Even if you work independently on the code itself, you still have to collaborate with others on the general design, the features, the interface, and maintaining the code.
Zed has plenty of criticisms of other people and developers. Most of them seem pretty valid on face value.
But I get the impression that if I sat down and offered to pay him for help on some library for our code he would listen for two minutes and then dominate the conversation. Two days later he'd deliver something better than I could write, that I could not understand, which probably did not fit my project, and which I could not maintain. Then he'd take his money and be on his way, patting himself on the back for his brilliance. I'm left with a high powered nail gun with a Japanese owner's manual when I wanted a beer can and a simple 'pull tab to open' instruction.
Unless you're writing the code solely for yourself, your ability to collaborate on some level is crucial.
According to this: http://www.nei.org/newsandevents/1999productionrecords/ the 103 commercial nuclear reactors in the US provide 20% of the nation's electrical energy, and wikipedia asserts that a typical nuke plant produces 1 GW of power. So presumably about 100 GW of electrical generation is 20% of our electrical needs.
Also from wikipedia, the Nevada Solar One project makes 64 MW from 400 acres of land, or about 0.62 square miles. So if we wanted something like Nevada Solar One to replace the roughly 400 GW of our non-nuclear electrical energy production, it would require (400 GW / 64 MW ) * 0.62 = 3875 square miles of solar power plants all across the country.
That's a mind boggling amount of land - a square about 62 miles on a side - but since the total amount of US land is 3,539,225 square miles, it represents about 0.1% of our land mass. That's not cheap, but it's not impossible either.
A lot of fatigue from unfamiliar exercise comes because your body isn't used to the motions. I bet if you did the same game for a few days, it wouldn't tire you nearly as much.
A perfectly fit jogger who doesn't jump rope will get tired quickly by jumping rope. Someone who jumps rope frequently but doesn't swim will get tired quickly by swimming. Someone who swims a lot but does not bike will tire quickly while biking. Etc... etc... A lot of the fitness you get is skill adaption to the motions involved and muscle strengthening.
FDA, FCC, court system, environmental protection, FBI, OSHA, Equal Opportunity, HIPAA (medical patient privacy laws), air transit authority - those are all federal. Reproducing each on the state level would be very expensive, and just make life more difficult for businesses, since they would have a broader range of regulations to meet for inter-state commerce.
Regardless, arguing for Federalism - moving agencies and responsibilities from the federal government to the states - is far different from arguing for the total dismantling of the federal government. I was responding to those who seem to imply that shutting the federal government down and not changing state agencies to compensate will leave us with some sort of free market utopia. It would be every bit as bad for most people as having the federal government assume ownership of every company in the country. Six of one, half dozen of the other: power concentrated away from the average person.
So...in Ron Paul's utopia:
Can't afford medical care? Die.
Lots more toll roads, or unusable roads.
Get a bad batch of medicine or food? Hope your family wins a lawsuit against the manufacturer.
You want police or fire protection? Hire your own. If you can't afford it? That's too bad.
Can't afford home heat this season? Freeze.
Did you sign an abusive contract? Well, you're stuck with it.
Someone won't hire or serve you because you're black, hispanic, asian, or lack a penis? Sucks to be you.
Your employer forces you to work overtime, or refuses to pay you? You better have enough money saved to file your own lawsuit.
Your insurance company refuses to pay on a policy claim? You better have enough money saved to file your own lawsuit.
And last but not least, pay no attention to those millions of kids with nothing better to do but get into trouble because the public schools are gone and their parents can't afford private school.
No question, the US government - and most governments - waste money like crazy. But the libertarian solution is as naive as the communist one. Reforming the government by dismantling it just returns us to the days when railroad owners worked the immigrants to death, mine owners used mercenaries to beat workers that tried to strike, men were free to beat their wives, and blacks weren't welcome in certain businesses. No thanks.
Re:High level != "automated memory management"
on
State of the Onion 11
·
· Score: 1
I'd argue that some million line C++ projects can be handled just fine by Perl. Perl supports modules, object orientation, threading, etc...
Now if raw performance is a serious consideration, then obviously Perl won't work.
Re:High level != "automated memory management"
on
State of the Onion 11
·
· Score: 1
I'd say a scripting language has two other advantages that are both incredibly handy.
