"I have a car, a sportscar, it's 8 years old. Do I want a new car? Hell no! Does my car have a back-up camera?"
FWIW, I just installed a backup camera in my 18 year old Toyota Camry. It's useful because they stopped making cars with adequate rear visibility about four decades ago. But I didn't have to buy a new car with no spare tire and a lot of truly obnoxious electronics in order to get the capability. (The camera was actually a bonus. I really bought the new radio mostly because the old one had no decent way to play mp3s).
My guess would be that at least 15% of US engineers would be fit to produce software by the Indian criterea. The other 85% can do website development or, if unfit for even that, can get an MBA and become managers.
"Only the bug reporter can find out whether what you fixed also fixed their problem. You can track that - and say the issue is either confirmed fixed, or unconfirmed fixed."
Basically, I agree. But I don't think it's simple at all. Any organization where developers and "stakeholders" aren't in close contact tends to function like an elaborate game of telephone/chinese whispers https://en.wikipedia.org/wiki/.... Written language without feedback is far from a perfect form of communication. For several years I was in charge of testing a very marginal system and ended up doing more of the user liaison than I (or anybody else) really wanted me to do. I used to call the author of every bug report after we released a fix. A remarkable percentage of the time the fix either didn't really work, or fixed the wrong problem or sometimes just made things worse.
I also used to call every bug report author whose complaint was rejected by developers because the system was "working as designed". (Something of a misstatement. Very little of the original design actually got implemented). That was not a whole lot of fun, but it did allow us to get a lot of what the users wanted into the (overly cumbersome) design change process.
Some of that. But I made quite a good living for several decades back in the 1960s, 1970s, and 1980s integrating large systems. Did some development also, and have some sympathy for developers. Good unit tests help. and so would automated testing. But let's get real. Few actual deployed systems have meaningful specs, and even those that started with specs probably didn't maintain the specs and have bugs in the original specs -- omissions (oh, you wanted the trig functions to be fast as well as accurate? should have said so). bad ideas, good ideas poorly specified, ambiguous equations. (how can equations be ambiguous? The, hands down best spec I ever saw had an equation of the form (A+B)/(C-D). Except that A,B,C,&D were all pretty complicated. But it was written in the spec the way you might write it on a blackboard without the parentheses and got programmed as ((A+B/C) -D -- oopsie) , etc,etc,etc. It all adds up.
IMHO, most folks grossly underestimate the number of yet to be discovered bugs lurking in virtually all software.
Which is why they assume that with just one or two more fixes, everything will be fine.
And is also why patching our way to security very likely will never really work.
Working with most systems is like trying to keep a 1970 car on the road. It may run fine on good days. It may even get decent gas mileage and not be terribly polluting. My 1979 Mazda GLC got better gas mileage than any of the cars in my driveway today. And it had not only a catalytic converter, but an air pump and EGR and lord knowns what else (The actual hardware was more complex than the shop manual description--a fact I discovered only when trying to reassemble the engine after replacing the head gasket... This hose from the mysterious gidgie near the wiper motors. Where does it go to?). The emissions numbers weren't bad at all even when compared to much newer cars. But between rust and the unavailability of repair parts, I eventually had to junk it.
It doesn't not work. More eyes do tend to catch out and out bugs. But some still slip through. And stuff that sounds good but busts user's workflow still gets through. And developers hate the delay. And more review costs more. And more testing costs even more than more review. So managers aren't a big fan of either.
I suppose the solution is never to work with millions of lines of code. But "This sucker is way to big and too complicated" is not an easy sell.
It's good to see that at least one of our prisons is teaching inmates marketable skills that will allow them to earn a decent living once they are released.
"How many billion dollars has "Thunderf00t" made with his space and car companies? Yes, it's the same number as you - none"
About the same number as Musk's high tech schema also. Solar Cities seems to be (have been?) a money sucking disaster whose troubles have been hidden by tucking it into Tesla . Many folks think it was/is a scam aimed a skimming off government subsidies and leaving homeowners with an incomprehensible, unbreakable lease on equipment that doesn't work as promised. Tesla looks likely to be a monumental bubble although it possibly has some chance of pulling through. The battery factory may well be OK although it faces REALLY stiff East Asian competition. Space X is an admirable effort and it may eventually show significant profits depending on how many of the reused rockets self destruct leaving the company with a bill for the destruction. But if you believe the WSJ, SpaceX profits were slim in 2014, negative in 2015 and probably minimal to non-existent in 2016.
