Well what is prudent and necessary to do is a matter of what threats exist in practice. If *all* those concerns you listed proved to be popular vectors for malware you'd have to check all those things, or admit to customers your software is untrustworthy.
Fortunately there are intermediate levels of paranoia between trusting everything blindly and building everything up from machine language. There are choices you can make about who to trust and what steps to take to verify that trust. This story demonstrates that. Apple caught the problem before allowing the app in their store, which successfully prevented the distribution of the malware in this case. The developer in turn figured out which of his vendors was at fault.
Is that a lot of effort doing something that wouldn't be necessary in an ideal world? Sure. Is it annoying? Absolutely. Is it beyond the ability of a skilled professional developer to deal with practically? No.
Actually the average or amortized time complexity of hashing insertion is much better than the worst case. In fact they're constant, provided you have enough space to make collisions rare. So the "use hashing for everything" trick is reasonable heuristic for many tasks, but of course not all of them. Knowing how to balance the concern about worst case against the concern about average case is a matter of judgment, which is frequently lacking in people who fetishize this stuff. There are times when a compact O(n^2) algorithm will outperform a complex O(n log(n)) algorithm for all relevant inputs.
I note that they refer to a 1000x improvement in "performance", and a 6.4x improvement in "latency". Latency is one time-related performance metric; throughput might be the other be alluded to.
Imagine two water hoses. One is ten feet long and one inch wide; the other is inches in diameter and a hundred feet long. Which can deliver water "faster"? Well, when you turn the spigot on water comes out of the ten foot first; but if you're filling up a swimming pool the hundred foot long hose is faster.
I'm actually old enough to remember when metal detectors and baggage screen were made mandatory in 1972. It made a HUGE difference; by 1970 or so major hijackings had become a multiple times/year event (check out Wikipedia's list). It took almost a decade but by the 80s hijackers successfully boarding at US airports became a rare event.
We need to think in terms of two things:
(a) marginal utility; and (b) patching specific vulnerabilities as they are exposed.
The thing is baggage screening, metal detectors and requiring ID were the low hanging fruit in terms of systematic improvements. Reinforcing cockpit doors was a no-brainer after 9/11. After that there just wasn't a lot of numerical improvement to be had when things like full body scanners came along. I actually believe attempting to fix parts of the system where very little marginal improvement is possible is a scam. I was working in stuff that was remotely applicable to bioterrorism and we were multiple times approached by politically connected VCs looking to repackage our stuff to cash in on the homeland security bonanza; we declined. Meanwhile other things like physically securing the airport tarmac were neglected.
According to the linked article there was something funky going on with the CO2/mileage certification process.
Granted the certification figures for all vehicles are optimistic, but that doesn't make them useless. I've spent many years working with environmental and scientific data, and it's often the case that you can't know certain things precisely. Nonetheless it is still important to measure these things in a consistent manner so you can compare figures to each other.
So suppose car A's test say it emits 282 g / km of CO2 traveled -- about 20 miles/gallon -- car B's test say it emits 188 g/mile (i.e. gets 30 mpg). While both cars in fact emit *more* CO2 than the tests say, I can at least rank them, and even get a rough sense that car B isn't just slightly more efficient than A; it's a lot more efficient. What's more if I have a maximum standard for CO2/km some cars may in practice emit a bit more than that standard, but since I can at least rank cars I can keep the most polluting cars off the road.
If the tests are somehow cooked, all that goes out the window and VW gets a fraudulent advantage over its competitors. So it does matter, even if test results are consistently more optimistic than real-world performance.
Now your point about hybrids is well taken; comparing a hybrid to a pure ICE car is an apples-to-oranges comparison because if they test approximately the same which is ranked better will depend on the specific kind of driving you do. However even then the comparison is not totally useless; presumably they use a model which they think accurately represents how an average driver might use a hybrid. It's very likely that you as an individual consumer won't be very close to that "average" driver, but the figures may be reasonable when aggregated by all the cars of that model on the road.
