You're confusing "engineered," as in orderly, organized and planned according to specifications with "legacy" systems. A little more time in industry will show the difference.
A CS graduate will know about a 50 ways to solve a problem and will have the ability to find the most optimal way to do it. He'd probably take twice as long as the hacker school graduate who'd solve the problem using the one canned method he was taught in hacker school.
"Getting away with it" == using another solution.
> A CS graduate will know about a 50 ways to solve a problem and will have the ability to find the most optimal way to do it. He'd probably take twice as long as the hacker school graduate who'd solve the problem using the one canned method he was taught in hacker school.
Disagree. I rarely find someone with a degree to be substantially better than someone self-taught on the job. I've even seen developers with degrees do far worse.
Not knowing recursion for 20 years is more "getting away with it" then just simply "using another solution"
>Disagree. I rarely find someone with a degree to be substantially better than someone self-taught on the job. I've even seen developers with degrees do far worse.
Sure in jobs where the most optimal solution doesn't matter there is no correlation with performance and degrees. This is probably the case for many jobs. However for a job where it does matter, say the programmers behind the V8 engine or people implementing Ruby and the python features, a degree or knowledge about theory makes a huge difference.
Its disingenuous to think everyone needs to obtain that level of knowledge, at substantial cost, when only a few are going to use that knowledge in the scenarios you described. Not everyone needs to know how to create programming language or write a complier.
We should not look down on those who are bootstrapping themselves in the industry without the exorbitant costs of a CS degree lest we retreat to the mindset of ancient guilds.
This knowledge is usually what separates someone with a degree and someone without one.
I don't look down on someone without a degree, but I have no doubt in my mind that such a person will have a higher chance of being less capable of implementing high concept applications like a ray tracer, a 3D engine, search algorithm, operating system, or a machine learning system, ... etc. These jobs are not so obscure that only a 'few' people need to know it, as you seem to indicate.
In fact there's a popular new buzz word in CS called "Big Data" where those involved need a huge amount of theoretical knowledge that cannot be gained from simply attending a hacker school.
Someone define recursion here... are we saying this guy didn't know what a recursive loop was for 20 years, or are you using "recursion" in some other way?
edit what i mean is this: did the guy have a hard time defining recursion in terms of CS, or could he flat out not demonstrate it with some pseudo code?
He was a python programmer. Typically in python, you shouldn't use recursion anyway.
You proved my point, though. Programming is only one small aspect of computer science. Engineering, mathematics and theory is the biggest part which you can't get at a 6 week school.
A CS degree is one way to learn about algorithms and core CS concepts. It is not the only way.
This is a dangerous idea to those who spent 4-6 years of their life in school before they started receiving real life experience.
Agreed. Sadly lots of people are getting sucked into the start up tech bubble. Sure you can get a really high paying tech job on limited knowledge as a web developer at some shitty start up, but it might be gone tomorrow and you'll find yourself SOL.
Having gone the no CS degree route myself I can say that I find myself in a constant loop where I have to prove my abilities even with the recommendation of former bosses and a nice resume.
EDIT: Actually, the fact that you mention "large scale engineered systems" is interesting. Generally, I have found large, monolithic, tightly coupled systems to be very unmaintainable. "Building a small app" (or a bunch of them), along with in-house libraries tends to be the way to go if you want maintainability. "Building a large scale engineered system" might actually be a red flag when it comes to code reusability and quality.