How to Keep America Competitive

pkbarbiedoll writes to tell us that in a recent Washington Post article, Bill Gates takes another look at the current state of affairs in computer science and education. According to Gates: "This issue has reached a crisis point. Computer science employment is growing by nearly 100,000 jobs annually. But at the same time studies show that there is a dramatic decline in the number of students graduating with computer science degrees. The United States provides 65,000 temporary H-1B visas each year to make up this shortfall — not nearly enough to fill open technical positions. Permanent residency regulations compound this problem. Temporary employees wait five years or longer for a green card. During that time they can't change jobs, which limits their opportunities to contribute to their employer's success and overall economic growth."

Re:Overworked?

[Skreems]

I don't know who you talked to, but that doesn't seem like a very fact-based view of the computer science field to me. But hey, what do I know? I just work in it...

Re:Ha ha

[morgan_greywolf]

...and Microsoft will do anything to solve this "crisis" except spend money on it.

Ever heard of the Bill and Melinda Gates Foundation? http://www.gatesfoundation.org/default [gatesfoundation.org]

Addressing educational inequities, especially in the United States, is exactly what they do.

Re:Ha ha

[MrAnnoyanceToYou]

What does that have to do with Microsoft? And how is that contributing to the problem Bill is whining about? I seem to remember a lawsuit a few years back attacking Microsoft over calling people who weren't engineers engineers. I seem to remember people being encouraged to not finish their CS degree so MS didn't have to pay them as much in the long run. I seem to remember twenty years of vicious market monopoly abuse. Two or three years of giving a little bit back doesn't make up for being a robber baron for twenty. In fact, I don't know if a hundred will, the way the foundation manages itself.

Re:Economics lesson for Billy

[0123456]

"Um, he majored in economics. At Harvard."

So it's no wonder he doesn't understand that 'if you pay peanuts, you get monkeys'.

I'm constantly amazed at how little 'economists' know about economics, and how poorly their predictions turn out.

Re:Au contraire

[bmajik]

I work at Microsoft.

I know very few 80hr/week employees. As in, i can't think of any right now.

Microsoft doesn't have a problem finding applicants. Microsoft has a problem finding _qualified_ applicants. I've done a bunch of interviews. We interview _way_ more people than we hire. And I don't even want to think about the people that _don't_ make it to me and don't even pass the HR and phone-screening stages of the process.

We want good people no matter where they come from. There is no particular focus on H1-B workers. Given the extra paperwork and overhead involved, and the legal restriction that they get the same pay, etc etc, don't you think we'd rather not deal with the extra hassle?

Bill Gates Wants IT Salaries Lower

[ProteusQ]

That's the alternate headline. It's basic economics that if there's a labor shortage, wages go up. But Bill wants to stay a billionaire, so he'd rather pull an Al Gore and create hysteria rather than acknowledge certain facts. Double IT salaries, or even just double benefits, and its doubtful that the shortage will continue.

Re:Economics lesson for Billy

[JesseMcDonald]

H-1B visas artificially increase the labor supply while decreasing wage growth.

Rather, the requirement for H-1B visas artificially restricts the labor supply, raising wages (but reducing wealth), whereas an increased supply of visas allows the labor supply and wages to grow toward their natural levels.

Re:Au contraire

[bmajik]

Qualified means "people we make offers to", pretty much by definition. We talk to a lot more people than the number of people we choose to extend offers to. Ergo, we have a problem finding qualified applicants.

The only caveat is that there are probably a set of people out there that would be qualified (i.e. we'd hire them) but they won't talk to us. I don't know how large that set of people is.

What I can tell you is that there are plenty of people who _do_ interview with us who we feel are not qualified to join us... at least at the time of the interview.

I've never paid attention to someone's degree status during an interview. I look at their resume and see what they say they've done. Then I ask them about it. Then I ask them a few other questions. I can't speak for the layers of recruiting that come before me - they may have an unhealthy fixation on university degrees. But I personally do not, and it's also something that never comes up amongst the other interviewers I talk to.

Re:Au contraire

[bmajik]

You'd be shocked at the number of people that can graduate college that can't answer a question like:

"implement $(randomly selected function in the C string library) on this white board. Use whatever language you like. It doesn't even have to be a real language if you can explain it to me and it's coherent"

You'd be shocked at the number of people that just draw a blank when you say:

"describe how you'd test a coffee maker"

These aren't especially interesting or novel questions. It's a shame you even have to ask them except that you still meet people that CANT ANSWER THEM!

As an aside, based on your knowledge of what makes a good candidate, and what makes a poor interview, what would you suggest asking potential candidates for developers? For testers?

Re:Au contraire

[gfxguy]

All good questions, and ones that Bill Gates should have answered in his commentary.

But since you mentioned Wall Street, you will probably find the other article [washingtonpost.com] very interesting, and I don't want to make you search for it. Here's a quote:

On average, American lawyers make 42 percent more than chemical engineers. At elite levels, huge pay gaps also exist. In 2005 the median starting salary for a new Harvard University MBA was $100,000. An MBA is a two-year degree. By contrast, a science or engineering PhD can take five to 10 years, with a few years of "post-doc" lab work. At a Business Roundtable press briefing, one CEO said his company might start this sort of scientist at $90,000. Does anyone wonder why some budding physicists switch to Wall Street?

Re:Don't even go there, Bill

[tom's a-cold]

Maybe the real reason that IT is "suffering" is that companies often don't treat their IT employees like real employees.