The syntax is not hard to unravel, although the combination of the syntax AND the jargon (the most common source of Slashdot incoherence) is a bit of a challenge. The problem is a stray "that". "That" is used to introduce a restrictive relative clause (i.e., a phrase which narrows down the universe of possible things that "the twitter account" might refer to). The tweeter began a sentence with what he thought was the subject, then introduced a restrictive relative clause to clarify the subject; when he discovered that what he wanted to say was in that clause he simply slapped a period on the end.
The simplest edit would be to make "pastebin" the subject of the sentence:
The pastebin released by the twitter account included government officials who clearly are not KKK.
On the other hand, the subject in a subject-verb-predicate sentence has a kind of pride of place. I suspect that the writer was trying to put the spotlight on what he saw was unethical behavior by the tweeter. So you could also write it this way:
The twitter account that released the pastebin included government officials who clearly are not KKK in that pastebin.
I remember being so blown away by OS2/Warp's ability to multi-task so many applications at once, with such a clean UI.
As an older programmer, let me suggest one ought to be reticent about saying things like that. I know that by any reasonable standard it should make you sound experienced and therefore worth listening to, but if you have any gray in your hair it's bound to have a very different effect. Like the time I sat next to a guy at a banquet who was reminiscing about when his department got an IBM 701. "Yep," he said with evident satisfaction, "that was a stored program jobbie."
Employers are looking for programmers who were in diapers while you were being blown away by OS/2, so ixnay on that kind of alktay. Instead practice saying things like "Node.js is so 2015." And when someone asks you what you mean, turn to them, raise one eyebrow, then literally turn your back on them.
Well, it's not 800 Euros a month hot off the printing press; it's 800 Euros that were taxed out of the economy then put back into the economy in a different place. It'll surely effect the prices of many things, but net there's no more total money in the economy.
I suspect the thinking is that many of the things that people on the lower income end of the spectrum have relatively inelastic demand: housing and foodstuffs. Things that are discretionary purchases for those people are bound to become more expensive.
Sure, but what matters is the performance of *systems*; and if you set aside battery one one hand and gas tank on the other, an electric powertrain is much, much lighter than a gasoline powertrain. If you got a 10x increase in energy to weight in battery technology, that'd come right off what is by far the electric car's heaviest component. The result would be dramatic improvements in the car's performance and efficiency.
If it also comes with the kind of cost reductions they're imagining, you could be talking about a car about the size and cost of a Nissan Leaf but with a 500 mile range and about 10% lighter.
However I'm not holding my breath. The air breathing battery technology idea isn't new; the problem is longevity. These guys have some ideas for improving that, but it remains to be seen whether those'll lead to a practical vehicle battery with the kind of performance they're hoping for.
I think the selling point is extensibility. I've been playing around with it and I detect a great deal of Emacs influence in the editor's features. I think the idea is to provide emacs-like power user features with a basic text editor-like entry level experience.
I think the primary reason is that they want the editor to be extensible using javascript, which of course is possible with a browser based app but then the people doing the customization have to work around browser security policies like single origin which complicate scripting for reasons that don't apply to a native editor.
It's a text editor, built with web-based technologies and packaged as an app-specific browser that looks like a native desktop app, which offers a user experience similar to most desktop programmer's editors but with emacs-like extensibility (only in javascript rather than elisp).
I've been the Hinton Lab in Jamaica Plain dozens of times. The people I worked with there from the lowest lab tech to the middle managers were outstanding, but they were in the epidemiology end of things. The drug testing lab was segregated on its own floor, and it was walled off like a fortress. But despite that superficial formidability of the drug testing lab, there was clearly a problem: back in 2007 the director of the crime lab resigned because mishandling DNA tests, and before that the lab had been in trouble for processing DNA too slowly. There were rape kits that had been waiting to be processed for eighteen years.
Yet despite the review of the crime lab's procedures that followed this scandal, Annie Dookhan was able to continue with her antics for an other four years before she was caught. It's odd that she was even hired with her phony degrees because that was the year it came out that Ralph Timperi, the Hinton Lab's overall director, got his PhD from a diploma mill. You'd think that'd trigger a little more scrutiny.