It is conceivable that a tube wall sturdy enough to hold up under 15 psi = 103kPa pressure won't leak a whole lot. I'm assuming sliding or dilating or something-ing barriers at the stations to keep trash, luggage, small children, etc from being sucked into the traveling portion of the tube.
OTOH, the cost of a hyperloop tube -- whether buried, run along the surface, or suspended from giant genetically modified vultures -- is likely to be pretty high compared to a rail line. It might be cheaper to have a SpaceX rockets fly you from LA to San Diego than to buy a hyperloop ticket.
"We already have a complete infrastructure in place that will allow you to go from just about any city to just about any other city at high speed."
Sure -- assuming that you don't count the hour getting to the airport, the ever shrinking seats, the two hour "security" delay, the overnight stay when you miss your connection in Chicago, Atlanta, or some other garden spot, and the hour collecting your luggage (if it takes the same flight(s)) and getting from your putative destination to where you really want to be. And also assuming that it's not raining or snowing anywhere in North America.
But what the Hell. Maybe The Donald will rein in the TSA and make the planes fly on time.
"Uber Said To Use 'Sophisticated' Software To Defraud Drivers, Passengers"
Really now. You wouldn't expect a high tech company like Uber to use unsophisticated software to rip off its customers and employees... ehrr.. independent contractors?
Interesting. But let's not forget that Moore's law is about exponential growth of transistor density, not NOT exponential improvement in performance. The transistors are (I believe) being fruitful and multiplying as they have been for four decades. What, if anything, useful is done with them is another issue.
Seriously, jetpacks and the like have a long history of leg and knee injuries. Dropping out of the sky with a heavy object strapped to your back will do that. This thing -- if it's not an April 1 joke -- probably just expands that problem to other extremities.
Computer controlled engines -- with ads and a wi-fi connection to the web. No need to tell the advertisers that the customer will very shortly be buried or seriously incapacitated..
Good questions. Based on prior experience, the answers are:
1,. YES, you can use this vulnerability to brick a TV 2. YES, the manufacturer is legally liable 3. NO, the manufacturers will not have to take security seriously. There is no force in the known universe capable of forcing a typical IOT vendor to take security seriously.
"When I can manufacture in China where I don't have to pay for air scrubbers or sewage treatment"
My understanding is that Chinese laws require most of that hardware. In the US you have to pay to keep it running. In China, not so much, but you have to pay the local officials to let you not fix it. At least that's what I've been led to believe.
Correct me if I'm wrong, but didn't Sears shut down their catalog business in the early 1990s. I recall thinking at the time that walking away from a captive audience of 8 or 10 million rural customers who didn't really want to treck beyond the local small town general store, garage, movie theatre, and bar to shop was weird. What sort of business walks away from 8 or 10 million customers?
"it makes less sense to locate manufacturing thousands of miles over an ocean from the market, and I imagine what will eventually happen is a good deal of manufacturing happening closer to major markets to bring down distribution costs, but you're not really going to see any significant increase in jobs."
Exactly. There may be some shipping, receiving, and shlepping jobs created as manufacturers and parts suppliers move back onshore. And some robot maintenance and repair jobs. But the days when protective tariffs protected jobs as well as profits are likely pretty much over. If buying, setting up and fixing robots costs less than the fully burdened cost of an employee, the jobs are going to go to the bots. And robots are very unlikely to engage in annoying practices like walking picket lines.
There may be countries that will adapt to this brave new world with minimal disruption. I somehow don't think the US is going to be one of them.
Slashdot Mirror
Excel? 70MB? Golly Gee whiz, what could the problem be here? Nothing for it, we'll just need to get you a faster computer ...
"I have a car, a sportscar, it's 8 years old. Do I want a new car? Hell no! Does my car have a back-up camera?"
FWIW, I just installed a backup camera in my 18 year old Toyota Camry. It's useful because they stopped making cars with adequate rear visibility about four decades ago. But I didn't have to buy a new car with no spare tire and a lot of truly obnoxious electronics in order to get the capability. (The camera was actually a bonus. I really bought the new radio mostly because the old one had no decent way to play mp3s).