In other words, they treat them like their other employees. The whole H1-B program is a scam to depress the salaries of American software people. I've done a lot of work in Silicon Valley and it's become increasingly clear to me that it's not about a skills shortage, it's about being able to pay a guy from (let's use a common example) India a lot less than an equivalently-skilled American would demand, and have that guy unable to leave your company without facing deoprtation, and yet have no job security. A win-win for an unscrupulous employer. No inconvenient unfair dismissal suits. No saying "screw you, I'm going to work for the competition." It's indentured servitude, plus a little "cut chicken to scare monkey."

As for skills, H1B people have a bell curve too. I have met a few great designers/coders, and a lot of schmucks who in a previous stage of IT evolution would have been COBOL drones. Everyone works hard, H1B or US citizen, but so what? My job is to clean up the product of hard work when it has been applied to piss-poor IT decisions. The real problems with IT in the US have to do with pig-ignorant short-term-obsessed management who believe contrary to all experience that free lunches exist, spineless IT managers who go along with the knee-jerk "can't afford to think strategically" mind set, consulting firms selling flavor-of-the-month solutions, and the received wisdom that IT people are in some sense interchangeable. If that's true, you're hiring the wrong ones.

It's not the H1-B person's fault. He/she is doing better by coming to the US despite the abuses, and there's the prospect of a green card a few years in the future. But it puts American professionals in a race to the bottom with respect to working conditions, wages and job security. The best thing for the US would be to shut the H1-B program down. The depressed wages provide a disincentive for anyone trying to enter IT in the US, the exploitation makes working conditions worse for everyone, and the H1-B workers who return home take the knowledge with them. If it's about money, then perhaps top executive compensation would be a better place to look for savings. Our foreign competitors don't reward non-performing execs nearly as lavishly, and there's no evidence that you get better work for $50M per year than you do for $30M. And there's a lot more savings to be found there than in putting the squeeze on IT people.

Me, I'm middle-aged and doing well in the business, but I've advised my son, who's more talented than I am, to do something else for a living since this gravy train has long ago moved on. US corporations and their kleptomanagers have already poisoned that well.

Re:Au contraire

[bmajik]

And unless you recently finished a CS class, then remembering a good implementation of any random C function is not going to be on the top of your head. So, it's all but guaranteed that the MS interviewer will rip apart the answer because it wasn't very good because it came about through a completely bad process. Big surprise.

This is not about _remembering_, it's about deriving. If someone knows the question off the top of their head, we try something different. If someone cannot derive an implementation of a string function, they're not an interesting candidate. _Especially_ if they're interviewing for a position with the BCL or other platform/framework type group.

"Ripping apart" answers isn't something we do. Rarely does someone issue a perfect answer on their first try - both in interviews and in the real world. For almost any answer someone gives, there is some possible drawback or "gocha". What is the memory consumption of your routine? How many conditional branch statements would it require? Asking these follow-on questions are what makes it a less-worthless question, and seeing how someone thinks about the implications of their decisions and describes the tradeoffs is what makes it worthwhile.

That's because the first thing they think is, "Wow, what a stupid question." And then, "What? Oh, he's actually serious." Then, "do I really want to work here if that is the best interview question they can come up with?"

That's a fine response to have, but i'd ask you to justify it. Why is it a stupid question? Obviously, i'd ask it as an allegorical question to the problem of how to test software. Fundamentally, a coffee maker is something many people are familiar with, so its something that doesn't require significant introduction.

It's not the "best" question. It is _a_ question. And i'll ask you again - justify why you think it is a poor/irrelevant quesiton?

Having actually conducted several interviews for senior developers, you need to do two things: see if you like them and see if they can actually do the job.

We agree so far. Although i'm not sure about "liking them".

You look over their resume and verify what's listed there.

Done.

You ask them about their previous positions and one major problem they encountered in each place. I then like to ask relevant questions that have come up recently in my own work

I'm with you there. Sometimes, these are college hires. Sometimes, these are people that haven't had previous work experience.

The questions are high-level, don't require writing on a whiteboard, and are difficult enough to tell whether they're bullshitting their way through an answer or not

I find that the opposite is true -- people that are unwilling to delve into the details of an answer.. people that keep things "high level" are bullshit artists. The saying "The devil is in the details" is a saying for a _reason_.

We don't have a perfect hiring philosophy. I'm not sure where your animosity comes from, however.

Re:How about the 17-year education lag?

[Vomibra]

General Education requirements at my college are pathetic; most of them can be skipped by taking AP tests in high school. A quick summary:

• Two writing classes
• One oral communication class
• Three classes total in mathematics and the natural sciences (this includes remedial mathematics and introductory sciences)
• One literature class
• One non-literature humanities (philosophy, music, foreign languages, etc)
• Two social studies classes (economics, history, psychology, sociology)

These are pretty basic requirements to function as a useful member of society; your crack at "Shakespeare appreciation" shows that you didn't really understand the purpose of literature courses--to teach reading comprehension.

Re:Au contraire

[bmajik]

I didn't get the first set of positions I interviewed for :)

I'd guess that was 50% me being a "normal" coder instead of a "legendary" one.. and the other 50% was my attitude.

One of the questions the HR lady asked me was

"so, how are you as a C programmer.. on a scale of 1-10?"
"9"
"Ok, what would make you a 10?"
"I donno.. i could have the lang spec memorized or something... i haven't written a C compiler yet"

Yeah. She knew that I'd be interviewing with people writing C compilers, but I was too arrogant to connect the dots. Maybe every snotty kid who's used to being "the computer guy" and who doesn't have to try very hard in school runs into reality.. I certainly did :)

The interviews didn't go very well. My up-until-then-invincible-ego thoroughly crushed, I was given a second set of interviews for a different type of position and that went much better.