It all makes the entire Hinton Lab sound like a hot mess, but with the exception of Timperi's phony degree all the problems were in the crime lab, which while located inside the Hinton Lab building was (IIRC) actually overseen by the Massachusetts State Police. Possibly some kind of responsibility thing was going on there. On the public health side of things the people at the state labs were among of the best public employees I've ever dealt with, and I've worked with state and county agencies across the country. It's a shame they've been tainted.
What would be doomed is his vision of what NASA should be doing ten to twenty years out. At present manned exploration of Mars is not even big enough to warrant a line item in NASA's 2016 budget request, which means it's not a big part of what NASA is doing now. Mars exploration *is*, though, with 412 million requested to do serious and productive science.
You know what *would* doom NASA? Gutting its Earth and planetary research to fund an astronomically expensive manned mission to Mars in twenty years. I think an excuse for gutting NASA's Earth science program is part of the political imperative for a manned Mars mission.
I'm not against a manned mission to Mars per se; I'm against the opportunity cost. For the price of keeping a human crew alive and healthy for six years in bathed in deep space radiation and setting them down on the surface of Mars with an acceptable safety margin (very hard to do), we could get so much more done with robots. Extrapolating from Apollo to a manned Mars mission is very risky; Mars is immensely harder than the Moon in almost every respect. I don't think America is ready to spend Apollo program dollars every year for two or three decades, so what we'll get is a half-assed program that starves everything else NASA ought to be doing with very dubious prospects of success.
Eventually the time will come when the marginal value of a human mission will exceed the marginal value of the additional robotic missions the same funding would secure. That will be the time to start planning a manned Mars mission. I expect we might even get humans there faster that way.
Not really. First of all he wasn't an electrician, he was something called a cable jointer; it doesn't require a license and normally they don't work on live stuff. Secondly, according to TFA he wasn't aware the line was live, which means someone in a supervisory position screwed up.
As to whose fault it is, that depends. If the company in charge of the site has a proper lock out/tag out policy and training program, it's probably some supervisor's fault -- unless the worker went into an area specifically against instructions in which case it's the worker's fault. If the company doesn't have a proper LOTO program (including training for all personnel on site) it's the company's fault.
Right, because when you change one thing, everything else has to stay the same, otherwise things would get to complicated to discuss in a post like yours.
And somehow the other part of macro (saving during booms) never gets the same airing.
Funny that.
I actually think is an important point. I've often felt that the best time for a government to tighten its belt is when there is full private sector employment. That's because when you tax a dollar out of the private sector at that point you're doing at a time when the private sector is using dollars as many dollars as it can get its hands on to create jobs. It doesn't happen because that's a time a politician can look like a hero without inflicting sacrifice on anyone.
Of course there are other reasons to spend besides macroeconomics of course. The reason you fund a meals-on-wheels for elderly shut-ins isn't because of the economic stimulus effect. Same goes for defense projects -- although clearly the most advantageous time to cancel a major program or to close bases would be during a full employment period.
My problem with these discussions is that people on both sides talk about "spending" as if all spending is fungible. In my experience it's not. Some things are smart to spend money on, other things are stupid to spend money on; and also it's very easy to spend money stupidly on things that are unquestionably necessary. It's one thing for everyone to agree that, say, defense is an important thing to spend money on; and yet another thing to assume that all defense spending is good.
I think it's quite possible for a country spend its way to solvency -- if it spent overwhelmingly wisely. It's also possible to economize your way to bankruptcy, if you cut overwhelmingly foolishly. Of course it's also possible to economize wisely or spend yourself into a crisis. I've seen every possible combination, both in the private sector and the public sector. Private enterprises can become uncompetitive when management is unwilling to spend the money it needs to compete.
Slashdot Mirror
Well what is prudent and necessary to do is a matter of what threats exist in practice. If *all* those concerns you listed proved to be popular vectors for malware you'd have to check all those things, or admit to customers your software is untrustworthy.
