I've had this position in other comments, but I'll say it again - College degrees don't teach you how to do a job, and don't necessarily equate to job performance. To that end, it really doesn't matter what kind of degree someone has. A college degree is about broadening horizons, teaching critical thinking, and exploring subjects in slightly more depth in a controlled environment. I had taken plenty what we now call "STEM" courses in the process of pursuing a Harvard undergraduate degree (of which they don't offer a "science" degree in the classical sense anyway!) A greenhorn college grad will have been exposed to many valuable situations, and a college degree says that they can think and have proven that to some number of accredited boards to their satisfaction. They will still need job training, additional learning, and just plan ole' experience. Some people will be just better at certain types of jobs, and not at others. Do people who choose a particular degree type self-select? Maybe, but there are plenty of medieval lit majors out there programming, and they do it just as well as an EE or CS major.

Oh, and grad school is about torture and the ego's of the adviser committee. It means you spent a lot of time as a serf eating ramen noodles. It may mean you know a lot about almost nothing...

We've been arguing this for more than 20 years. Not much has changed, and it's not a new question. Code Slinger vs. Book Knowledge. College of Hard Knocks vs College of Ivy. I'm a greybeard now, and while I won't pretend to answer the whole question, I will provide some perspective...

I was a code slinger type - Right out of high school with some programming knowledge, some commercial success (with the C64), and whole lotta balls. I did some college, but it wasn't for me at the time. It didn't connect with what I wanted to do, which was code. I joined a contracting house, and they sent me all over the country. I learned more in 10 years doing that than any college would ever teach. Databases, Integration, GUI's, network programming, mulithreaded programming, and real-world problems, both programming and political. C, C++, Cobol, Fortran, BASIC, assembly (various), and eventually Java.

In the late 90's, I went back to school. Why? Not to learn programming - I was already at the top of my game. I went back to learn all the other stuff, and to do other things. I took psych courses, math courses, art classes, electronics, music, law, languages (Living: French, Dead: Nahuatl) ... I did it on my terms (Harvard Extension, no time limits.) I will graduate next year.

Do colleges teach some basics? Sure - Data algorithms and Graphics programming were very useful. Are they realistic? Not really - sometimes horribly so. Massively Parallel Programming was a mess of math decomposition problems I dropped quickly. Did I need them to enter a career of commercial programming? Nope.

I would say college education is not a prediction of coding ability. Having a college degree when you are entering the field can be useful, but having a CS degree IMHO is not any more useful than a general BA or BS. If you go to college, go to get a general education, learn how to think critically, expose yourself to some interesting things - but it is NOT a training program for coders. Technical schools are a whole 'nother thing, and I would avoid them like crazy. My experience is that they do train you, but the training is narrow and short-sighted. In the end, it would be throw-away time, and the student would have very little gained.

College? Sure - go do it. You will be a better person, and you will have some great social experiences. But if you want to code, you need to put the time in yourself and learn the skills. College won't teach you that.

No, I can imagine how electronics can be designed to handle audio in a PC environment. But I realize that it is not often not done well. Yamaha made a very nice series of audio cards in the 1990s that were clean and well designed, for instance. I think in the end it comes down to the price/performance tradeoff - there is not a need to provide top-notch audio on a motherboard because mostly people would not appreciate it enough to pay for it. It's simple enough to provide the digital chips, but the analog part costs a bit more, in both space and money. And my top-notch preamps that I use with some performances occupy boards almost as large as a typical small-factor motherboard all by themselves.

Integrated audio isn't good enough, isn't great, and isn't for me. I have a pro-level sound studio, and there's no way your going to tell me that the noisy environment that is the system motherboard is going to give me results I can be proud of. Not even for gaming, thanks.

Discreet card? Ok, maybe, but generally you need to jump up to RME or some such before you can really call it good. I have a an RME RayDAT - This means that that all my AD and DA happens somewhere else, and not in the computer. It all goes digital over ADAT to my mixer (a Yamaha DM2000) where the conversion happens. Or it goes digital over ethernet (audinate Dante) to an X32, again where the conversion happens.

