAUmEEJ2iPI
Github https://github.com/SarCoptU/my-CS-education

I am thinking on learning how to deploy it and create for it a register/login page so myself and other people, maybe in OSSU community can use it without needing to have access to CS50 ide just to have a look at it functioning.

Best wishes,

Remus

[spamegg1]

@SarCoptU That's awesome! This is what I requested a few months ago, when someone in Gitter chat said they could lend their time and expertise to help out OSSU. You can deploy it using a free service such as Heroku following the instructions from CS50, which is what I used

[SarCoptU-zz]

@spamegg1 Great stuff! I'll do that. Thank you very much.

[AbdesamedBendjeddou]

@spamegg1 congrats, and thanks for all the helpful information, I admire your commitment, I wish you all successes and I hope that I can do it too.

[spamegg1]

@AbdesamedBendjeddou Very appreciated!

[spamegg1]

@Andreilg Hard to say. I think it's not "directly" applicable, that's probably true. It teaches you tons of important stuff in simplified format, on a "toy" computer. You wrap your head around how a computer works from the bottom up. It's true that spending too much time on it would be a waste.

But if you ever have to write a compiler, parser, or assembly/VM code, the ideas you learn will definitely help. When you get to OSTEP you'll definitely see it. For example it helped in Hack the Kernel where I have to implement first-fit or best-fit memory allocation. It also helped writing the system calls (necessary to understand how registers work). Maybe take a look at the Computation Structures courses in Advanced Systems? I'll do that in the future and report whether N2T was applicable or not.

It's my favorite course among all, but mostly because of the fun and how it made me feel. When I was a kid I wanted to be an electrical engineer and make hardware, but that didn't happen, so these courses were like a childhood dream for me. Also it made me feel connected to computing history, and made me feel overpowered like I could do anything.

Interesting, I didn't like their machine language either (very cumbersome) but really liked the assembly code (with the D, M, A). The least intuitive was the VM code. But I never learned x86 so I don't know. I'm not surprised that x86 would be more straightforward, because it's a much richer instruction set, with probably much more convenient commands. In this case "simpler" actually means "harder".

If you think about it from the perspective of the creators of the course, they must have had to decide on a trade-off between simplicity/accesibility and features/difficulty. To me it feels like they made a good trade-off.