Slashdot Mirror
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.
Computer Science needs to die (well, shrink a lot). Industry does not need computer scientists. It needs software engineers, human interface engineers, and programmers.
A slashdotter who didn't build his own computer is like a Jedi who didn't build his own lightsaber.
All you have to do is print lots of stories about how people are making hundreds of thousands of dollars a year "just because they can operate this computer". You know, kind of like back when I was a kid. The kind of thing that suckered me into getting into a field that is destroying itself.
"No problem. I have the capacity to do infinite work so long as you don't mind that my quality approaches zero."-Dilbert
I am frustrated at how many people persuing CS degrees don't properly understand basic data structures. Arrays, stacks, queues, vectors binary trees and the like are not really thought about as much as they should be. In too many places CS is becoming increasingly about a little bit of "CS theory" and a lot of "MS Applications".
Want to save CS? Put "Computer Science" back into it.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
If you didn't know that studying Computer Science was mainly studying a branch of applied mathematics, then you obviously didn't do enough research into the program you applied to.
And what sort of university offers "Web development" as a major? Web development is the sort of thing you learn at a community college, or on your own time with the help of several books. You don't take three or four years at a university to learn web development.
Science is hard and not sexy. There are also too few electrical engineers (not VHDL programmers), semiconductor scientists, material scientists, physicists and what not is needed to feed the entire information technology chain.
On the other hand - the other posts are probably right about the common misconception of computer science and programming.
A generation ago, back in the 70s, the science department on my highschool got a basic computer . I wrote a program that would show time with analog hands (calculated with sin(t),cos(t)).
I tried to get my son interested in programming by showing him how to write som simple software that could draw stuff.
His response was basically: "Why cant we make something like 'Grand theft auto'; This is boring"
don't cut it off www.mgmbill.org
CS is dying because it is several different disciplines wrapped up into one, making it hard for students to get the education they want (or need). Some want to focus on the mathematical and theoretical aspects of computer science. And yes, we need these people because they are the ones who come up with the new encyption methods / exponentially faster alorithems / proofs that one way to route traffic is better than another / and so on. Some want to be software engineers (learning how to program and program in groups). Still others want to focus on user interface design or software design in general, without dealing with all the programming details. And of course there are niche fields like 3D graphics and AI that are important but not really large enough to split off on thier own. In any case the point I am making is that, by cramming all these together under one degree, CS programs tend to suck because you are forced to learn stuff that you don't want to, and so the degree you earn isn't necessarily relevant to what you want to do. Students are catching on to this and are thus migrating away from the standard CS degree, some of them never to come back.
Philosophy.
We've talked about this on /. before. Many of the things I'm about to say will probably be in other comments, or you've seen in the past. I'll try to give a whole picture though. I should also mention where I'm coming from. The article is from the UK, but I'm in the US (almost dead center). I graduated with a degree in CIS (Computer Information Services) around June, and I've had a job as a software engineer for about 6 months now with a great little company.
Now I'm the kind of person who has always been interested in computers. Like many /.ers, I would probably have pursued this field if it paid next to nothing. While my salary is nothing to sneeze at, it's nothing compared to the 60-70k number people seemed to like to throw around during the bubble.
When I entered college in 2001, there were TONS of people who were in it for the money. That was clear by what they knew, how hard they tried, etc. There were more who seemed to think it would be interesting but weren't sure they wanted to do computers. There were others like me who breezed through the early programming courses because we were self-taught already in such simple things (basic loops, etc).
As I went through school, the bubble burst and the idea of instant riches from computers disappeared. Biomed seems to be the new instant riches career.
The biggest attraction to the field I see now for the average person is games. Everyone wants to make games. You like video games? Why not make them! You can get a CS degree or go to one of the many colleges offering game focused degrees (both accredited and fly-by-night). If you're on Windows, you have no chance at being exposed to programming. When I was younger we had HyperCard on the Mac (fantastic), BASIC on DOS/Windows, and you could learn. Today, Windows doesn't come with anything to learn programming. There is free stuff out there, but it doesn't come on the computer. Combine this with the fact that in the DOS days you could make something decent looking with BASIC or Hypercard that looked somewhat comparable with "real" software. Try that with today with anything. GUIs aren't easy. Even VB requires some rather abstract concepts (like events).