My guess would be that at least 15% of US engineers would be fit to produce software by the Indian criterea. The other 85% can do website development or, if unfit for even that, can get an MBA and become managers.
"Only the bug reporter can find out whether what you fixed also fixed their problem. You can track that - and say the issue is either confirmed fixed, or unconfirmed fixed."
Basically, I agree. But I don't think it's simple at all. Any organization where developers and "stakeholders" aren't in close contact tends to function like an elaborate game of telephone/chinese whispers https://en.wikipedia.org/wiki/.... Written language without feedback is far from a perfect form of communication. For several years I was in charge of testing a very marginal system and ended up doing more of the user liaison than I (or anybody else) really wanted me to do. I used to call the author of every bug report after we released a fix. A remarkable percentage of the time the fix either didn't really work, or fixed the wrong problem or sometimes just made things worse.
I also used to call every bug report author whose complaint was rejected by developers because the system was "working as designed". (Something of a misstatement. Very little of the original design actually got implemented). That was not a whole lot of fun, but it did allow us to get a lot of what the users wanted into the (overly cumbersome) design change process.
Some of that. But I made quite a good living for several decades back in the 1960s, 1970s, and 1980s integrating large systems. Did some development also, and have some sympathy for developers. Good unit tests help. and so would automated testing. But let's get real. Few actual deployed systems have meaningful specs, and even those that started with specs probably didn't maintain the specs and have bugs in the original specs -- omissions (oh, you wanted the trig functions to be fast as well as accurate? should have said so). bad ideas, good ideas poorly specified, ambiguous equations. (how can equations be ambiguous? The, hands down best spec I ever saw had an equation of the form (A+B)/(C-D). Except that A,B,C,&D were all pretty complicated. But it was written in the spec the way you might write it on a blackboard without the parentheses and got programmed as ((A+B/C) -D -- oopsie) , etc,etc,etc. It all adds up.
IMHO, most folks grossly underestimate the number of yet to be discovered bugs lurking in virtually all software.
Which is why they assume that with just one or two more fixes, everything will be fine.
And is also why patching our way to security very likely will never really work.
Working with most systems is like trying to keep a 1970 car on the road. It may run fine on good days. It may even get decent gas mileage and not be terribly polluting. My 1979 Mazda GLC got better gas mileage than any of the cars in my driveway today. And it had not only a catalytic converter, but an air pump and EGR and lord knowns what else (The actual hardware was more complex than the shop manual description--a fact I discovered only when trying to reassemble the engine after replacing the head gasket ... This hose from the mysterious gidgie near the wiper motors. Where does it go to?). The emissions numbers weren't bad at all even when compared to much newer cars. But between rust and the unavailability of repair parts, I eventually had to junk it.
"That doesn't work"
It doesn't not work. More eyes do tend to catch out and out bugs. But some still slip through. And stuff that sounds good but busts user's workflow still gets through. And developers hate the delay. And more review costs more. And more testing costs even more than more review. So managers aren't a big fan of either.
I suppose the solution is never to work with millions of lines of code. But "This sucker is way to big and too complicated" is not an easy sell.
It's good to see that at least one of our prisons is teaching inmates marketable skills that will allow them to earn a decent living once they are released.
"How many billion dollars has "Thunderf00t" made with his space and car companies? Yes, it's the same number as you - none"
About the same number as Musk's high tech schema also. Solar Cities seems to be (have been?) a money sucking disaster whose troubles have been hidden by tucking it into Tesla . Many folks think it was/is a scam aimed a skimming off government subsidies and leaving homeowners with an incomprehensible, unbreakable lease on equipment that doesn't work as promised. Tesla looks likely to be a monumental bubble although it possibly has some chance of pulling through. The battery factory may well be OK although it faces REALLY stiff East Asian competition. Space X is an admirable effort and it may eventually show significant profits depending on how many of the reused rockets self destruct leaving the company with a bill for the destruction. But if you believe the WSJ, SpaceX profits were slim in 2014, negative in 2015 and probably minimal to non-existent in 2016.
It is conceivable that a tube wall sturdy enough to hold up under 15 psi = 103kPa pressure won't leak a whole lot. I'm assuming sliding or dilating or something-ing barriers at the stations to keep trash, luggage, small children, etc from being sucked into the traveling portion of the tube.
