Exactly right. I've got >10K cores and >10M LOC. "Hardware fault" typically means a datacenter caught fire, or was flooded, or an undersea cable got cut.
If someone pitches a cheaper solution (e.g. power savings,) I'm happy to listen for 10 minutes. Then I just want to know how fast I can see results: a dev costs $50K/month here, so I'll give it a week or two: if you don't have a test farm ready to go with full compilers, a data security plan, etc, I'm going to just reject. If you can get traction with universities, great, come back and pitch again in a year.
I've found CS degrees to be mostly uncorrelated with programming skill. Out of our best programmers, one has a CS masters degree, the others are a mix of math, philosophy, no-degree, etc. The common trait is that all of them know the field back-to-front: after four hours of drinking at the local bar, the topic will still be monads, graph theory, quantum computing, or something equally geeky.
Nope, that's not what the summary is saying. It's stating that we have a steady stream of scientists and engineers, but it seems like they choose another career path when they realize they'll just be overworked and underpaid.
Which is another way of saying their firm realized they are a useless waste of money incapable of doing the most basic jobs that their degrees implied they were qualified for. What a surprise that they eventually seek another career path.
If programmers were lawyers, 90% would be disbarred. Half the statisticians I interview can't solve a basic statistics problem correctly. 2/3 of the aeronautical engineers couldn't solve for lift on a wing. Half the math guys needed hints to prove that if x-1 and x+1 are prime, x is divisible by 6.
Fuck em. A McDonalds worker is welcome to a sociology degree, but giving him a math degree doesn't make him a mathematician.
We need good people in the math/sci fields, not dumbasses who got a degree because it was "going to make them rich."
Ya know... I find the normal developers to be the weirdest.
Exactly right. All developers are a bit quirky, but the seriously weird always seem to be the wannabes in the middle of the pack.
The best developers I know are odd: they ignore a lot of life-stuff, but concentrate on making really good technical and biz decisions. They seem strange, but have no trouble finding hot girlfriends and good jobs. The second tier is a bit of a cargo-cult: they imitate the strange aspect, then get confused when the chicks and money don't arrive.
Pretty much right. Although there doesn't really need to be an ordering in time: dump money in infrastructure, public beautification projects, scientific research, social services, education.
It all is net positive while inflation remains low. Some people (mostly the rich) will always complain that "their" tax dollars are being wasted, but the real waste is that skilled workers deskill (e.g. accountants cooking their own meals, or plumbers darning their socks.) The economy is not a zero sum game.
How many people do you employ directly? What is your position? And do you have a family yet?
a. 0. Luckily, my boss has 15 or so direct reports, so I just call up a peer and ask if I need something done (either how I should do it, or ask them to do it.) b. zzz Director. Enough that I can get a plane ticket to where I want to go without needing approval, can cut red tape, etc. Nothing more. c. Wife, two kids, full time nanny, and part-time help.
Unfortunately you do not usually find out if someone codes in their free time until after you employ them so is not doing this a sackable offence?
Once sometime is hired, productivity is all I really care about. I don't care if they write video games at work, browse the web endlessly, etc. All I want is that they do stuff they said they were going to do. Ideally, on-time, elegantly, and professionally.
Hmm, I code in my free time. My boss codes in his free time. So does his boss. We also read books about random stuff and discuss our biz, etc. The end result is that we trust each other to make intelligent choices and decisions at work: you can just do a $1MM project with a few minutes of explanation. Try doing that where the management chain doesn't really understand code.
I wouldn't hire, or work for, a person who treats programming as a 9 to 5 activity. Life is short, and the craft so long to learn.
The real message is a bit more subtle. Yes, you have to follow your passions, but you also have to become really good at what you are passionate about. That takes years, or decades, of enjoyable, but very hard work.
Oh, and when you get to that top.01%, the bottom 90% are still going to whine about how you got a few lucky breaks and don't deserve the money you are now being paid.
Seek out a few people who are, or were, at the top of their field (olympic athletes, trial lawyers, coders, etc) and you'll find they had talent (probably in the top 1%,) but you'll also find they practiced ruthlessly on getting better. Day after day, for hours on end, regardless of their current professional short-term situation.
Note that Benford's law does not apply to specific standard distributions (e.g. numbers from a gaussian m=6, sd=1 will not obey the law.) That applies to all digits, btw.
I'll concede the numerical distributions where base 10 is important, e.g. your $10K tax cutoff, are not going to obey Benford's law. Given that, however, you can't just throw out the first digit and claim Benford's law applies to the rest of digits.
Benford's law (and I'm probably going to piss off a lot of IRS forensic accountants here) works in any base (e.g. if I convert your tax return numbers into base 7, I expect a certain first-digit distribution.) If your numbers stick out over a Benford test of bases 3,5,7,10,11, and 13, I am going to think you are cheating and look very closely.
