When he says deep TCP/IP knowledge he probably means he needs a systems programmer. TCP/IP is such a common standard that anyone with the required background will already be familiar with it, so the training will not begin there.
If it were an entry level programming position teaching someone to program might be worthwhile, but there are also plenty of programmers with a security clearance. I imagine part of the problem he is having is that the security clearance does not mean as much to someone who has this skill set. The major draw to having one is the job security and higher salary, but if you can sit down and write a network stack or code to analyze packets in interesting ways in real time you already have an easy time finding employment, and the ability to make real money if you desire.
You seem to be stuck in the distant past with your opinion of C.
Yes, it entirely sucked when people tried to use it for general applications as a default language (same with C++, maybe even more so there). It is not at all an appropriate language when you are trying to deal with a ton of general business code implemented by a large team of varying quality (Java and C# do very well here). It is not really appropriate for someone who wants to learn to code as quickly as possible, nor does it make a particularly good prototyping language.
I think a few people already pointed you to all of the public linux stuff. MS also writes the kernel of Windows in C, all device drivers... also I did a bit of programming for industrial hardware, all of it was C. There are a few other areas where it ends up being very useful or nearly required, such as dealing with applications using wildly different hardware, a task which is at the limits of what a modern computer can do, or if more speed is always good (the last time I wrote a significant amount of assembly was actually a financial project, you would be surprised at where you end up needing the bag of tricks you can use with C and assembly to greatly speed up a process).
You hire expensive senior programmers by reference generally, not by posting on monster.com or whatever (unless you are really out of options).
C++ is not common in life critical software, maybe ADA (although I have never seen it, I also do not do government work). You will be applying KISS if you are designing that kind of system, in which case a large and complicated piece created externally is right out, so no C++ compiler. The fact that C has such a simple compiler is an important consideration.
It has never been lower than position 2. It has been first or second on that list since they started tracking it, as you can see if you scroll down a little bit for the chart over time (which goes back 15 years on that page).
If all of your jobs involve javascript, that likely makes you a web developer. Using some web frameworks to put together websites is about as far away as you can get from systems programming, so of course nobody is asking you to write a bunch of C code. It is not your profession, and the results are likely to be horrific.
The vast majority of important code is written in C, including new projects today. It dominates few industries, but is present in almost all of them (because when you run into a problem which needs C you have few other options, none of them easier). C# is pretty popular for general business programming, matlab is popular for scientific stuff, I still see a fair amount of cobol for big financial systems, php and like half a dozen others are popular for web development, etc. The thing is they have pretty narrow niches, so while they may be dominant in one industry, outside of that you basically do not see it.
I never see anyone write javascript for a big financial program (except maybe for the front end), write data analysis code for an experiment in C#, or write their website backend in matlab. I do however see financial programmers who think the world is entirely Java, C# and SAP, academics who think matlab can solve anything, and web developers who think PHP and Javascript are all you will ever need.
When your problem requires new hardware of any sort, you care about speed, or there is the possibility that someone may be harmed if the program does not work correctly, the language of choice is usually C.
I will also throw out there that I do a lot of code review, and the C programs I am sent have by far the lowest defect rate of what I review (PHP being the worst.)
The long term website tracking popularity (as well as you can do that for programming languages) is at http://www.tiobe.com/tiobe_index
C is pretty much always 1 or 2 in the list, and notice how it has twice the ranking of the third in the list (C++... which owes much of its popularity to the ability to compile most C programs). C is a very popular language, and is pretty much the language of choice whenever you need to do something novel, involving hardware, or care about execution time. It is not usually the best choice for a big CMS or other tasks where there are domain specific languages available, but that is mostly due to staffing concerns.
It tends not to have a ton of people in any given industry using it as it requires a greater understanding as to how the computer operates, which means you do not train decent C programmers in a few months like you do PHP or such (and you do not throw it at academics like MATLAB). It takes years to get good, and even then it turns out a large number of people will struggle with memory management no matter how hard they try.