There are a ton of good external boxes to handle sound - some quite reasonable. Stay away from the onboard and cheap USB sound dongles. If you have the speakers to handle it, then why put up with bad sound?

I'm not an expect in functional programming, but I am an expert in other (object, etc) styles. While I appreciate the functional toolbox in languages such as Scala (which I use every day), I don't really see a way to do my day to day job in a purely functional way. Others have mentioned the I/O dilemma, but I think it goes deeper than that. Functional != Efficient for many of the tasks I perform, which are rather iterative. For many of my tasks, the overhead of the functional structures required are either much more memory intensive, or impose a run-time overhead that isn't acceptable. In the end, when what I have to do is move 300 fields from one data structure to another with edits, COBOL would be sufficient...

In related news, Mr. Greenspan has no clue about inequity in stratified markets. If you push on the top, you just compress the layers into smaller layers, with the bottom filling until it can absorb no more. Then you get slums, riots, and chaos. The only way the market works is with a strong middle class with buying potential. Without that there is no market, and hence no profits or growth. Once that contract is broken, it's not a long way to the bottom for most.

FAIL. Try Interplanetary.
Listen to the Beastie Boys...

Interstellar would be a cool trip also, and more probable of finding life than under 100 miles of ice on Europa. Of course there is that extra mileage charge on the rental, and the roaming fees would bankrupt you...

I was a big fan, and a game developer for the C64. Those were the days that a machine could be fully understood by an untrained person with a knack for programming. When the C128 came out, I was interested, especially in the 80 column screen and CP/M software compilers. But there were too many limits on the machine (no hard drive easily added, no real OS, etc.) and it didn't feel like enough of an advancement over the C64. My grandfather did buy one, and I had some time with his, but that never really sparked much either. My next machine would be the Amiga, and as soon as that become somewhat affordable by a college student (the A500), I never looked back.

RM

This is all well and fine, until they herd us all into some kind of processing center and then hook us up like some kind of "D" cell in series to power the mastermind machine...

Richmond Science Museum, on the E&S DigiStar projector - we could play a space-war variant on the dome. No color, of course, but the resolution was pretty good if my memory serves me right. Plus the dials of the control panel were just about perfect for controls.

Drain the fuel, set it upright, patch it up, tow it to Atlantic City - Profit!
(Drop it Lake Mead - Profit!)
(Park it outside Boston - Lawsuit!)

Among my customer base? Yes, it's used internally. A lot of them are IT shops dealing with very old equipment, like 10 year old PC's. Some of them have internal intranet apps that only work on IE6. It will be awhile before those move.

C

I'll pop the cork when my customers get off IE6. Until then I need to sink development resources into maintaining and testing on IE6, no matter how painful it is.
Unfortunately my customers' IT departments are slow moving and not motivated in moving quickly off XP and IE6. Most of them are understaffed and underfunded and dealing with PC's that are sometimes more than 10 years old. I suppose they have more pressing problems, given that...

C
 

I think the fallacy in this argument is not that quality doesn't win out, but that quality isn't always important.
The problem is that the determination process is flawed.
I might make the decision that I need lesser quality (whatever that means) for an internal time-keeping application than I do for something customer-facing, such as my sales portal. The article is of course arguing that I shouldn't be making that decision based on initial cost but on longer-term factors, but on the management side of things as long as I've got a fixed budget rooted in the short-term I can't make that decision equally. Like many financial equations, X dollars today vs. X dollars tomorrow is in play.

C

Yeah, but there's rules for them:

1. Don't optimize.
2. Don't optimize YET.

Rule 1 is for type 1 - and is generally the best case. Then you can come along and after rule 2 has expired, make the improvement where it matters. Type 2 fools skip both rules and make a mess.