1. Your script is your code. You don't need to track your source and your compiled binary code separately. You can't lose the source code for a particular version of your software. During development, you just save your code in your editor (Ctrl-S,:w, whatever) and run it. If you forgot which version of a particular app is on a particular machine and what features it might have, you can just open the files directly in view-only mode and read the change log or the code itself.
2. In addition to their own merits, the first class data structures, first class functions, and dynamic typing all contribute to code brevity. If you can express complex ideas in fewer lines of code while keeping the code easier to understand, I think it improves your productivity and also makes it easier for others to maintain your code. I'm sure there are thousands of C, C++, and maybe even Java developers than can accomplish anything I can do in 20 lines of Perl with 15 lines of their own code. But in most cases, it would be damn difficult to read.
I'm certainly not an expert on damage to the human body, but I imagine getting tossed a 2, 3, or 10 yards backwards from a small improvised explosive device is preferable to having most of you remain in one place and having some or all of a limb go sailing away.
You're still toast if a train hits you, but I imagine your survivability from lesser impacts goes way up.
I'm saying that the rural areas get more transit funding per person than the cities already. Some highway in the middle of nowhere Pennsylvania - and I grew up in the middle of nowhere Pennsylvania - gets a lot more money per local resident or per vehicle that travels on that road each year than the big highways in the cities.
Rural area folk like to think of themselves as rugged individuals whose tax money is wasted in the big cities, but the truth is that the government spends more money on infrastructure supporting rural people than it does on the much more population dense cities. You have it backwards.
Thanks for the correction. But I think my general point still stands - you will need security on the trains. The explosion that derails a hijacked plane might kill plenty of people inside, even if the derailment itself does not.
The rural areas already suck all the cash. In Philadelphia you might have 500 inhabitants for every mile of (state and federally funded) road, and in the rural areas you might have 50, and in some cases just 5 inhabitants for every mile of (again, state and federally funded). Fast Eddie Rendell wants to collect tolls on the major roads through rural areas, just to shift a fair portion of state transit funding back into the cities.
The sensible thing to save costs is to allocate the same amount of transit money per citizen to each of the areas - but that would leave rural areas with no new roads and inadequate maintenance of existing roads. Is that what you want?
Thus, maybe the USA will finaly have some user-friendly mass transportation that is
BOTH 1. very fast *AND* 2. doesn't require mandatory body cavities search because of some post 9/11 paranoia.
If a train traveling at 150 mph derails, even if it only hits sand most of the people on board are going to be injured or killed. You still need extensive security - not as much as for a plane, but you can't let a handful of guys with suitcases full of explosives just stroll on in.
What you're saying makes sense, but I think you miss the major items in the article.
Three major points:
1. Easily digestible carbohydrates stoke your appetite. Fats and proteins suppress your appetite. For example, pick a day and eat as many cookies as you want. Pick another day and eat as much cheese or beef as you want. You'll eat less than half as many calories on the meat or cheese day.
2. For some people, strenuous exercise tends to stoke the appetite. It does no good to burn 500 calories on the treadmill or do squats, pulldowns, and shoulder press until your arms and legs can't move if you go home hungry enough to eat two cheese steaks. When I was lifting weights 6 hours a week, a typical lunch was two cheese steaks with a salad and a dessert. I didn't get any fatter - but I didn't get any thinner either. (I also didn't get much stronger - I seem to get much better strength gains from strength training from just two or three 45 minute workouts per week.) Exercise is extremely helpful and important, but if strenuous exertion stokes your appetite, take that into consideration and only exercise moderately. Thin + mild exercise is better than Obese + strenuous exercise.
3. (Most important) Cutting fat in your diet has not been linked to reducing your risk of heart disease. When some guy who lives on McDonalds Super Size meals has a heart attack, we've been taught it's because of the fat, period. But in fact, he's had tons of sugars in his soda, burger buns, and condiments, tons of starch in the fries, trans-saturated (i.e. not naturally occurring) fats used to cook the food, and tons of salt on the fries and in the sandwiches. (Ground beef and chicken have relatively little salt naturally, but fast food burgers and chicken are almost all positively loaded with it.) The beef in his burger probably had nothing to do with his problems.
My bet is that the mother in the previous post might find her appetite reduces further if she put more fats and protein in her diet, and if she did moderate exercise (but not so strenuous as to send her appetite through the roof), it would help even more.