If it takes you 3 or 4 times as long to write something in C, that is a good indication you should not be writing C. In contrast it is my favorite language and I am far more productive in it than Java, PHP, etc. (not that I use it all the time, I am well aware that if I write something in C I will likely need to continue maintaining it, so I usually need a reason to do so).
Unfortunately while it includes the fundamentally unsafe ones like gets(), it also deprecates a bunch of functions which are not fundamentally unsafe, such as the simple and super common strcpy (which is just fine if you know there is a null byte within the buffer length, generally because you as the programmer put it there).
The solution they use is also pretty bad, adding a buffer length parameter to the function is useless in practice as it still lets you overflow the buffer by getting this wrong, and it seems most code which would contain a buffer overflow ends up calculating this in a way which ends up passing the wrong buffer length. When combined with how common some of those functions are in existing code (where usage is perfectly safe!), I end up just turning it off by default in my compiler. It is an annoyance for no real gain.
I do not consider those useful warnings as MS implements them. It's like someone with minimal understanding as to why you would want to control your own memory layout who never works on code for an OS other than Windows was asked to fix buffer overflows in C, and came up with the great idea that if you had the programmer specify the length twice they might not screw it up in exactly the same way twice (which they do, in practice it is rare to see the size of the buffer provided correctly if it would have been an error without it, so it fixes very little).
A company valued at $19 billion will pay it one way or another.
There is effectively zero chance they could not get a loan, or if they willingly ignored the order and the company or some assets were sold at auction as a result, that it would not make that much.
They are absolutely good for any judgement rendered against them, as it will not be in the billions.
GA state minimum is $25,000 per person in bodily injury, likely not close to enough in this case. Sure they can go after the kid too and eventually garnish much of what they make, but that is also likely not to cover it.
The obvious target for a civil suit is the party with the ability to pay.
I did a chmod 700 / home/me on a production server a couple of years ago as root after setting another user up (note the space typo).
Total downtime was about two minutes, as I figured it out pretty quickly when it did not complete immediately and I hit up to look at my last command.
A chmod 777 which I let get about as far made it work temporarily, and fortunately RPM actually has a setting to restore permissions... but it was very inconvenient as there are rather a lot of permissions not set by that. I ended up setting up another server and not daring to restart that one.
rm I am careful enough with that I have never had a problem, but I also stare at that command for a minute before I hit enter if I am root.
I think if they actually start lobbing effective ICBMs our way the response would be much more immediate than an invasion. The US is not going to sit around and wait to see if they can prep another for launch quickly or deliver it some other way after the first strike, and an invasion takes time.
We can mostly neutralize the threat within minutes, and that is exactly what we will do if they nuke us.
You... think China would just pile on the nuke war?
Something tells me they would prefer to sit that one out. I bet the US would decide to sit it out if Mexico ran for the bomb and hit China for some reason too, that is not the kind of trouble you borrow.
Not really, you could design it such that it stabilized itself during entry.
If you have enough pressure to cause serious heating, you have enough pressure that it will naturally want to point the heavier side prograde, and present a minimal profile to the air it is punching through.
My bet would be that the AI will refuse to drive if it is not maintained well enough to do so safely. These systems will be redundant, so one failure should not cause an immediate crash.
As far as I know planes fly just fine, and batteries are only lacking in density (there are purely electric planes today, but the range is poor). If anyone can pull the maybe 3x improvement necessary to make this into a reality for short flights he can.
It is certainly not that far out there that I would call it a limitation derived from basic physics.
If it were anyone else, I would be a bit skeptical...
We are talking about the guy who owns the largest electric car company, is about to own the largest battery factory (employing many battery chemists and other specialists), and owns the most successful of the private space launch companies.
I think he is well aware of the tyranny of the rocket equation (as well as how it applies to aircraft as they burn their fuel), the limitations of a battery, and the differences between theoretical maximums and what you get in practice. More importantly he is advised by those who really are experts, and past success suggests that he does listen and understand.
Tripling battery density would make it viable on shorter flights, perhaps even cost effective. If anyone is in a position to make that happen, he is.