OTOH, the cost of a hyperloop tube -- whether buried, run along the surface, or suspended from giant genetically modified vultures -- is likely to be pretty high compared to a rail line. It might be cheaper to have a SpaceX rockets fly you from LA to San Diego than to buy a hyperloop ticket.
"Were the Japanese subjected to medical experiments or put to death?"
Of course not. We used African-Americans for experiments. Asians, we mostly just tried to keep out.
"We already have a complete infrastructure in place that will allow you to go from just about any city to just about any other city at high speed."
Sure -- assuming that you don't count the hour getting to the airport, the ever shrinking seats, the two hour "security" delay, the overnight stay when you miss your connection in Chicago, Atlanta, or some other garden spot, and the hour collecting your luggage (if it takes the same flight(s)) and getting from your putative destination to where you really want to be. And also assuming that it's not raining or snowing anywhere in North America.
But what the Hell. Maybe The Donald will rein in the TSA and make the planes fly on time.
"Nazi Germany reminds me of Japanese American internment camps"
Not one of our finer moments, OTOH, we didn't try to exterminate the Japanese (or German or Italian) internees.
"Uber Said To Use 'Sophisticated' Software To Defraud Drivers, Passengers"
Really now. You wouldn't expect a high tech company like Uber to use unsophisticated software to rip off its customers and employees ... ehrr .. independent contractors?
"Once the jet is in the air, they will turn off the jet engine, and turn on the battery motor to keep the plane up."
By flapping the wings?
Pilot? You think there's going to be a pilot?
Interesting. But let's not forget that Moore's law is about exponential growth of transistor density, not NOT exponential improvement in performance. The transistors are (I believe) being fruitful and multiplying as they have been for four decades. What, if anything, useful is done with them is another issue.
"a fast edit-test-debug cycle"
There is testing and debugging in the cycle? who knew? I thought the industry standard was "worked once-ship it -- we'll fix it in production"
Seriously, jetpacks and the like have a long history of leg and knee injuries. Dropping out of the sky with a heavy object strapped to your back will do that. This thing -- if it's not an April 1 joke -- probably just expands that problem to other extremities.
"There should be computer controlled engines"
Computer controlled engines -- with ads and a wi-fi connection to the web. No need to tell the advertisers that the customer will very shortly be buried or seriously incapacitated..
It's no crazier than any other IOT device.
"In the event of an unavoidable crash the control is RELEASED to you"
So, your Belchfire 5000 gets itself stuck on railroad tracks with a freight train barreling toward you at 100kph ... and it turns control over to you.
This will come to be known as the "Hasta la vista, Baby" patent
Good questions. Based on prior experience, the answers are:
1,. YES, you can use this vulnerability to brick a TV
2. YES, the manufacturer is legally liable
3. NO, the manufacturers will not have to take security seriously. There is no force in the known universe capable of forcing a typical IOT vendor to take security seriously.
You got my vote, mate.
"When I can manufacture in China where I don't have to pay for air scrubbers or sewage treatment"
My understanding is that Chinese laws require most of that hardware. In the US you have to pay to keep it running. In China, not so much, but you have to pay the local officials to let you not fix it. At least that's what I've been led to believe.
Correct me if I'm wrong, but didn't Sears shut down their catalog business in the early 1990s. I recall thinking at the time that walking away from a captive audience of 8 or 10 million rural customers who didn't really want to treck beyond the local small town general store, garage, movie theatre, and bar to shop was weird. What sort of business walks away from 8 or 10 million customers?
"it makes less sense to locate manufacturing thousands of miles over an ocean from the market, and I imagine what will eventually happen is a good deal of manufacturing happening closer to major markets to bring down distribution costs, but you're not really going to see any significant increase in jobs."
Exactly. There may be some shipping, receiving, and shlepping jobs created as manufacturers and parts suppliers move back onshore. And some robot maintenance and repair jobs. But the days when protective tariffs protected jobs as well as profits are likely pretty much over. If buying, setting up and fixing robots costs less than the fully burdened cost of an employee, the jobs are going to go to the bots. And robots are very unlikely to engage in annoying practices like walking picket lines.
There may be countries that will adapt to this brave new world with minimal disruption. I somehow don't think the US is going to be one of them.