Some schools are not much better. The school I attended (DeVry) has scraped their computer program (which wasn't bad) and has replaced it with the "tracks" system. Now you don't get a CS degree, you can get a degree that focuses on database programming, or computer forensics ("It's computers, combined with CSI! Fight crime!"), or something else. It is nowhere near as general and well rounded as it was.
CS degrees seem to be being dumbed down (which seems at least due to trying to attract more people during the bubble). My local state school (which I attended for a while) had a pretty good CS program, but they've were dumbing it down as I left (putting off harder classes, using "easier"/trendier languages, etc.)
But like the article said. Computers aren't magic boxes any more. They are a normal part of life. They are like cars. Most people don't care how to make a car, only some people will try to do that for a living. We may be near that point with computers. Most of the children I've met in the last few years may use computers a ton, but don't care much about learning how to make stuff for them. They don't even have a passing interest in trying to find out the beginning. I may not know enough to make a car (far from it), but I understand some of the principles behind it. I know about the internal combustion engine.
I don't expect them to want to know about RCU, radix-trees, elevator schedulers, memory mapping, and other relatively esoteric things. But many don't even know about programs/operating systems/processes, or even really understand the filesystem hierarchy. They can get around quite well, and they've been trained in how to make flashy Powerpoint presentations about pointless things (I can't tell you how great a skill I think that is that the public school taught my 13 year
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
At its basis, computer science is nothing more than discrete algerbra and graphy theory. It is applied mathemtics, pure and simple. For what you want to do, you're right. You'd better off getting an art degree and picking up a few programming course on the side.
Actually, I would say that CS programmes need MORE math and theory classes. Sorry, but too many IT people are walking around with little to no understanding of fundamental data structures, calculus or logic. People think that I'm talking in circles at work, and its mainly b/c I have a physics background -- we solve problems with graphs, formulas and rigors that many IT "professionals" fail to grasp. If anything, I'm pursuing a CS masters so I can grasp even more of the high level basis of modern information technology. Its only in seeing the full theorem can you understand how it all fits.
And in what universe does a CS undergrad not learn how to program? That was a pre-req for majoring in the subject when I was in college -- and I graduate in 2000!! Must be older than I thought.
Look, you can teach most people how to hack code, troubleshoot a switch, or run a tech support call. But the underlying theorems that glue all of IT together you can only get from a classical Computer Science or engineering education. Its alot easier for a CS, Physics or Math major to switch between a networking job and a programming job, b/c he knows all of the background stuff that makes it all work. An IT-certified pro may struggle a bit, only b/c he's not going to know the basics.
My big complaint about CS majors is that most HS in the states don't prepare young people for the kind of head-scratching work that the major requires.
Computer Science is an academic program in the traditional sense and teach you how to think abstractly about the very basis of an intellectual field. Programming, database management, information services and the like are essentially trade degrees that prepare you to do jobs. Think of the difference between a degree in economics and a degree in business. Or physics vice engineering. One isn't better than another but they are fundamentally different.
IT work isn't computer science. Plumbing isn't physics. Programming is perhaps the closest to computer science as you can get as an IT worker (IT means information technology, not information science), because a good programmer who does groundbreaking work often requires knowledge which comes from computer science (and math). Big advances are increasingly happening on the mathematical front, simply because the datasets are becoming much larger as computational capacity increases. Sure, UI is very visible and thus easily recognized as important, but the nicest user interface doesn't create fast search engines, realtime renderers, distributed filesystems, voice recognition software, efficient routing, weather forecasts, encryption software, GPS receivers, etc. Try writing anything in those fields without a background in computer science...
Public breast-beatings like these are generally political maneuvers by people in the field, who want more power and funding. I'm a physicist, and in my field you hear the same kind of thing: boo hoo, the number of students majoring in physics is dropping, it's a national crisis, please throw money at us to cure the problem. Usually the people complaining are faculty who produce 20 grad students over the course of their careers, and tell all 20 of them that they're failures if they can't make their way into careers exactly like their adviser's: teaching and doing research at a school that has a high-powered graduate program. There's always the nationalism, too: watch out, because the Russians (or Chinese, or Indians, or, ...) will beat us. They always leave out some of the relevant facts: that the U.S.'s graduate-level educational system is the envy of the rest of the world; that the number of people the U.S. is trying to educate at high levels is higher than anything that's ever been attempted before in all of history. People misuse statistics like crazy, too. For instance, they compare the number of students graduating in India with the number graduating in the U.S., but the degree programs in India they're including are basically like AA degrees, not programs that are comparable to a U.S. bachelor's degree.