Fortunately there are intermediate levels of paranoia between trusting everything blindly and building everything up from machine language. There are choices you can make about who to trust and what steps to take to verify that trust. This story demonstrates that. Apple caught the problem before allowing the app in their store, which successfully prevented the distribution of the malware in this case. The developer in turn figured out which of his vendors was at fault.
Is that a lot of effort doing something that wouldn't be necessary in an ideal world? Sure. Is it annoying? Absolutely. Is it beyond the ability of a skilled professional developer to deal with practically? No.
Well, neither are human drivers.
That's like worrying about whales pissing in the ocean.
Thank you so much cheap fossil fuel prices.
There. Fixed that for you.
Well-spotted -- from an AC too.
Actually the average or amortized time complexity of hashing insertion is much better than the worst case. In fact they're constant, provided you have enough space to make collisions rare. So the "use hashing for everything" trick is reasonable heuristic for many tasks, but of course not all of them. Knowing how to balance the concern about worst case against the concern about average case is a matter of judgment, which is frequently lacking in people who fetishize this stuff. There are times when a compact O(n^2) algorithm will outperform a complex O(n log(n)) algorithm for all relevant inputs.
One out of ten of the Internet's top web sites doesn't leak your information!
I note that they refer to a 1000x improvement in "performance", and a 6.4x improvement in "latency". Latency is one time-related performance metric; throughput might be the other be alluded to.
Imagine two water hoses. One is ten feet long and one inch wide; the other is inches in diameter and a hundred feet long. Which can deliver water "faster"? Well, when you turn the spigot on water comes out of the ten foot first; but if you're filling up a swimming pool the hundred foot long hose is faster.
I'm actually old enough to remember when metal detectors and baggage screen were made mandatory in 1972. It made a HUGE difference; by 1970 or so major hijackings had become a multiple times/year event (check out Wikipedia's list). It took almost a decade but by the 80s hijackers successfully boarding at US airports became a rare event.
We need to think in terms of two things:
(a) marginal utility; and
(b) patching specific vulnerabilities as they are exposed.
The thing is baggage screening, metal detectors and requiring ID were the low hanging fruit in terms of systematic improvements. Reinforcing cockpit doors was a no-brainer after 9/11. After that there just wasn't a lot of numerical improvement to be had when things like full body scanners came along. I actually believe attempting to fix parts of the system where very little marginal improvement is possible is a scam. I was working in stuff that was remotely applicable to bioterrorism and we were multiple times approached by politically connected VCs looking to repackage our stuff to cash in on the homeland security bonanza; we declined. Meanwhile other things like physically securing the airport tarmac were neglected.
According to the linked article there was something funky going on with the CO2/mileage certification process.
Granted the certification figures for all vehicles are optimistic, but that doesn't make them useless. I've spent many years working with environmental and scientific data, and it's often the case that you can't know certain things precisely. Nonetheless it is still important to measure these things in a consistent manner so you can compare figures to each other.
So suppose car A's test say it emits 282 g / km of CO2 traveled -- about 20 miles/gallon -- car B's test say it emits 188 g/mile (i.e. gets 30 mpg). While both cars in fact emit *more* CO2 than the tests say, I can at least rank them, and even get a rough sense that car B isn't just slightly more efficient than A; it's a lot more efficient. What's more if I have a maximum standard for CO2/km some cars may in practice emit a bit more than that standard, but since I can at least rank cars I can keep the most polluting cars off the road.
If the tests are somehow cooked, all that goes out the window and VW gets a fraudulent advantage over its competitors. So it does matter, even if test results are consistently more optimistic than real-world performance.
Now your point about hybrids is well taken; comparing a hybrid to a pure ICE car is an apples-to-oranges comparison because if they test approximately the same which is ranked better will depend on the specific kind of driving you do. However even then the comparison is not totally useless; presumably they use a model which they think accurately represents how an average driver might use a hybrid. It's very likely that you as an individual consumer won't be very close to that "average" driver, but the figures may be reasonable when aggregated by all the cars of that model on the road.