More proteins and fat, and less simple sugars = a lower appetite. That makes it easier to Eat Less And Exercise More.
Really, for most purposes high-end laptops are just as powerful as a desktop.
Sure, but you have to get someone to pay for it. My employer budgets the same fixed amount for employee computers regardless of whether they are workstations or laptops. If my current workstation dies in some spectacularly irretrievable fashion, I can get a respectable new desktop or a low level laptop.
There's also the issue of disk access speeds. On a desktop, you can get super fast 10,000 RPM hard drives pretty cheaply, and if you're really ambitions you can put them in a RAID mirroring and striping array to increase performance even further without unduly risking data loss. You can't do that with a laptop. Since I'm moving around databases and database files a lot, that makes a big difference.
Most people do not use every last byte of a storage device to store data, so the initial argument is extremely plausible, as it applies to the majority of people. The latter argument is possible, but unlikely, especially if there is other evidence against the suspect.
Absolutely. I made an 80 GB encrypted TrueCrypt partition when I got my PC years ago. I mount the drive when I'm surfing for pornography (bad pun unintended), and save anything I like to the drive. It's still much less than half full - and I don't think I can be reasonably expected to have some massive hidden partition.
No it does not violate the Law of Energy, because we don't have to replenish the Sun. The Sun provides the external energy into the system, so you do get a net energy yield because it is not a closed loop.
The problem is that corn yields relatively little fuel per acre after you consider the energy that goes into tilling the land, planting the seeds, watering the field, harvesting the crop, and converting the harvest into fuel. I don't know the exact numbers, but I would imagine that covering the same acre with solar panels or windmills yields more net energy. I'm certain other crops exist that would work better.
There are researchers working on the twin holy grails of renewable fuels: an efficient process that can convert any organic material into fuel, and hardy, high yield strains of algae that can be grown in vats to yield massive amounts of bio-fuels.
I'd say the most valuable mathematical skill to possess is the ability to prove concepts.
An enormous piece of any pursuit, whether it's teaching, software development, marketing, or construction, is finding logical ways to attack new problems and finding ways of approaching existing problems that lead to more efficient solutions.
Even if you never need to prove any piece of geometry, trigonometry, or calculus in your actual job, developing skills at proving will lend itself to learning the skills you actually do need at a much faster pace.
t may appear that way, but IIRC several surveys have shown far and away the opposite -- the quality of work you get after a certain threshold (somewhere in the 40-50 hour range) goes down so rapidly that you're substantially better off throwing more people at the problem than more time.
Thanks for posting that. I agree with it. I think (hope) I write well designed, well documented, well tested, usable code. But honestly, my employer probably gets 5 hours of solid work a day from me and then my concentration starts to rapidly deteriorate.
When I work a marathon session to get something out the door, more than half the time I end up kicking myself because of bugs I missed or crucial design flaws that require rewrites from scratch to add new features. One of the best features in my current job is close work with tech support. Our support staff (1 guy, it's a tiny company) couldn't program to save his life. But I take database changes, end user typical use cases, and sample GUIs to him and we spend at least 20 minutes every day discussing new features and fixes. We figure things out in a ten minute chat that I could spend hours chasing the wrong way on my own.
And if your property taxes on a $200,000 house are $3600 a year (and some places, like New Jersey, can easily hit three times that level of taxes for a $200,000 home) and you're also smart enough to have homeowner's insurance and flood insurance (where floods are a risk), and your monthly payment is around $1600.
Still doable on a $60,000 gross income, but not by as much of a margin as you'd think.
And property taxes are a tough one. Most parents have three choices: pay cheap property taxes in a district and send your kids to mediocre public schools, pay high property taxes and send your kids to decent public schools, or pay cheap property taxes and then spend lots of extra money sending your kids to decent private schools.
How did you research which fields were receiving the students formerly pursuing Computer Science? I'm very curious to see what other areas these people have entered.
Perhaps I'm foolishly optimistic, but from my research there are many thousands, perhaps tens of thousands, of good computer science jobs available inside the US. The problem for Computer Scientists is two-fold. For new graduates, a disproportionately small number of employers are willing to hire people without work experience. Once you have a few years of work under your belt, from what I have seen finding new work is easy. Getting your foot in the door is hard.
