Teaching Myself Computer Science, what should I learn?

Question

I've already been programming for quite some time and have a firm grasp on programming itself, OOP, and a few other programming related things. However, I'm interested in learning the same things that would be taught to a computer science graduate, and was wondering what I need to cover?

In case it's relevant, I've programmed with PHP, Java, Python, C & C++ and I'm looking into Lisp/Scheme.

Answer 1

Computer Science courses obviously vary from university to university, but generally speaking they don't focus too much on languages. Learning a bunch of different languages may help you find a job, but won't teach you much about algorithms, the theory of computation, artificial intelligence, compilers, and the other subjects university courses cover.

MIT Open CourseWare would be worth looking at. It might also be worth your time to look through your local university's website and see exactly what courses and subjects are taught, and what textbooks are required.

Answer 2

Why, Knuth's Art of Programming series of books of course! Read them cover to cover.

Also, being able to list off languages you know (the syntax of) is not that impressive. If you go for a job that even requires programming expect questions on Algorithms and Data Structures.

Answer 3

Look up the curriculum for a CS program you respect. Course websites will often reference a textbook. Buy the textbooks and work through them. If nothing else, do do this with one or two algorithms textbooks. Pay attention to theory, which is what programming experience will not necessarily give you. Learn Lisp, Prolog, Haskell and one or two assembly languages. Get at least a passing acquaintance with the lambda calculus and pi calculus. Write yourself a minimal OS kernel and a compiler. Computability and complexity theory is a must. Database theory, too, even if you know SQL in practice.

Answer 4

If you want to learn for yourself, I'd suggest you work through some books:

• Tanenbaum (Networks, OS, distributed systems, ...)
• Some algorithm book (I liked Algorithms by S. Dasgupta et al.)
• Languages, Complexity - (I guess CS undergrad read (at least partially) something by Alfred V. Aho)

Also - as already mentioned - ocw.mit.edu or webcast.berkeley.edu are great audio-visual resources.

Answer 5

Code Complete is definitely worth reading

http://www.amazon.com/Code-Complete-Practical-Handbook-Construction/dp/0735619670

http://www.cc2e.com/

It will teach how you should code, not how to code.

Answer 6

Most colleges have this kind of information online these days. Look at the required courses and electives in the CS programs at a few schools, and then synthesize something that meets your interests.

Some schools (e.g., MIT) have a lot of their courseware online. Watching recorded lectures might suit you better than just reading the textbooks.

Definitely make sure you get the textbooks and work through the assignments. It will give you a much deeper understanding than just hearing or reading the material.

Answer 7

Much of the computer science curricula leaves you with "programming literacy". Here are the classes I remember that build on top of that:

• Operating Systems: Mostly covering threading and the low level interlock code necessary to make it work. We implemented out own "green" threads on top of some solaris primitives.
• Compiler Construction: Lexing and Parsing; We built a non-optimizing compiler for a C-like language.
• Fundamental Algorithms: Sorting and Searching; Divide and Conquer; Balanced Trees; Big-O Notation; Graphs
• Discrete Math: This is a post-algebra course that isn't calculus. Basically essential to understand what the heck is going on in...
• Computability Theory: Regular Languages; Finite-state automata; Turing machines; Proofs.

(There were also a bunch of electives that were "learn a technology; build something in it." Graphics, AI, and Networking come to mind. These don't really require the courses I've described above.)

Answer 8

algebra:

programming: - structured programming - object oriented programming - data structures and algorithms (qs, lists, trees, indexes, quicksort, dijkstra, etc...)

data bases: - relational - non relational

os: - linux fundamentals

software engineering:

electronics and phisics: may be optional - digital systems - automated systems (VHDL) - basic phisics on diodes and transistors, Op Amp

Answer 9

Pick up a book on Discrete Math - I used Rosen and thought it was pretty good.

After that pick up Sipser's Introduction to the Theory of Computation, which in my opinon may be the most readable book on Computer Science ever.