Nate Silver does great analysis at the first order multiple-linear-regression level -- he outperformed all the other polls/predictors in 2008 iirc.
He sucks at meta-analysis though, in that he just doesn't understand the math. His 2008 monte-carlo stuff gave good results, but was just a bad reinvention of averaging. His recent foray into analyzing stock returns was interesting but 0-information (i.e. useless.)
Now he's mentioning Benford's law, but playing with trailing digits. Then he handwaves a non-normal result with an appeal to "it looks wrong." Come on, give us some real math here!
That said, he's probably right, but he's given us no math to support his claim.
Sadly, yes. Peter Norvig felt 16/23 were "invisible or simpler" in a higher-level language (Lisp in his case.)
A big system is going to have a couple more layers of complexity (each with its own 'design patterns'. ) A functional language is a required element at this point.
I still we're still writing the guide. My approach (for a massive Python installation,) has evolved towards a series of unittests:
Group 1. These test the normal behavior of the module (add an order, cancel an order, modify on order, etc.) Should be easy enough for another developer to read and get a sense of how the module is used. Gets written concurrently with the module itself. Group 2. Edge, error, and pathological cases. E.g. add an order with no items, race condition on simultaneous modifies of an order. Again, should be understandable by other developers. Group 3. Tests that catch every real bug or problem that was reported after initial release. I don't change the code until I've added a test here that fails under the old code, then I fix the code, then the test passes.
1 & 2 form part of the documentation. 3 is the historical record of the module.
Actually, there was a lot of government intervention in the tulip bubble. The govt basically caused the bubble by restricting investments in most other products. Then, when things started getting out of control, they trashed the market with yet more intervention.
Oh, for those of you thinking that investing in tulips was objectively insane, consider this: Holland still controls the worldwide flower market because they built the infrastructure during that "bubble." A long-term investor with a diversified portfolio of firms in the flower industry in that period would have outperformed the stock market over the long term.
The pictures show a couple of strands of hair across copper terminals. And that gets you 18W? Sunlight at around 1KW/sq m - sorry, the math doesn't add up.
Oh, and with your body dumping around 100W of waste heat, a free power source of even 10W would have animals evolved into electrically powered beasts in no time.
We know obesity is inversely correlated with income, and income is positively correlated with measured intelligence. Ergo, obesity is correlated positively with stupid.
And, no, I'm NOT saying it's a perfectly linear relationship, and, NO, one does not cause the other. There are many explanations for the relationship, e.g. conservatives are fat, thus they have to adopt stupid points of view to get laid, so they can't use their brains much, so they lose brain mass (Glenn Beck and Rush are merely extreme examples.)
You're quite right, of course. The goal is big fanout + channelling to get the mob density down, plus gentle downhills to nudge the people in the right directions. A fire exit killer is often not the pure size of the exit, it is that that the first 100 people out form a crowd outside and slow the egress of the next 400 people. You can deal with the risks of people running downhill as long as the running reduces bottlenecks and trample risk - it's the difference between panic inside a burning nightclub and people fleeing the scene of a massacre.
Oh, and design the exit assembly areas so as to encourage dispersion from the final choke point at the exit: ideally, have the exit open to an amphitheater like shape so people will walk/run downhill/in various directions. Add attractors to get them away from the choke point fast: like, big sign advertising free beer 100 yards off to the side (seriously.)
The weird thing is that people who actually design stuff for crowd control have known this since at the least the 1980s. The goal is to get people ordered into efficient lines heading towards the goals and make sure people understand the process is fair and nothing is to be gained by jumping lines. For a real world example, see Heathrow's newer terminals versus its older ones, or any third world airport: if you make it easy to cheat by changing lines, and other people can see you do it, you get a mob in short order. So, keep lines narrow, and hard to switch from one to another, and people move faster. That means barriers - big ones. Just think Disneyworld, airports, good stadia.
So the program is ideal. Screw $1,500 of kayaking - think $15,000 of "exotic roleplaying." Man, this spa thing could be a winner - 45 days at a hotel-like spa with outside activities? Set it up right and Blue Cross will even pay for it.
I am so there! Sorry, boss, I need 45 days off to cure my internet addiction - I'll be kayaking, deep-sea fishing, and TFing hookers. No prob, I hate to do it, but it's the only way to reconnect with reality. Oh, and you are paying, don't make me invoke the "Americans with disabilities" law.
This is pretty much settled law. She published public information. So what?
That it was about a public official makes her actions more reasonable, not less. She didn't incite anyone to go kill the guy.
Compare that to, say, people publishing abortion doctors' info on "wanted dead or alive" type sites, and it seems reasonably clear she is covered by the 1st amendment.
Slashdot Mirror
Exactly right. I've got >10K cores and >10M LOC. "Hardware fault" typically means a datacenter caught fire, or was flooded, or an undersea cable got cut.