The syntax is not hard to unravel, although the combination of the syntax AND the jargon (the most common source of Slashdot incoherence) is a bit of a challenge. The problem is a stray "that". "That" is used to introduce a restrictive relative clause (i.e., a phrase which narrows down the universe of possible things that "the twitter account" might refer to). The tweeter began a sentence with what he thought was the subject, then introduced a restrictive relative clause to clarify the subject; when he discovered that what he wanted to say was in that clause he simply slapped a period on the end.
The simplest edit would be to make "pastebin" the subject of the sentence:
On the other hand, the subject in a subject-verb-predicate sentence has a kind of pride of place. I suspect that the writer was trying to put the spotlight on what he saw was unethical behavior by the tweeter. So you could also write it this way:
I remember being so blown away by OS2/Warp's ability to multi-task so many applications at once, with such a clean UI.
As an older programmer, let me suggest one ought to be reticent about saying things like that. I know that by any reasonable standard it should make you sound experienced and therefore worth listening to, but if you have any gray in your hair it's bound to have a very different effect. Like the time I sat next to a guy at a banquet who was reminiscing about when his department got an IBM 701. "Yep," he said with evident satisfaction, "that was a stored program jobbie."
Employers are looking for programmers who were in diapers while you were being blown away by OS/2, so ixnay on that kind of alktay. Instead practice saying things like "Node.js is so 2015." And when someone asks you what you mean, turn to them, raise one eyebrow, then literally turn your back on them.
Your dream is to retire with an income of $10,568.40 a year?
Well, it's not 800 Euros a month hot off the printing press; it's 800 Euros that were taxed out of the economy then put back into the economy in a different place. It'll surely effect the prices of many things, but net there's no more total money in the economy.
I suspect the thinking is that many of the things that people on the lower income end of the spectrum have relatively inelastic demand: housing and foodstuffs. Things that are discretionary purchases for those people are bound to become more expensive.
Sure, but what matters is the performance of *systems*; and if you set aside battery one one hand and gas tank on the other, an electric powertrain is much, much lighter than a gasoline powertrain. If you got a 10x increase in energy to weight in battery technology, that'd come right off what is by far the electric car's heaviest component. The result would be dramatic improvements in the car's performance and efficiency.
If it also comes with the kind of cost reductions they're imagining, you could be talking about a car about the size and cost of a Nissan Leaf but with a 500 mile range and about 10% lighter.
However I'm not holding my breath. The air breathing battery technology idea isn't new; the problem is longevity. These guys have some ideas for improving that, but it remains to be seen whether those'll lead to a practical vehicle battery with the kind of performance they're hoping for.
More mainstream than elisp, though.
I think the selling point is extensibility. I've been playing around with it and I detect a great deal of Emacs influence in the editor's features. I think the idea is to provide emacs-like power user features with a basic text editor-like entry level experience.
I think the primary reason is that they want the editor to be extensible using javascript, which of course is possible with a browser based app but then the people doing the customization have to work around browser security policies like single origin which complicate scripting for reasons that don't apply to a native editor.
It's a text editor, built with web-based technologies and packaged as an app-specific browser that looks like a native desktop app, which offers a user experience similar to most desktop programmer's editors but with emacs-like extensibility (only in javascript rather than elisp).
I've been the Hinton Lab in Jamaica Plain dozens of times. The people I worked with there from the lowest lab tech to the middle managers were outstanding, but they were in the epidemiology end of things. The drug testing lab was segregated on its own floor, and it was walled off like a fortress. But despite that superficial formidability of the drug testing lab, there was clearly a problem: back in 2007 the director of the crime lab resigned because mishandling DNA tests, and before that the lab had been in trouble for processing DNA too slowly. There were rape kits that had been waiting to be processed for eighteen years.