Another issue that people tend to sweep under the rug is that there is a pipeline at work, and the reason people drop out of the pipeline is usually a good one. At every step along the way, some people are dropping out of the pipeline simply because their genes don't make them good at the field. Others are dropping out because they're low on motivation. Others are dropping out because they don't enjoy it, and can tell that they're not going to enjoy it once they're out of school and in a job. Still others are dropping out because they see the field as being incompatible with the family lifestyle they want.
And finally, these fluctuations in enrollment are usually driven by Mom and Dad. There is always a small core of people who were born to do a certain thing, whether it's music or plumbing or CS or physics research; they're in the field because they love it, and they love it because it's what they're naturally suited for. Layered on top of that core is always a much bigger number of people who majored in something because Mom and Dad told them they could make a lot of money at it. When times are good, the core still does well, but the wannabes bail out, because it's not turning out to be a good way to earn big bucks doing something that they're marginally talented at.
Find free books.
I read the article, and I disagree utterly.
The problem is a much deeper-lying one. Universities are selling themselves as steps towards getting jobs. With very rare exceptions (divinity, for instance) this was never the case, nor was it intended to be. They are not vocational institutions, nor are they designed as such.
I have seen, and been in, vocational institutions. They are very fine places (called vocational schools, technical colleges, technikons or what have you) where pupils are drilled in particular modes of work to accomplish given tasks. They are very good at what they do, and they often work alongside other teaching systems such as apprenticeship schemes. They are not interested, institutionally speaking, in research, nor in high-flown theory. They are there to tell little Johnny that if he pulls the lever on the drill-press smoothly and evenly, it will produce an accurate, regular hole with little risk of breaking the bit. People who want to learn to be Java programmers would be well served by attending such courses. They will learn to crank out Java well, repeatably, and quickly. They won't learn in-depth knowledge about garbage collection strategies; that isn't why they are there.
Universities are not about drilling students. They are set up to expand minds. In principle a university could be a few comfortable seating areas around a vast library, with students exploring under the guidance of other people interested in expanding human knowledge. Add a few laboratories, maybe a few lecture halls for guest presentations, and you're there. In the computer science world, where the point is to have students truly understand on a deep level what is going on inside the computer, and even inside computers which only have theoretical expression, drilling them in Java would be a total waste of time.
The writer of the article wants student numbers up, and shows little or no interest in the raison d'etre of the courses and departments in the first place. His agenda, as revealed by the article, is for universities to be, or to become, vocational institutions. This is in line with the existing trend for universities to beg for students, tempting them with airy promises of gainful employment. The problem can be phrased as a question: where will those who wish for the services of universities, rather than vocational institutions, go?
Right now, the best bet would appear to be a library, or perhaps the web, because only there is pertinent information available with a minimum of time-wasting distractions. At this rate we bid fair, at least in computer science, to leave behind the benefits of university courses and return to a pre-academic level of support for research. I won't go so far as to say definitely that this is a bad thing, but I do think that to present what the author is suggesting as a university course is bordering on the fraudulent.
The real "Libtards" are the Libertarians!
Too many employers look for checklists of skills rather than overall knowledge of an area. In a job interview I was once asked why I didn't ever get an A+ certification. I told them that since I had 6 years experience in the field, I didn't need it. They still pressed me to take the A+ test after I was hired.
Similarly, the fact that I'll have a related degree in the field won't matter to a lot of HR drones. They care more about MSCE and CCNP certifications than they do a Bachelor's degree. I know the underlying concepts of networking, routing, etc., but since I haven't worked directly with Cisco routers, I'm apparently useless to them. Who cares that I can learn whatever software package they're using in a week or so?
No wonder no one wants a degree in "CS". They just want a job in the field, and there are easier ways to get there than a 4-year degree.
No. There is no better way to handle it than making you do the work. Basically what I'm hearing is "but math is hard!" Math is not only a big part of CS, it is the foundation for CS, much in the way that physics is the foundation for engineering (but even moreso). Most of the math courses they make you take are most likely going to be directly applicable, as well. Discrete mathematics? Number theory? Linear algebra? Those are all used everyday in computer science. If you really want to do computer science, knowing how to prove things mathematically is going to help you as well.
