The Death Of CS In Education?

JohnnyKimble writes "A provocatively titled article recently appeared in the 'Future of Computing' section of the British Computer Society website. 'The Death Of Computing' was written by a lecturer at De Montfort University in the UK, and considers the problem of falling interest in computer science courses in the UK and what needs to be done to encourage more students to take the courses." This ties in well with our discussion last night about Why Software is Hard.

Comments

Re:If their CS programs are like ours...

[metlin]

Let's see now:

• Almost anything in graphics programming is entirely mathematics (or physics, which once again, brings us to math).
• Networking - network optimization algorithms and the like use queuing theory, routing algorithms, graph theory and related stuff, which is math.
• Almost everything in quantitative AI is math - machine learning, neural nets and the like.
• Multimedia - codecs and the like use transforms, compression algorithms etc. which are, once again, math.
• Cryptography - crypto is almost entirely number theory. Sure, implementation needs coding skills, but the fundamentals of crypto itself is entirely math.
• Data strutures, compilers and the like are once again discrete math and graph theory, stacks, queues and trees.
• Information theory is almost entirely applied math.
• Distributed computing, parallel computing etc. once again use routing mechanisms, load distribution algorithms and other things which are heavily dependent on math.
• Theoretical computer science by itself is, well, applied math (look at computational complexity).

If you wanted to do "web development" (heh) you are in the wrong area. Other than parts of systems programming (and even there, paging algorithmsm memory management etc. are mathematical), Computer "Science" is, well, math oriented.

Re:no subject

[MBCook]

That depends. Are you talking about history, which doesn't help you with programming but makes you a more well rounded person; or are you talking about some kind of complex math which you may not see the value of but really does have a ton to do with programming?

If it's the former, I understand your grief but we all have to go though with it, and you just may discover other subjects you're interested in during the course of taking those courses.

If it's the later, maybe this isn't the field for you. Maybe you want to do something else related that doesn't require that kind of knowledge, like light system administration, computer repair, or maybe another field entirely. But even if you specialize yourself in computers (like DB training only), you will still need that stuff if you want to be really good at your job.

Computers, and especially computer science, are NOT for people who don't like Math.

Re:Blame employers

[SkyDude]

Too many employers look for checklists of skills rather than overall knowledge of an area.
You couldn't be more on target if you shot it with a .357 Magnum. While any employer wants to make sure they hire the right person for a given position, far too many rely on the degrees listed on the resume rather than the practical application of the supposedly accrued knowledge. When it comes right down to it, earning a degree only shows one's ability as a student, not the real world use of that knowledge. Why else are there so many semi-competent people in various fields, programming being just one of them? Hell, anyone can probably list a dozen different CEOs who are running their companies into the ground.

I'm not down on education or earning advanced degrees, but several years ago, I remember reading about the explosion of MBAs. In the article, the author pointed out that less than 15% of CEOs in Fortune 500 companies had advanced degrees. I don't know if that still holds true today, but it proved to me that real world knowledge was far more important than a degree in a frame.

Re:Computer science is a branch of mathematics.

[AuMatar]

Computer engineering is a hybrid EE/CS degree. Please don't put us in the group of ignorants who think CS is useless.

The Meaning of a "CS" needs to be re-thought

[FlyingGuy]

When i was in HS ( '73 - '77 ) being a Computer Scientest was meaningfull because they were the guys creating what today we call a "computer" along side guys with "EE" degrees. The EE guys built logic circuits that the CS guys wrote, by todays standards, primitive code that made them work. Compliers were extremely rare and we barely had "high" or "mid" level languages. Most stuff was writtin in machine code.

Now contrast that with today. Compilers, good ones even, are really a dime a dozen. Linkers and assembler are the same. The very talented have created languages, structures and frameworks that take most of the "programming" out of what people do today. Look at Java, Delphi, C#, C++, Ruby, Python, Perl, C, VB, all of them. How much really guts low level programming to the vast majority of programmers really do?

There are libraries and frameworks for practicaly everything. You need a database? Go download MySQL, Firebird, Oracle, DB2, Interbase etc. You want to build a UI? There is the entire MS-Windows API, Gnome, Aqua, KDE and numerous others. Need to talk TCP/IP, there are libraries for that on every platform, with simple invocations for just about every language. Almost everything low level these days has had a wrapper for your favorite dialect put around it.

The vast majority of programmers these days are more or less scripters. Yes you use the vocabulary of your favorite language, but lets be real here for a moment. Lets say you want to represent a list of files to a user via some UI. Are you going to go out and write the very low level code that will determine, with a mathematical proof, that you are reading the file entires on the disk drive to make sure you are doing it as fast as possble? Nope. In windows you are going to use the FindFirst / FindNext API. In *nix you might just spawn off a find thread and get its results back through STDIO. Thats not what a lot of people would say is programming in its classicle sense of the word.

A lot of the first programs i wrote that had a user interface sent me into long nights of just handeling field input, because at that time I was programming in Turbo Pascal 3.x and there were no librairies or API's that did that for you. So I was writing loops, capturing keyboard input, checking to see if was a function key that was pressed and if not then, well most of you know the drill. I had to build it all myself. But the best thing about that was that I had total control of the user expirience and I had total control of the way the software worked. There was very little in between me and the hardware.

These days its hard to even find the hardware, much less interact with it. Everything is burried under virtual methods or its being controlled by the underlaying OS which cannot give you direct control over it, because 8 other programs are all trying to use the same bit of hardware. I used to be able to stuff the keyboard buffer, now I stuff the message queue and its harder to deal with then the keyboard buffer.

The market forces really have not changed, as others have asserted, the nature of the beast has changed. I am 48 years old and 25 years ago there was barely a thing called a network, these days its ubiquitous. 25 years ago you had to either be one very smart mofo or you had to have a degree in Computer Science to be able to do anything other then what you got on a floppy. I was not one of the latter, and I worked HARD to understand what was happening inside tht box. I spent many many nights laerning about interrupt controllers, about drive controllers ( MFM anyone? ) about starting drive diagnostics with debug and understanding what the hell I was doing. I cursed IBM daily for dropping all the memory mapped hardware into the TOP of the address space instead of the bottom, OHHHH how I cursed them. I learned the LIM spec and how to shuffle chucks of memory around. but I digress...

Business embraced the beast and the beast grew and matured. Todays business does not need a person with a CS d