How Can Computer Science Be Defined?

[Monster92]

Because I'm not totally sure what it really is.

Hal Abelson in this MIT lecture tells the student that Computer science is not a science. He then suggests that it could be art or engineering. He then goes on to say it's not really to do with computers in the same way Biology isn't about microscopes.
http://www.cosmolearning.com/video-lectures/overview-and-introduction-to-lisp/

A few CS graduates told me that perhaps CS is just Software Engineering. I then researched and compared some universities SE courses and CS courses and found a lot of symmetry. However, in further research I found professors stating that Cs is not a programming degree.

The Wikipedia article says that perhaps it should be called algorithmic and a university in Holland renamed their CS department Dataology.

What do you think it is? What does CS mean to you?

[MHD]

CS is the study of Universal Computational Systems.
That is, Lambda Calculus or Turing Machines

I can't really think of any simpler definition.

[WarDaft]

It's math, with slightly different goals.


Specifically, its about what you can and can't do, rather than what is and isn't true. You can have truths and proofs about what you can or can't do, or how fast you can do it, but it's just nonsense to try to decide if an algorithm is true or false. Conversely, we could say that math is a non-deterministic algorithm constructed so that if one accepts certain concepts as true, then every branch of execution results in another true statement. Math and computation are really very deeply linked.

[Ankit1010]

Computer science is about the study of algorithms to analyze how problems which can be solved at a high level. It also has a certain amount of programming to help you really understand those algorithms, because coding it yourself requires a good understanding of how they work, and develops the skill of turning those high level ideas into concrete programs. The algorithms part is logic and math, and the implementation is programming and understanding how computers work.

[mcvoid]

The confusing part is that industry doesn't care to differentiate between the two. I think computer science in its purest form is applied mathematics. This becomes especially apparent beginning with analysis of algorithms. In that sense, it has nothing to do with programming: it's general enough to be done by an electrical engineer making the logic in a circuit board rather than a program, or even by a manager devising a new business process. However, programming is where CS is most applied itself. Correctness, Turing-completeness, automata, and complexity are computer science, but these are the tools that assist programming. Programming in its purest form is pretty much the same job as back in the early 19th centry with punch card-driven looms. Good programmers, however, utilize the theory in computer science to be more effective.

[Moose Hole]

Definition of COMPUTER SCIENCE
: a branch of science that deals with the theory of computation or the design of computers

[zombie_monkey]

Scott Aaronson calls CS "Quantitative epistemology".

[Yakk]

Computer science is mathematics that cares about costs?

[WarDaft]

Eh, not quite. In math, the whole point is that you can never have certain finite statements in your system, and all of the ones you can have, you associate with being true.

In programming languages, you can have every finite statement, and the question is more how quickly you can transition from one class of statements to another, and which classes can be transformed into which other classes.

[Yakk]

I'm talking about Curry-Howard.

Models of computation and models of logic are the "same type" of thing, and proofs and program outputs correspond to each other.

In computer science, the models of computation have "costs" associated with each step: that doesn't correspond directly to anything in mathematical logic. I'm saying that the "memory and time" costs are what distinguish analysis of a program in a given model of computation from the corresponding Curry-Howard proof in the corresponding proof system.