The second problem as you mentioned, is staying current with trends and technologies. My friends who teach and who are medical professionals have some token requirements for continuing education, but most can meet the requirements with one or two seminars per year and the occasional class or correspondence course. With software development and IT in general, more work is required.
If you read the parent post more closely, he's speaking of a non "zero sum game" in terms of software creation. There is not a finite demand for good software. New companies open, new business models are discovered, new requirements are added, new devices are created with new software tools for controlling them. More games are created, more tax software is created, more music libraries, more operating systems, more everything.
It's still much cheaper to employ an Indian software developer than an American - but you pay much more for the fellow in India or Chinda today than you did five years ago. As demand keeps increasing eventually they will be no cheaper than an American.
The market definitely does suck in the short term, and I don't mean to downplay the terrible experiences of people whose jobs were outsourced. My previous employer farmed my job out to India, and while I have nothing personal against the people who replaced me, the experience was terrible.
But as India and China's economies grow by leaps and bounds, the demand for their own software developers continues to increase. Eventually they won't be substantively cheaper to employ than we are in the US and Western Europe. Then job opportunities will get better and better for all of us until there's another short term dip because of a new cheap economy opening up - perhaps Africa. And again, the resulting reduction in demand will only last until the new territory's native economy starts catching up with the rest of the world.
I'm fond of the log4j library. Set it up, then litter your code with log statements - logger.error for serious events, logger.debug for reporting nearly every step of the code. First, it documents the code right in the code, and it's much easier to keep the log statements updated when you make code modifications. Second, when something breaks you just change your logging configuration file to turn all logging on for the affected code and re-run your tests. Third, in those (hopefully rare) occasions when a bug appears in production that didn't appear in development, you can turn up the logging level on your production software too.
The 2008 Honda Accord 4-cylinder has 190 horsepower, which is better than a 2006 Toyota Camry V6. By the time my 3 year old is asking for his first car, there'll probably be economy cars that can outrun a 2006 350z.
Unless 75% of the people your PCs encounter use Mind Blank and Non-Detection as a matter of course, the mere fact that they've cast it indicates that they're hiding something. It may be something unrelated to the overarching plot, but it's still something.
Slashdot Mirror
Even if you work independently on the code itself, you still have to collaborate with others on the general design, the features, the interface, and maintaining the code.
Zed has plenty of criticisms of other people and developers. Most of them seem pretty valid on face value.
But I get the impression that if I sat down and offered to pay him for help on some library for our code he would listen for two minutes and then dominate the conversation. Two days later he'd deliver something better than I could write, that I could not understand, which probably did not fit my project, and which I could not maintain. Then he'd take his money and be on his way, patting himself on the back for his brilliance. I'm left with a high powered nail gun with a Japanese owner's manual when I wanted a beer can and a simple 'pull tab to open' instruction.
Unless you're writing the code solely for yourself, your ability to collaborate on some level is crucial.
According to this: http://www.nei.org/newsandevents/1999productionrecords/ the 103 commercial nuclear reactors in the US provide 20% of the nation's electrical energy, and wikipedia asserts that a typical nuke plant produces 1 GW of power. So presumably about 100 GW of electrical generation is 20% of our electrical needs.
Also from wikipedia, the Nevada Solar One project makes 64 MW from 400 acres of land, or about 0.62 square miles. So if we wanted something like Nevada Solar One to replace the roughly 400 GW of our non-nuclear electrical energy production, it would require (400 GW / 64 MW ) * 0.62 = 3875 square miles of solar power plants all across the country.
That's a mind boggling amount of land - a square about 62 miles on a side - but since the total amount of US land is 3,539,225 square miles, it represents about 0.1% of our land mass. That's not cheap, but it's not impossible either.
A lot of fatigue from unfamiliar exercise comes because your body isn't used to the motions. I bet if you did the same game for a few days, it wouldn't tire you nearly as much.
A perfectly fit jogger who doesn't jump rope will get tired quickly by jumping rope. Someone who jumps rope frequently but doesn't swim will get tired quickly by swimming. Someone who swims a lot but does not bike will tire quickly while biking. Etc... etc... A lot of the fitness you get is skill adaption to the motions involved and muscle strengthening.
FDA, FCC, court system, environmental protection, FBI, OSHA, Equal Opportunity, HIPAA (medical patient privacy laws), air transit authority - those are all federal. Reproducing each on the state level would be very expensive, and just make life more difficult for businesses, since they would have a broader range of regulations to meet for inter-state commerce.