Put it this way: why are you in college? Are you there to learn how to make web pages, apps, or video games? You may be wasting your time. Actually, let me put it another way: if you aren't in a CS program because you like learning about computers for the sake of learning about computers, you're in the wrong place. Get out of that degree and apply yourself.
Did you ever notice that *nix doesn't even cover Linux?
So you signed up for the wrong course. Computer Science, or Computing Science, is really Computing / Computational Mathematics. If you didn't research what was involved in computer science then you can't blame the university, only yourself.
Computer science (and possibly some related disciplines) are dying because there are more and more dumbed down courses which people apply for instead because they are at universities that are more accesisble. People are generally lazy - the explosion in people asking homework questions on programming forums and IRC is testament to that - and if they think they can get away without doing something, generally they will.
Hard computer science problems require lots of math, you can't escape that, whether it be graphics, artificial intelligence, classification, whatever. The more I work, the more I realise just how much math I really should know, and how little that I do. There are areas of using and designing for computers that don't require math - but these are mainly psychology related. Once you get down to what is going on under the hood, it is math.
There is a reason that major employers are picky about the grades, courses and universities that they recruit from.
I've programmed as a hobby all my life and have a computing degree from Carnegie Mellon. I'll tell you what is hard: Finding a job coming out of college. Everyone looks at you like you have no idea how to code because you have no experience. It makes me mad I went to college when I coulda just coded for some startups in the mid 90s and been fine.
God spoke to me.
Computer science does have a lot to do with math but is it reasonable to expect 24 hours of math when 36 hours gets you a degree in math?
Well, let's see, "science" in this context is basically the process of using math to model physical phenomenon and "computing" is using physical phenomena to model math.
So that would be a "Yes."
Which also leaves me to wonder which part you didn't understand, the "computer" or the "science"? I'm dead serious.
Oh, I don't blame you entirely, although one might think you had looked into a field and just what it entailed before signing up for it as a major, but the really troubling part is that you had prior education and career counseling. It was not done with any apparent degree of competence.
The really, really troubling part is that I would be more surprised if it had been. Such is the state of things today.
So, don't think I'm really picking on you or anything, I'm just having a bit of a sad giggle to myself in public while banging my head against the wall.
So yes, math is a big part of CS but there has got to be better ways to handle it.
No, there do not. Computer science is about math. Because that's what it is about. If you want to be a programmer or a web designer, study those. You seem to have finally found your path. It's merely a shame there was noone about competent enough to steer you toward it in the first place.
And that's not your fault.
KFG
(Disclaimer, I'm an Electrical Engineer).
I normally see just much of what the parent has listed as being part of an Electrical Engineering disciple. In my experience, Computer Science just does not take on these areas... (CS people studying information theory *today*, HA that's a good one!)
Isn't work in multimedia codecs typically done by Electrical Engineers (signal processing, embedded systems)? The design/implementation of MPEG video codecs requires background in signal processing, VLSI techniques, etc....
My somewhat biased view is this: if it involves calculus, (mathematical) optimization, advanced probability, adaptive algorithms, etc, it is usually part of electrical engineering. On the other hand, if it involves abstract algebra, applied linear algebra, heuristic algorithms (i.e. those not based upon mathematical optimization), discrete math, compiler design, it often falls under computer science.
I've taken a sizable number of CS classes. Case in point: the Fourier Transform is apparently a new concept to CS Graduate students in a highly ranked ("Top 20") US University. Even more disturbing: deriving the DFT of a simple sine-wave was considered overly difficult! Yes, I realize most CS majors don't do this every day... Then again, its only simple calculus, and is taught to EE sophomores/juniors! This is not the only example and I could go on and on....
I'm not trying to start a flame-war, I just don't see CS as being "math-based" compared to other fields. For me, CS is somewhere between Information Technology and Engineering in terms of math.
I went to college to get an education, not training. If the education was a good one and you're reasonably talented, you should be able to fill in any holes in your knowledge yourself.
Of course it helps to find an employer who can recognize this and is willing to invest in you. Rather than, say, one who might be trying to pay bottom dollar for recent grads while simultaneously expecting them to already know everything about programming in a business environment?
(just my $.02)
then they might be failing because they are more math degrees than what I would consider "computer science". That is why I changed my major to web development here.
Obviously you don't know what computer science is.
I didn't want a math degree and that is exactly what I was getting.
No. You didn't want computer science. You wanted a trade school that would teach you how to use an application.
