Blog » In Praise of Elitism
Posted on 15 Feb 2009 12:17
A respondent to one of my earlier posts complained that approaching Haskell by way of category theory is a "radical approach that would lead to elitism", since such an approach is likely to scare off all but the mathematical sophisticate.
That may or may not be true - in my opinion a well-written CT-oriented intro to Haskell need be no more terrifying than any other - but I take mild exception to the implied assertion that "elite" is somehow bad.
The community of Haskell programmers (and functional language programmers in general) is a natural, self-selecting elite. There's nothing wrong with that; on the contrary, the elite status of the Haskell Master should be celebrated. Haskell is hard, there's no sense in denying it, but so are most things worth doing. It's hard to master the piano, or fly-fishing, or brain surgery, or a foreign language, and nobody sensible would pretend otherwise. Why should the highly sophisticated art of functional programming be any different? Elite does not mean snide, or exclusionary, or condescending, or selfish; it means excellence, and the truly elite are open-minded, inclusive, and generous with their hard-earned knowledge.
I came to Haskell with years of imperative programming under my belt. I became a pretty "advanced" programmer, but I never really felt like I was doing anything a clever monkey couldn't do. A C programmer spends most of his time doing much what an auto mechanic does - diagnosing and fixing problems in a sophisticated machine. There's nothing wrong with being an auto mechanic or a C programmer; but it's not exactly an elite.
On my third try I've finally begun to appreciate the beauties of Haskell. The turning point for me was category theory, which lead to a sea change in my thinking. I'm not sure I'll ever become a master Haskell programmer - one of the great things about it is you can acquire a pretty good understanding of how the language works without writing Haskell programs - but it's clear that those who do master it are a different kind of programmer. Show me a great C programmer and I'll show you a Thomas Edison; show me a great Haskell programmer and I'll show you a J. S. Bach. (Idea for a logo: a measure of musical notation with a lambda clef, and notes tied by arrows.)
The message to newcomers should not be "don't worry, it's just like PHP, only different". It would be more honest to state right up front that this is a different game: entree to the elite is open to anybody, but not for free. You have no right to demand excellence without effort. You have to work; you have to master some mathematical ways of thinking that can in fact be difficult even for those with a good deal of math. You can do this; many others have. But if you're the type to throw a fit the first time you discover that your customary ways of thinking won't work, then please, go away. This is for grown-ups.
If you value excellence, and are motivated by the prospect of rising above the common dross and attaining excellence yourself, then you are in for a great intellectual adventure of deep insights and dazzling discoveries. And you will find lots of help. There is a plethora of material on the web that will help you, and the community of those who have gone before you is open and more than willing to help those in their apprenticeship.
In my experience the best, most imaginative and talented people are stimulated by a challenge to excel. They want the hard stuff, and they're the ones who should be evangelized.
The message to business people should be similar. No, Haskell is not PHP on steroids, it's a different beast altogether. It's a major paradigm shift, with major benefits that are well worth the costs. You hire expensive, highly-trained MBAs to professionally manage your company's financial assets; why would you not hire elite programmers to manage its information assets? The importance of husbanding knowledge assets to gain competitive advantage is only going to increase; treat it as mere data processing and your competitors will eat you alive. Yes, it's harder to find a good Haskell programmer than it is to find 100 mediocre PHP hackers, but aside from being more productive, the Haskell programmer comes with a disciplined, highly sophisticated way of thinking about information structures and algorithms. This makes it much more likely that your Haskell staff will do a much better job of figuring out how to turn your raw data into real capital that returns value.
(Fortunately I don't have any money riding on my sales pitch.)
Like this entry?