Sucking at CS is a depressingly easy thing to do. If you have trouble with math, logic, or writing, your odds of excelling in computing science are pretty small. If you don’t like pulling things apart until you can see their fundamental structures, you’re in trouble. If you didn’t play with Lego and don’t see why anyone would, you aren’t doing so well.

There’s nothing inherently wrong with sucking at CS, of course. Many disciplines are at least as hard, and as easy to suck at, as CS — math, music, physics, painting, and motorcycle racing are the first five to come to my mind. The world is full of terribly difficult pursuits. Most of the people who excel at these pursuits are deeply and intrinsically motivated to them — you’d no more be able to keep Knuth from writing code than you’d be able to keep Mozart from writing music.

I’m not going to address the question of whether these talents are innate or acquired. One can surely be born with a mind that works well at computing science, just as one can surely develop such a mind through diligent effort. Similarly, mutatis mutandis, with intrinsic motivation.

The problem is that, thanks to the Computer Revolution and more recently the Dot-Com Boom, a thundering herd of otherwise perfectly talented people have decided that an education in computing science would get them a great (or at least stable) job and oodles of money in stock options that might (one day) vest into something more useful than ass-paper. These people are extrinsically motivated, and while there’s nothing wrong with wanting a decent job that might make you an unlikely bazillionaire that wanting isn’t going to make you a competent computing scientist.

I’ve written about this before, of course. Joel Spolsky started the ball rolling (for me, at least) with:

• The Perils of JavaSchools

More recently, other people have weighed in, chiefly due to a paper by Robert Dewar and Edmond Schonberg:

• Where Are the Software Engineers of Tomorrow? (original paper)
• Who Killed the Software Engineer? (Datamation)
• Who did Kill the Software Engineer? (Ekinoderm)

In short: we have too many CS students, and most of them suck.

So the thundering herd mentioned above thunders its way into CS departments around the continent, dramatically raising enrollment figures. This is, essentially, Christmas in September for those departments: more students means more money (from the university administrations and, in many cases, the government) and more prestige. It looks great on paper for the department and for the university.

Of course, there’s a problem: most of these students can’t handle the tough and often mindfucking subjects that are central to good computing science — formal languages and automata, propositional and predicate logic, data structures, algorithms, calculus, linear algebra, graph theory, and above all the mental discipline of decomposing a problem into a complete set of mutually synergistic and tractably implementable subproblems (among others). Department administrators can’t allow these students to fail: to do so would damn them just as high enrollments exalted them. The students themselves are rarely motivated to throw themselves into these problems: they just want a good job, not (necessarily) an intellectual challenge, and while the theory of computing science is apposite to their jobs it doesn’t seem immediately “relevant”.

When “relevance” (as measured by reading job advertisements and counting the buzzwords) meets curricular revision (with an eye towards lowering the failure rate), they have drunken monkey sex and the end result is a JavaSchool. Lest you think that Java is the only problem, I’ll quote from the Datamation article:

To sum up Dewar’s argument: today’s college computer science programs aren’t rigorous enough, and don’t promote in-depth thinking and problem solving. Instead, in an effort to boost enrollment, CS programs focus on easily accessible curricula, and so fail to prepare students to compete with their international peers.

One of the article’s main points (one that was misunderstood, Dewar tells me) is that the adoption of Java as a first programming language in college courses has led to this decline. Not exactly. Yes, Dewar believes that Java’s graphic libraries allow students to cobble together software without understanding the underlying source code.

But the problem with CS programs goes far beyond their focus on Java, he says.

“A lot of it is, ‘Let’s make this all more fun.’ You know, ‘Math is not fun, let’s reduce math requirements. Algorithms are not fun, let’s get rid of them. Ewww – graphic libraries, they’re fun. Let’s have people mess with libraries. And [forget] all this business about ‘command line’ – we’ll have people use nice visual interfaces where they can point and click and do fancy graphic stuff and have fun.”

Dumb down the curricula (often at the explicit request of many of the students), and you dumb down the graduating class. You’re still going to graduate some really sharp computing scientists — the sort of people who go on to make a bazillion dollars by applying that recondite CS theory — because you simply cannot stop them from constantly and aggressively improving their art. On the other hand, you’ll graduate a thundering herd of half-assed programmers, who all go out and get jobs at Microsoft and IBM and the like.

You run into real problems when Microsoft and IBM and all their little friends realize that they can hire better programmers for less money by outsourcing their projects to places like India. This doesn’t trouble the really good programmers — they’re all trying to get jobs at Fog Creek or money from Y Combinator or finish their Ph.D. dissertations. It does trouble the thundering herd, though, because they went into CS as a career, not as a productive and occasionally unhealthy addiction.

And just like that, CS enrollment drops through the floor.

“Quick! Enrollment’s dropping! Take out the hard courses! Make CS more fun!“

Those of us who’re into the theory recognize this as a positive feedback loop. It’s not a good place to be. Lower enrollment means dumber programmes means less-employable graduates means lower employment….

The irony of all this is that it’s at best a sideshow to the people who’re inexorably drawn to computing science. Those are the people who drive computing progress and create improbably successful startups every decade or so; they are the people whom the thundering herd seeks to imitate. They’re still signing up for CS degrees, they’re still relentlessly improving themselves, and they’re still doing the vast majority of important, innovative computing.

(And they’re probably still playing with Lego.)