If someone pitches a cheaper solution (e.g. power savings,) I'm happy to listen for 10 minutes. Then I just want to know how fast I can see results: a dev costs $50K/month here, so I'll give it a week or two: if you don't have a test farm ready to go with full compilers, a data security plan, etc, I'm going to just reject. If you can get traction with universities, great, come back and pitch again in a year.
I've found CS degrees to be mostly uncorrelated with programming skill. Out of our best programmers, one has a CS masters degree, the others are a mix of math, philosophy, no-degree, etc. The common trait is that all of them know the field back-to-front: after four hours of drinking at the local bar, the topic will still be monads, graph theory, quantum computing, or something equally geeky.
Quite right. That's the "extra credit" part of the question :)
Nope, that's not what the summary is saying. It's stating that we have a steady stream of scientists and engineers, but it seems like they choose another career path when they realize they'll just be overworked and underpaid.
Which is another way of saying their firm realized they are a useless waste of money incapable of doing the most basic jobs that their degrees implied they were qualified for. What a surprise that they eventually seek another career path.
If programmers were lawyers, 90% would be disbarred. Half the statisticians I interview can't solve a basic statistics problem correctly. 2/3 of the aeronautical engineers couldn't solve for lift on a wing. Half the math guys needed hints to prove that if x-1 and x+1 are prime, x is divisible by 6.
Fuck em. A McDonalds worker is welcome to a sociology degree, but giving him a math degree doesn't make him a mathematician.
We need good people in the math/sci fields, not dumbasses who got a degree because it was "going to make them rich."
Exactly right. All developers are a bit quirky, but the seriously weird always seem to be the wannabes in the middle of the pack.
The best developers I know are odd: they ignore a lot of life-stuff, but concentrate on making really good technical and biz decisions. They seem strange, but have no trouble finding hot girlfriends and good jobs. The second tier is a bit of a cargo-cult: they imitate the strange aspect, then get confused when the chicks and money don't arrive.
Pretty much right. Although there doesn't really need to be an ordering in time: dump money in infrastructure, public beautification projects, scientific research, social services, education.
It all is net positive while inflation remains low. Some people (mostly the rich) will always complain that "their" tax dollars are being wasted, but the real waste is that skilled workers deskill (e.g. accountants cooking their own meals, or plumbers darning their socks.) The economy is not a zero sum game.
Actually, every programmer in my group was paid six figures. The hardest thing is just explaining that there is no upper bound :)
a. 0. Luckily, my boss has 15 or so direct reports, so I just call up a peer and ask if I need something done (either how I should do it, or ask them to do it.)
b. zzz Director. Enough that I can get a plane ticket to where I want to go without needing approval, can cut red tape, etc. Nothing more.
c. Wife, two kids, full time nanny, and part-time help.
Once sometime is hired, productivity is all I really care about. I don't care if they write video games at work, browse the web endlessly, etc. All I want is that they do stuff they said they were going to do. Ideally, on-time, elegantly, and professionally.
Hmm, I code in my free time. My boss codes in his free time. So does his boss. We also read books about random stuff and discuss our biz, etc. The end result is that we trust each other to make intelligent choices and decisions at work: you can just do a $1MM project with a few minutes of explanation. Try doing that where the management chain doesn't really understand code.
I wouldn't hire, or work for, a person who treats programming as a 9 to 5 activity. Life is short, and the craft so long to learn.
The real message is a bit more subtle. Yes, you have to follow your passions, but you also have to become really good at what you are passionate about. That takes years, or decades, of enjoyable, but very hard work.
Oh, and when you get to that top .01%, the bottom 90% are still going to whine about how you got a few lucky breaks and don't deserve the money you are now being paid.
Seek out a few people who are, or were, at the top of their field (olympic athletes, trial lawyers, coders, etc) and you'll find they had talent (probably in the top 1%,) but you'll also find they practiced ruthlessly on getting better. Day after day, for hours on end, regardless of their current professional short-term situation.
Note that Benford's law does not apply to specific standard distributions (e.g. numbers from a gaussian m=6, sd=1 will not obey the law.) That applies to all digits, btw.
Um, basically no.
I'll concede the numerical distributions where base 10 is important, e.g. your $10K tax cutoff, are not going to obey Benford's law. Given that, however, you can't just throw out the first digit and claim Benford's law applies to the rest of digits.
Benford's law (and I'm probably going to piss off a lot of IRS forensic accountants here) works in any base (e.g. if I convert your tax return numbers into base 7, I expect a certain first-digit distribution.) If your numbers stick out over a Benford test of bases 3,5,7,10,11, and 13, I am going to think you are cheating and look very closely.
Nate Silver does great analysis at the first order multiple-linear-regression level -- he outperformed all the other polls/predictors in 2008 iirc.