Regardless, arguing for Federalism - moving agencies and responsibilities from the federal government to the states - is far different from arguing for the total dismantling of the federal government. I was responding to those who seem to imply that shutting the federal government down and not changing state agencies to compensate will leave us with some sort of free market utopia. It would be every bit as bad for most people as having the federal government assume ownership of every company in the country. Six of one, half dozen of the other: power concentrated away from the average person.
So...in Ron Paul's utopia:
Can't afford medical care? Die.
Lots more toll roads, or unusable roads.
Get a bad batch of medicine or food? Hope your family wins a lawsuit against the manufacturer.
You want police or fire protection? Hire your own. If you can't afford it? That's too bad.
Can't afford home heat this season? Freeze.
Did you sign an abusive contract? Well, you're stuck with it.
Someone won't hire or serve you because you're black, hispanic, asian, or lack a penis? Sucks to be you.
Your employer forces you to work overtime, or refuses to pay you? You better have enough money saved to file your own lawsuit.
Your insurance company refuses to pay on a policy claim? You better have enough money saved to file your own lawsuit.
And last but not least, pay no attention to those millions of kids with nothing better to do but get into trouble because the public schools are gone and their parents can't afford private school.
No question, the US government - and most governments - waste money like crazy. But the libertarian solution is as naive as the communist one. Reforming the government by dismantling it just returns us to the days when railroad owners worked the immigrants to death, mine owners used mercenaries to beat workers that tried to strike, men were free to beat their wives, and blacks weren't welcome in certain businesses. No thanks.
I'd argue that some million line C++ projects can be handled just fine by Perl. Perl supports modules, object orientation, threading, etc...
Now if raw performance is a serious consideration, then obviously Perl won't work.
I'd say a scripting language has two other advantages that are both incredibly handy.
:w, whatever) and run it. If you forgot which version of a particular app is on a particular machine and what features it might have, you can just open the files directly in view-only mode and read the change log or the code itself.
1. Your script is your code. You don't need to track your source and your compiled binary code separately. You can't lose the source code for a particular version of your software. During development, you just save your code in your editor (Ctrl-S,
2. In addition to their own merits, the first class data structures, first class functions, and dynamic typing all contribute to code brevity. If you can express complex ideas in fewer lines of code while keeping the code easier to understand, I think it improves your productivity and also makes it easier for others to maintain your code. I'm sure there are thousands of C, C++, and maybe even Java developers than can accomplish anything I can do in 20 lines of Perl with 15 lines of their own code. But in most cases, it would be damn difficult to read.
I'm certainly not an expert on damage to the human body, but I imagine getting tossed a 2, 3, or 10 yards backwards from a small improvised explosive device is preferable to having most of you remain in one place and having some or all of a limb go sailing away.
You're still toast if a train hits you, but I imagine your survivability from lesser impacts goes way up.
I'm saying that the rural areas get more transit funding per person than the cities already. Some highway in the middle of nowhere Pennsylvania - and I grew up in the middle of nowhere Pennsylvania - gets a lot more money per local resident or per vehicle that travels on that road each year than the big highways in the cities.
Rural area folk like to think of themselves as rugged individuals whose tax money is wasted in the big cities, but the truth is that the government spends more money on infrastructure supporting rural people than it does on the much more population dense cities. You have it backwards.
Thanks for the correction. But I think my general point still stands - you will need security on the trains. The explosion that derails a hijacked plane might kill plenty of people inside, even if the derailment itself does not.
The rural areas already suck all the cash. In Philadelphia you might have 500 inhabitants for every mile of (state and federally funded) road, and in the rural areas you might have 50, and in some cases just 5 inhabitants for every mile of (again, state and federally funded). Fast Eddie Rendell wants to collect tolls on the major roads through rural areas, just to shift a fair portion of state transit funding back into the cities.
The sensible thing to save costs is to allocate the same amount of transit money per citizen to each of the areas - but that would leave rural areas with no new roads and inadequate maintenance of existing roads. Is that what you want?
Thus, maybe the USA will finaly have some user-friendly mass transportation that is BOTH 1. very fast *AND* 2. doesn't require mandatory body cavities search because of some post 9/11 paranoia.