What they are offering is not internet access, it is a place in his walled garden (or maybe fenced pasture is better). Wikipedia and such are the sugar that he was hoping would help them swallow the bitter facebook medicine.
There are a lot of secondary effects to allowing something like this, from crushing any local competition with facebook before it begins, to letting them choose and price what it takes to reach the populace. If he is only going for the betterment of everyone involved, he can straight up donate the money to do this, or allow all sites which meet certain bandwidth requirements.
Mine would be easily circumvented if someone wanted to do that intentionally, but I consider my job done if it takes an intentional bypass of two safety systems to mangle yourself.
If it only decimates it I will be sorely disappointed, I expect automated cars to do much better than that.
Incidentally, I have programmed that robot in a cage. Mine stops moving if you trip an optical sensor on the way in (possibly damaging the robot due to the application of too much force in the process).
It will not drive off a cliff if it is aware of the cliff under any circumstances, it will instead come to a stop before the road ends.
If stopping in time is impossible as something was basically dropped into its path, it will end up hitting the object at the lowest speed it can achieve. It will never intentionally hit anything for any reason at all, and my expectation is that they will be very good at this. Accidents so far always involve the automated car being struck by rather than striking an object for a reason.
The "drive off of a cliff" vs "run over some kids" scenario is not something a car will be aware of at all, it is something humans with no programming experience relevant to this kind of project suggest in an attempt to make it something they can relate to.
If there is no way to avoid an accident, the car will attempt to stop in its lane as quickly as possible. There is no other conceivable way this could work due to the extreme liability any other decision would imply.
This will in most cases greatly minimize the forces involved in a collision as well.
Slashdot Mirror
When he says deep TCP/IP knowledge he probably means he needs a systems programmer. TCP/IP is such a common standard that anyone with the required background will already be familiar with it, so the training will not begin there.
If it were an entry level programming position teaching someone to program might be worthwhile, but there are also plenty of programmers with a security clearance. I imagine part of the problem he is having is that the security clearance does not mean as much to someone who has this skill set. The major draw to having one is the job security and higher salary, but if you can sit down and write a network stack or code to analyze packets in interesting ways in real time you already have an easy time finding employment, and the ability to make real money if you desire.
You seem to be stuck in the distant past with your opinion of C.
Yes, it entirely sucked when people tried to use it for general applications as a default language (same with C++, maybe even more so there). It is not at all an appropriate language when you are trying to deal with a ton of general business code implemented by a large team of varying quality (Java and C# do very well here). It is not really appropriate for someone who wants to learn to code as quickly as possible, nor does it make a particularly good prototyping language.
I think a few people already pointed you to all of the public linux stuff. MS also writes the kernel of Windows in C, all device drivers... also I did a bit of programming for industrial hardware, all of it was C. There are a few other areas where it ends up being very useful or nearly required, such as dealing with applications using wildly different hardware, a task which is at the limits of what a modern computer can do, or if more speed is always good (the last time I wrote a significant amount of assembly was actually a financial project, you would be surprised at where you end up needing the bag of tricks you can use with C and assembly to greatly speed up a process).
You hire expensive senior programmers by reference generally, not by posting on monster.com or whatever (unless you are really out of options).
C++ is not common in life critical software, maybe ADA (although I have never seen it, I also do not do government work). You will be applying KISS if you are designing that kind of system, in which case a large and complicated piece created externally is right out, so no C++ compiler. The fact that C has such a simple compiler is an important consideration.
It has never been lower than position 2. It has been first or second on that list since they started tracking it, as you can see if you scroll down a little bit for the chart over time (which goes back 15 years on that page).
If all of your jobs involve javascript, that likely makes you a web developer. Using some web frameworks to put together websites is about as far away as you can get from systems programming, so of course nobody is asking you to write a bunch of C code. It is not your profession, and the results are likely to be horrific.
The vast majority of important code is written in C, including new projects today. It dominates few industries, but is present in almost all of them (because when you run into a problem which needs C you have few other options, none of them easier). C# is pretty popular for general business programming, matlab is popular for scientific stuff, I still see a fair amount of cobol for big financial systems, php and like half a dozen others are popular for web development, etc. The thing is they have pretty narrow niches, so while they may be dominant in one industry, outside of that you basically do not see it.