He sucks at meta-analysis though, in that he just doesn't understand the math. His 2008 monte-carlo stuff gave good results, but was just a bad reinvention of averaging. His recent foray into analyzing stock returns was interesting but 0-information (i.e. useless.)
Now he's mentioning Benford's law, but playing with trailing digits. Then he handwaves a non-normal result with an appeal to "it looks wrong." Come on, give us some real math here!
That said, he's probably right, but he's given us no math to support his claim.
Sadly, yes. Peter Norvig felt 16/23 were "invisible or simpler" in a higher-level language (Lisp in his case.)
A big system is going to have a couple more layers of complexity (each with its own 'design patterns'. ) A functional language is a required element at this point.
I still we're still writing the guide. My approach (for a massive Python installation,) has evolved towards a series of unittests:
Group 1. These test the normal behavior of the module (add an order, cancel an order, modify on order, etc.) Should be easy enough for another developer to read and get a sense of how the module is used. Gets written concurrently with the module itself.
Group 2. Edge, error, and pathological cases. E.g. add an order with no items, race condition on simultaneous modifies of an order. Again, should be understandable by other developers.
Group 3. Tests that catch every real bug or problem that was reported after initial release. I don't change the code until I've added a test here that fails under the old code, then I fix the code, then the test passes.
1 & 2 form part of the documentation. 3 is the historical record of the module.
Actually, there was a lot of government intervention in the tulip bubble. The govt basically caused the bubble by restricting investments in most other products. Then, when things started getting out of control, they trashed the market with yet more intervention.
Oh, for those of you thinking that investing in tulips was objectively insane, consider this: Holland still controls the worldwide flower market because they built the infrastructure during that "bubble." A long-term investor with a diversified portfolio of firms in the flower industry in that period would have outperformed the stock market over the long term.
The article is clearly bullshit.
The pictures show a couple of strands of hair across copper terminals. And that gets you 18W? Sunlight at around 1KW/sq m - sorry, the math doesn't add up.
Oh, and with your body dumping around 100W of waste heat, a free power source of even 10W would have animals evolved into electrically powered beasts in no time.
Why go so complex?
We know obesity is inversely correlated with income, and income is positively correlated with measured intelligence. Ergo, obesity is correlated positively with stupid.
And, no, I'm NOT saying it's a perfectly linear relationship, and, NO, one does not cause the other. There are many explanations for the relationship, e.g. conservatives are fat, thus they have to adopt stupid points of view to get laid, so they can't use their brains much, so they lose brain mass (Glenn Beck and Rush are merely extreme examples.)
You're quite right, of course. The goal is big fanout + channelling to get the mob density down, plus gentle downhills to nudge the people in the right directions. A fire exit killer is often not the pure size of the exit, it is that that the first 100 people out form a crowd outside and slow the egress of the next 400 people. You can deal with the risks of people running downhill as long as the running reduces bottlenecks and trample risk - it's the difference between panic inside a burning nightclub and people fleeing the scene of a massacre.
Oh, and design the exit assembly areas so as to encourage dispersion from the final choke point at the exit: ideally, have the exit open to an amphitheater like shape so people will walk/run downhill/in various directions. Add attractors to get them away from the choke point fast: like, big sign advertising free beer 100 yards off to the side (seriously.)
The weird thing is that people who actually design stuff for crowd control have known this since at the least the 1980s. The goal is to get people ordered into efficient lines heading towards the goals and make sure people understand the process is fair and nothing is to be gained by jumping lines. For a real world example, see Heathrow's newer terminals versus its older ones, or any third world airport: if you make it easy to cheat by changing lines, and other people can see you do it, you get a mob in short order. So, keep lines narrow, and hard to switch from one to another, and people move faster. That means barriers - big ones. Just think Disneyworld, airports, good stadia.
Oops, my boss going be really pissed about my "drying out" trip to Little Palm Island :(
So the program is ideal. Screw $1,500 of kayaking - think $15,000 of "exotic roleplaying." Man, this spa thing could be a winner - 45 days at a hotel-like spa with outside activities? Set it up right and Blue Cross will even pay for it.
I am so there! Sorry, boss, I need 45 days off to cure my internet addiction - I'll be kayaking, deep-sea fishing, and TFing hookers. No prob, I hate to do it, but it's the only way to reconnect with reality. Oh, and you are paying, don't make me invoke the "Americans with disabilities" law.
This is pretty much settled law. She published public information. So what?
That it was about a public official makes her actions more reasonable, not less. She didn't incite anyone to go kill the guy.
Compare that to, say, people publishing abortion doctors' info on "wanted dead or alive" type sites, and it seems reasonably clear she is covered by the 1st amendment.
Hey, I think we are arguing the same point here :)
(my points should be anded together, not ored)