Hacker News Reader

Comment by JoachimSchipper | Hacker News Reader

Why do you feel that's unfair? If you've only ever done Ruby on Rails, never wrote any algorithmic code and never done any algorithm course, you will indeed score low in the "Computer Science" section. Apparently you managed without those skills, but that doesn't mean that the assessment is not accurate.

View on HN

Not what I meant. I'm familiar with several dozen languages and I try to get acquainted with at least one topic and programming language outside my comfort zone every six months. I've been keeping with this routine for about 3 years now and what I've realized is that almost all those things he mentions have absolutely nothing to do with being a good programmer. There are certain meta-cognitive patterns that are far more relevant for being a good programmer and it's hard to gauge these meta-cognitive abilities by just running through a grid of requirements.

But if you're filling a RoR web development position does it matter?

Almost certainly not - which is why these kinds of tables are so silly.

You'd be surprised at how often a cleverly chosen data structure or algorithm can improve a web app. For instance, where I work our frontend code uses tries, heaps, and other data structures (and using these makes the user experience noticeably and measurably better than using slower data structures).

Um, if you don't care about the CS section, do not ask the CS section questions.

That is why the matrix is conveniently broken into specific sections.

But if you're filling a RoR web development position does it matter?

Yes, it most certainly does. In the long run, the developer with the better grasp of fundamentals (e.g., discrete maths, data structures, algorithms and so on) will produce better code. NB, I'm not saying that the "naive" web coder can't produce good code. Nor am I implying that CS education always makes you a better developer. On average, however, that's exactly the case.

Even with generic web 2.0 stuff.

The fact that people with a CS education are better coders does not imply that the CS education made them better coders. Without more evidence it's just as likely that people who are, or will become, good coders anyways tend to get CS educations. Do you have any evidence besides correlation that a CS education makes you a better web developer?

I guess you're either asking one of two questions, both which seem strange:

1) Do you receive magic fairy dust during college that makes you more productive and removes bugs from your code? No, I don't think so.

2) Will someone who knows the things taught in a typical 4-year undergraduate CS education be a better web developer? Obviously (?) yes; you can look at any of them and clearly see that they are good things to know. Will you point to some particular thing that you think is useless? Let me list some that seem reasonable for a lot of web development: non-trivial data structures, for understanding databases; experience with multiple languages; continuation-passing style; general system and network architecture (e.g. the OSI model) -- and so on.

If by "web developer" you mean writing http://www.hasthelhcdestroyedtheearth.com/ then I guess no, you don't need to know much about anything for that.

I wasn't arguing that CS fundamentals wouldn't make you a better developer. I was just asking for better evidence than a mere assertion, which you provided by citing specific examples. So, thanks, that's informative.

Reply on news.ycombinator.com