The programming talent myth

> What is the point of stating that though?

That, contrary to positions that portray programming as a skill only valuable to those who would be able to use it to become the equivalent of an exceptional artist, that the number of people who could derive benefits (professional/financial -- not to consider other personal benefits that might accrue) from learning programming is quite large.

> It's pretty hard to know ahead of time and talent is rarely worth more than sheer effort.

I'm not sure that there is a clear line between talent for a particular task and the ability to maintain concentrated effort directed at it, nor do I think that it is easy to pull apart the effects of each. But, sure, I'd agree that you can't know who can benefit the most from lots of things (programming included) ahead of them trying it.

Which is another strike, IMO, against those who would dismiss efforts at universalizing exposure.

"or that 10x chef"

I like the chef example. I know from past discussions here that there is a large contingent of HN posters who are terrified of their own kitchen, which is a shame.

I think in small organization you might need a lot of talent, but once organizations (or projects) get bigger, you need process over talent and that, well, also means good management. There are things you can do with a small group and things you can only do with a bigger group. I guess Django is at a point where it's big enough that it can take advantage also from the less talented coders. Maybe, and I am wild guessing here, part of the issue comes when startups grow and still look for the 0.1% talents, when at that size that can easily live with the 40-50%.

Especially if one don't even define what is the exact quantity measured. IQ is a normal distribution by definition, (IQ-100)^(1/3) is "bi-modal", it's just a simple transformation and both quantity correlates (supposedly) with the vague term "intelligence".

The question is not in bi-modality though, it's whether you can go from the left side to the right or the middle over time or you should have been born as a rock star.

Programming talent isn't bimodal, in fact it isn't a question of traditional talent at all.

There are only two levels of programming ability: "Bad" and "Good Enough". In all but the rarest of edge cases, everything else is a matter of preference. You think that someone's code is elegant, simple, clean, optimized, etc? That's great, but only other programmers might care about those things. The only question that the majority of people care about is, "Does it work?"

Obviously code can be written so poorly that it causes noticeable performance problems, or doesn't work at a scale past the initial requirements. But that falls under the "Does it work?" question. Basically if I'm using it and it isn't annoyingly slow, whoever wrote it was "Good Enough". I don't care if they were a l33t 10X rock star ninja or some guy who just finished a "$language For Dummies" book.

It is genuinely cool to see code that is well designed and well written. And doing that consistently does require the innate talent of a special mind. But let's not confuse elegance with doing the job. Doing the day in day out work of programming is like being a journalist: the most important thing is getting the facts straight, not writing something pretty. The truly talented can do it with the rhetorical flourish of a poet, which is awesome, but not required.

All that most people ask of code is that it's good enough to do whatever it's supposed to do. If we want to turn it into a poetry contest then it's definitely a bimodal distribution, and most of us will end up on the wrong end of the curve.