If a train traveling at 150 mph derails, even if it only hits sand most of the people on board are going to be injured or killed. You still need extensive security - not as much as for a plane, but you can't let a handful of guys with suitcases full of explosives just stroll on in.
What you're saying makes sense, but I think you miss the major items in the article. Three major points:
1. Easily digestible carbohydrates stoke your appetite. Fats and proteins suppress your appetite. For example, pick a day and eat as many cookies as you want. Pick another day and eat as much cheese or beef as you want. You'll eat less than half as many calories on the meat or cheese day.
2. For some people, strenuous exercise tends to stoke the appetite. It does no good to burn 500 calories on the treadmill or do squats, pulldowns, and shoulder press until your arms and legs can't move if you go home hungry enough to eat two cheese steaks. When I was lifting weights 6 hours a week, a typical lunch was two cheese steaks with a salad and a dessert. I didn't get any fatter - but I didn't get any thinner either. (I also didn't get much stronger - I seem to get much better strength gains from strength training from just two or three 45 minute workouts per week.) Exercise is extremely helpful and important, but if strenuous exertion stokes your appetite, take that into consideration and only exercise moderately. Thin + mild exercise is better than Obese + strenuous exercise.
3. (Most important) Cutting fat in your diet has not been linked to reducing your risk of heart disease. When some guy who lives on McDonalds Super Size meals has a heart attack, we've been taught it's because of the fat, period. But in fact, he's had tons of sugars in his soda, burger buns, and condiments, tons of starch in the fries, trans-saturated (i.e. not naturally occurring) fats used to cook the food, and tons of salt on the fries and in the sandwiches. (Ground beef and chicken have relatively little salt naturally, but fast food burgers and chicken are almost all positively loaded with it.) The beef in his burger probably had nothing to do with his problems.
My bet is that the mother in the previous post might find her appetite reduces further if she put more fats and protein in her diet, and if she did moderate exercise (but not so strenuous as to send her appetite through the roof), it would help even more.
More proteins and fat, and less simple sugars = a lower appetite. That makes it easier to Eat Less And Exercise More.
Really, for most purposes high-end laptops are just as powerful as a desktop.
Sure, but you have to get someone to pay for it. My employer budgets the same fixed amount for employee computers regardless of whether they are workstations or laptops. If my current workstation dies in some spectacularly irretrievable fashion, I can get a respectable new desktop or a low level laptop.
There's also the issue of disk access speeds. On a desktop, you can get super fast 10,000 RPM hard drives pretty cheaply, and if you're really ambitions you can put them in a RAID mirroring and striping array to increase performance even further without unduly risking data loss. You can't do that with a laptop. Since I'm moving around databases and database files a lot, that makes a big difference.
Most people do not use every last byte of a storage device to store data, so the initial argument is extremely plausible, as it applies to the majority of people. The latter argument is possible, but unlikely, especially if there is other evidence against the suspect.
Absolutely. I made an 80 GB encrypted TrueCrypt partition when I got my PC years ago. I mount the drive when I'm surfing for pornography (bad pun unintended), and save anything I like to the drive. It's still much less than half full - and I don't think I can be reasonably expected to have some massive hidden partition.
No it does not violate the Law of Energy, because we don't have to replenish the Sun. The Sun provides the external energy into the system, so you do get a net energy yield because it is not a closed loop.
The problem is that corn yields relatively little fuel per acre after you consider the energy that goes into tilling the land, planting the seeds, watering the field, harvesting the crop, and converting the harvest into fuel. I don't know the exact numbers, but I would imagine that covering the same acre with solar panels or windmills yields more net energy. I'm certain other crops exist that would work better.
There are researchers working on the twin holy grails of renewable fuels: an efficient process that can convert any organic material into fuel, and hardy, high yield strains of algae that can be grown in vats to yield massive amounts of bio-fuels.
I'd say the most valuable mathematical skill to possess is the ability to prove concepts.
An enormous piece of any pursuit, whether it's teaching, software development, marketing, or construction, is finding logical ways to attack new problems and finding ways of approaching existing problems that lead to more efficient solutions.
Even if you never need to prove any piece of geometry, trigonometry, or calculus in your actual job, developing skills at proving will lend itself to learning the skills you actually do need at a much faster pace.