I never see anyone write javascript for a big financial program (except maybe for the front end), write data analysis code for an experiment in C#, or write their website backend in matlab. I do however see financial programmers who think the world is entirely Java, C# and SAP, academics who think matlab can solve anything, and web developers who think PHP and Javascript are all you will ever need.
When your problem requires new hardware of any sort, you care about speed, or there is the possibility that someone may be harmed if the program does not work correctly, the language of choice is usually C.
I will also throw out there that I do a lot of code review, and the C programs I am sent have by far the lowest defect rate of what I review (PHP being the worst.)
All caps by many conventions means a precompiler define, this is a common way to set static buffer sizes in case they must change later.
The long term website tracking popularity (as well as you can do that for programming languages) is at http://www.tiobe.com/tiobe_index
C is pretty much always 1 or 2 in the list, and notice how it has twice the ranking of the third in the list (C++... which owes much of its popularity to the ability to compile most C programs). C is a very popular language, and is pretty much the language of choice whenever you need to do something novel, involving hardware, or care about execution time. It is not usually the best choice for a big CMS or other tasks where there are domain specific languages available, but that is mostly due to staffing concerns.
It tends not to have a ton of people in any given industry using it as it requires a greater understanding as to how the computer operates, which means you do not train decent C programmers in a few months like you do PHP or such (and you do not throw it at academics like MATLAB). It takes years to get good, and even then it turns out a large number of people will struggle with memory management no matter how hard they try.
If it takes you 3 or 4 times as long to write something in C, that is a good indication you should not be writing C. In contrast it is my favorite language and I am far more productive in it than Java, PHP, etc. (not that I use it all the time, I am well aware that if I write something in C I will likely need to continue maintaining it, so I usually need a reason to do so).
Unfortunately while it includes the fundamentally unsafe ones like gets(), it also deprecates a bunch of functions which are not fundamentally unsafe, such as the simple and super common strcpy (which is just fine if you know there is a null byte within the buffer length, generally because you as the programmer put it there).
The solution they use is also pretty bad, adding a buffer length parameter to the function is useless in practice as it still lets you overflow the buffer by getting this wrong, and it seems most code which would contain a buffer overflow ends up calculating this in a way which ends up passing the wrong buffer length. When combined with how common some of those functions are in existing code (where usage is perfectly safe!), I end up just turning it off by default in my compiler. It is an annoyance for no real gain.
I do not consider those useful warnings as MS implements them. It's like someone with minimal understanding as to why you would want to control your own memory layout who never works on code for an OS other than Windows was asked to fix buffer overflows in C, and came up with the great idea that if you had the programmer specify the length twice they might not screw it up in exactly the same way twice (which they do, in practice it is rare to see the size of the buffer provided correctly if it would have been an error without it, so it fixes very little).
A company valued at $19 billion will pay it one way or another.
There is effectively zero chance they could not get a loan, or if they willingly ignored the order and the company or some assets were sold at auction as a result, that it would not make that much.
They are absolutely good for any judgement rendered against them, as it will not be in the billions.
GA state minimum is $25,000 per person in bodily injury, likely not close to enough in this case. Sure they can go after the kid too and eventually garnish much of what they make, but that is also likely not to cover it.
The obvious target for a civil suit is the party with the ability to pay.
I did a chmod 700 / home/me on a production server a couple of years ago as root after setting another user up (note the space typo).
Total downtime was about two minutes, as I figured it out pretty quickly when it did not complete immediately and I hit up to look at my last command.
A chmod 777 which I let get about as far made it work temporarily, and fortunately RPM actually has a setting to restore permissions... but it was very inconvenient as there are rather a lot of permissions not set by that. I ended up setting up another server and not daring to restart that one.
rm I am careful enough with that I have never had a problem, but I also stare at that command for a minute before I hit enter if I am root.