I agree with your comment, but, I think this is a symptom of the tendency to treat College like a trade school. Computer science was a trendy major for a while when it looked like everyone might just be the next internet millionaire - not necessarily that a lot of people were really interested in computer science beyond having a career. Heck most people have no idea what computer science is.
Brett
Try switching languages and you'll see the CS majors (from good schools) shine.
True, in my school the courses the languages that we wrote in *were* incidental. Above the 200 level there were no "X Language" course. I used to say that the profs announced that "you'll be writing your next assignment in [.. roll the dice..] Java." But most languages are similar and fall into families. Scheme is like Lisp. Assembly is like C. Java is like C# or C++.
I've had input in hiring people. Personally, I'm not so concerned that someone has the stellar skills in the language of the day. Ability in C language impresses me. Clean code impresses me. Even for web stuff. Most programmers can hack things together on the fly and that impresses the brass; they don't see that this job is also a profession. The world doesn't need more hackers (classical sense of the word) because it causes never ending debugging and refactoring sessions. Nobody wants to spend their career fixing someone's crap. And the explanation that a product that they wanted was so rushed into place that it cannot become something slightly different without major work is painful for everyone. I know. So calculate your ROI over two or three years not just one.
Not saying that someone w/o a CS degree cannot code well. I'm saying that it is more likely that someone with a CS degree will want to. Unless they're bored out of their skull and then they need to be assigned some other project/language/platform or design work for a challenge.
The best preparation for the real world is an ability to learn and ability to think. Any actual trade skills learned in college will not be useful for too long. That is why experimental physics education is one of the best ones you can get for a very large variety of jobs. You learn how to learn.
<^>_<(ô ô)>_<^>
There's still optimisation at Google and other search engines. Speech recognition and synthesis at Microsoft. Handwriting recognition at Microsoft. Noise cancellation at Microsoft. Translation at Google.
CG at Pixar, Disney, nVidia and ATI. (I'm not sure how much software-based work there is at the last two, though...)
Plus all the technologies I mentioned are simultaneously being developed by many other companies!
Well, as someone with 20 years experience as a software engineering on everything from desktops to supercomputers I will tell you that the math and theory is what separates the men from the boys when it comes to computers. I don't go back to CS theory that often - maybe once or twice a year. However, when I do need to pull that stuff out it is invaluable.
When I was in college I often lamented the lack of practical experience I was getting. Today I'm glad that I got a solid grounding in the theory of CS rather than a lot of classes about how best to optimize PDP-11 assembly code. Technology and training goes obsolete. A solid theoretical base lets you keep up with the constant change in the computer industry and keep your knowledge of the technology current.
For Americans, and other aliens, I will now divert into some history. In the UK, we used to have two kinds of higher education establishments; Universities and polytechnics. Universities concentrated on academic courses, while polytechnics taught vocational courses. Then the Conservative government decided that it was discriminatory that not everyone got a 'university degree,' and so turned all of the polytechnics into universities. Now, instead of teaching good vocational courses, they teach bad academic ones. Worse, as appears to be the case with this guy, teach their old vocational course with an academic-sounding name. It sounds like he was teaching an IT course back in the polytechnic days and is now teaching the same course but calling it computer science.
My course had about a 30-50% attrition rate[1], although most of those who dropped out did so very early on in the first year, once they had realised that they didn't want to be doing Computer Science at all; they wanted to be doing IT, or possibly Software Engineering.
You can impart the fundamentals, which is what a degree in computer science is supposed to be about. Exactly right. The article said: What has changed is the need to know low-level programming or any programming at all. Who needs C when there's Ruby on Rails? A computer scientist doesn't need to know C or Ruby on Rails. A Computer Scientist needs to know about Unlimited Register Machines, Turing Machines and Lambda Calculus. A Computer Scientist needs to know about type theory and graph theory. Once you know these things, C and RoR are just syntax. Someone completing the first year of a Computer Science degree should be able to pick up a new language in a couple of days. Beyond that, being a good programmer is simply a matter of practice. Someone doing an IT degree will get a lot of practice in one or two languages, and so they will probably be a better programmer (in, say, Java or C#) than a fresh Computer Science graduate, but the Computer Science graduate will have the edge when you change the language that you need to use.[1] Note: unlike the American system, students in UK universities are enrolled on a particular degree scheme for 3-4 years. If they drop out, then they can not simply change their modules and get a different degree, they have to re-apply the next year to a different department.
I am TheRaven on Soylent News