There was a news article about this earlier this year, which Slashdot discussed: http://www.sfgate.com/cgi-bin/article.cgi?f=/n/a/2007/07/26/national/w111044D35.DTL
t may appear that way, but IIRC several surveys have shown far and away the opposite -- the quality of work you get after a certain threshold (somewhere in the 40-50 hour range) goes down so rapidly that you're substantially better off throwing more people at the problem than more time.
Thanks for posting that. I agree with it. I think (hope) I write well designed, well documented, well tested, usable code. But honestly, my employer probably gets 5 hours of solid work a day from me and then my concentration starts to rapidly deteriorate.
When I work a marathon session to get something out the door, more than half the time I end up kicking myself because of bugs I missed or crucial design flaws that require rewrites from scratch to add new features. One of the best features in my current job is close work with tech support. Our support staff (1 guy, it's a tiny company) couldn't program to save his life. But I take database changes, end user typical use cases, and sample GUIs to him and we spend at least 20 minutes every day discussing new features and fixes. We figure things out in a ten minute chat that I could spend hours chasing the wrong way on my own.
And if your property taxes on a $200,000 house are $3600 a year (and some places, like New Jersey, can easily hit three times that level of taxes for a $200,000 home) and you're also smart enough to have homeowner's insurance and flood insurance (where floods are a risk), and your monthly payment is around $1600.
Still doable on a $60,000 gross income, but not by as much of a margin as you'd think.
And property taxes are a tough one. Most parents have three choices: pay cheap property taxes in a district and send your kids to mediocre public schools, pay high property taxes and send your kids to decent public schools, or pay cheap property taxes and then spend lots of extra money sending your kids to decent private schools.
I sure hope you're right. But we've been looking for big breakthroughs in battery technology for decades without luck.
How did you research which fields were receiving the students formerly pursuing Computer Science? I'm very curious to see what other areas these people have entered.
Perhaps I'm foolishly optimistic, but from my research there are many thousands, perhaps tens of thousands, of good computer science jobs available inside the US. The problem for Computer Scientists is two-fold. For new graduates, a disproportionately small number of employers are willing to hire people without work experience. Once you have a few years of work under your belt, from what I have seen finding new work is easy. Getting your foot in the door is hard.
The second problem as you mentioned, is staying current with trends and technologies. My friends who teach and who are medical professionals have some token requirements for continuing education, but most can meet the requirements with one or two seminars per year and the occasional class or correspondence course. With software development and IT in general, more work is required.
If you read the parent post more closely, he's speaking of a non "zero sum game" in terms of software creation. There is not a finite demand for good software. New companies open, new business models are discovered, new requirements are added, new devices are created with new software tools for controlling them. More games are created, more tax software is created, more music libraries, more operating systems, more everything.
It's still much cheaper to employ an Indian software developer than an American - but you pay much more for the fellow in India or Chinda today than you did five years ago. As demand keeps increasing eventually they will be no cheaper than an American.
The market definitely does suck in the short term, and I don't mean to downplay the terrible experiences of people whose jobs were outsourced. My previous employer farmed my job out to India, and while I have nothing personal against the people who replaced me, the experience was terrible.
But as India and China's economies grow by leaps and bounds, the demand for their own software developers continues to increase. Eventually they won't be substantively cheaper to employ than we are in the US and Western Europe. Then job opportunities will get better and better for all of us until there's another short term dip because of a new cheap economy opening up - perhaps Africa. And again, the resulting reduction in demand will only last until the new territory's native economy starts catching up with the rest of the world.
I'm fond of the log4j library. Set it up, then litter your code with log statements - logger.error for serious events, logger.debug for reporting nearly every step of the code. First, it documents the code right in the code, and it's much easier to keep the log statements updated when you make code modifications. Second, when something breaks you just change your logging configuration file to turn all logging on for the affected code and re-run your tests. Third, in those (hopefully rare) occasions when a bug appears in production that didn't appear in development, you can turn up the logging level on your production software too.
Of course, unit testing is also good.
The 2008 Honda Accord 4-cylinder has 190 horsepower, which is better than a 2006 Toyota Camry V6. By the time my 3 year old is asking for his first car, there'll probably be economy cars that can outrun a 2006 350z.
Unless 75% of the people your PCs encounter use Mind Blank and Non-Detection as a matter of course, the mere fact that they've cast it indicates that they're hiding something. It may be something unrelated to the overarching plot, but it's still something.