You should see the button, it is quite intimidating. Lesser men cower in fear, and only a true hero is able to make himself approach to press it.
Or at least that is how the union rep describes the position.
I had to show ID to vote in the primary recently, I seem to recall doing that at all elections in recent years.
It may differ by state.
I think if they actually start lobbing effective ICBMs our way the response would be much more immediate than an invasion. The US is not going to sit around and wait to see if they can prep another for launch quickly or deliver it some other way after the first strike, and an invasion takes time.
We can mostly neutralize the threat within minutes, and that is exactly what we will do if they nuke us.
You... think China would just pile on the nuke war?
Something tells me they would prefer to sit that one out. I bet the US would decide to sit it out if Mexico ran for the bomb and hit China for some reason too, that is not the kind of trouble you borrow.
Not really, you could design it such that it stabilized itself during entry.
If you have enough pressure to cause serious heating, you have enough pressure that it will naturally want to point the heavier side prograde, and present a minimal profile to the air it is punching through.
My bet would be that the AI will refuse to drive if it is not maintained well enough to do so safely. These systems will be redundant, so one failure should not cause an immediate crash.
Which physical limitations are we speaking of?
As far as I know planes fly just fine, and batteries are only lacking in density (there are purely electric planes today, but the range is poor). If anyone can pull the maybe 3x improvement necessary to make this into a reality for short flights he can.
It is certainly not that far out there that I would call it a limitation derived from basic physics.
If it were anyone else, I would be a bit skeptical...
We are talking about the guy who owns the largest electric car company, is about to own the largest battery factory (employing many battery chemists and other specialists), and owns the most successful of the private space launch companies.
I think he is well aware of the tyranny of the rocket equation (as well as how it applies to aircraft as they burn their fuel), the limitations of a battery, and the differences between theoretical maximums and what you get in practice. More importantly he is advised by those who really are experts, and past success suggests that he does listen and understand.
Tripling battery density would make it viable on shorter flights, perhaps even cost effective. If anyone is in a position to make that happen, he is.
There are things which would kill it, but antibiotics are not one of those things.
The problem is that most of the things which would wipe out all of the yeast would also kill you.
That would certainly save a lot of effort, and I doubt anyone would object to that.
This is more like giving out free candy if you get into the car than working a soup kitchen.
What they are offering is not internet access, it is a place in his walled garden (or maybe fenced pasture is better). Wikipedia and such are the sugar that he was hoping would help them swallow the bitter facebook medicine.
There are a lot of secondary effects to allowing something like this, from crushing any local competition with facebook before it begins, to letting them choose and price what it takes to reach the populace. If he is only going for the betterment of everyone involved, he can straight up donate the money to do this, or allow all sites which meet certain bandwidth requirements.
If his motivations are entirely philanthropic in nature, there is an easy solution to their concerns:
Remove Facebook from the free list, and let the government decide what qualifies for the plan.
Mine would be easily circumvented if someone wanted to do that intentionally, but I consider my job done if it takes an intentional bypass of two safety systems to mangle yourself.
If it only decimates it I will be sorely disappointed, I expect automated cars to do much better than that.
Incidentally, I have programmed that robot in a cage. Mine stops moving if you trip an optical sensor on the way in (possibly damaging the robot due to the application of too much force in the process).
It will not drive off a cliff if it is aware of the cliff under any circumstances, it will instead come to a stop before the road ends.
If stopping in time is impossible as something was basically dropped into its path, it will end up hitting the object at the lowest speed it can achieve. It will never intentionally hit anything for any reason at all, and my expectation is that they will be very good at this. Accidents so far always involve the automated car being struck by rather than striking an object for a reason.
The "drive off of a cliff" vs "run over some kids" scenario is not something a car will be aware of at all, it is something humans with no programming experience relevant to this kind of project suggest in an attempt to make it something they can relate to.
This response is exactly what will happen.
If there is no way to avoid an accident, the car will attempt to stop in its lane as quickly as possible. There is no other conceivable way this could work due to the extreme liability any other decision would imply.
This will in most cases greatly minimize the forces involved in a collision as well.