Yet despite the review of the crime lab's procedures that followed this scandal, Annie Dookhan was able to continue with her antics for an other four years before she was caught. It's odd that she was even hired with her phony degrees because that was the year it came out that Ralph Timperi, the Hinton Lab's overall director, got his PhD from a diploma mill. You'd think that'd trigger a little more scrutiny.
It all makes the entire Hinton Lab sound like a hot mess, but with the exception of Timperi's phony degree all the problems were in the crime lab, which while located inside the Hinton Lab building was (IIRC) actually overseen by the Massachusetts State Police. Possibly some kind of responsibility thing was going on there. On the public health side of things the people at the state labs were among of the best public employees I've ever dealt with, and I've worked with state and county agencies across the country. It's a shame they've been tainted.
What would be doomed is his vision of what NASA should be doing ten to twenty years out. At present manned exploration of Mars is not even big enough to warrant a line item in NASA's 2016 budget request, which means it's not a big part of what NASA is doing now. Mars exploration *is*, though, with 412 million requested to do serious and productive science.
You know what *would* doom NASA? Gutting its Earth and planetary research to fund an astronomically expensive manned mission to Mars in twenty years. I think an excuse for gutting NASA's Earth science program is part of the political imperative for a manned Mars mission.
I'm not against a manned mission to Mars per se; I'm against the opportunity cost. For the price of keeping a human crew alive and healthy for six years in bathed in deep space radiation and setting them down on the surface of Mars with an acceptable safety margin (very hard to do), we could get so much more done with robots. Extrapolating from Apollo to a manned Mars mission is very risky; Mars is immensely harder than the Moon in almost every respect. I don't think America is ready to spend Apollo program dollars every year for two or three decades, so what we'll get is a half-assed program that starves everything else NASA ought to be doing with very dubious prospects of success.
Eventually the time will come when the marginal value of a human mission will exceed the marginal value of the additional robotic missions the same funding would secure. That will be the time to start planning a manned Mars mission. I expect we might even get humans there faster that way.
Not really. First of all he wasn't an electrician, he was something called a cable jointer; it doesn't require a license and normally they don't work on live stuff. Secondly, according to TFA he wasn't aware the line was live, which means someone in a supervisory position screwed up.
As to whose fault it is, that depends. If the company in charge of the site has a proper lock out/tag out policy and training program, it's probably some supervisor's fault -- unless the worker went into an area specifically against instructions in which case it's the worker's fault. If the company doesn't have a proper LOTO program (including training for all personnel on site) it's the company's fault.
LOTO may be a PITA, but it's not rocket science.
Right, because when you change one thing, everything else has to stay the same, otherwise things would get to complicated to discuss in a post like yours.
And somehow the other part of macro (saving during booms) never gets the same airing.
Funny that.
I actually think is an important point. I've often felt that the best time for a government to tighten its belt is when there is full private sector employment. That's because when you tax a dollar out of the private sector at that point you're doing at a time when the private sector is using dollars as many dollars as it can get its hands on to create jobs. It doesn't happen because that's a time a politician can look like a hero without inflicting sacrifice on anyone.
Of course there are other reasons to spend besides macroeconomics of course. The reason you fund a meals-on-wheels for elderly shut-ins isn't because of the economic stimulus effect. Same goes for defense projects -- although clearly the most advantageous time to cancel a major program or to close bases would be during a full employment period.
I think Elizabeth Warren and Ed Markey have already got the memo, seeing as they were the ones that originally sent it.
My problem with these discussions is that people on both sides talk about "spending" as if all spending is fungible. In my experience it's not. Some things are smart to spend money on, other things are stupid to spend money on; and also it's very easy to spend money stupidly on things that are unquestionably necessary. It's one thing for everyone to agree that, say, defense is an important thing to spend money on; and yet another thing to assume that all defense spending is good.
I think it's quite possible for a country spend its way to solvency -- if it spent overwhelmingly wisely. It's also possible to economize your way to bankruptcy, if you cut overwhelmingly foolishly. Of course it's also possible to economize wisely or spend yourself into a crisis. I've seen every possible combination, both in the private sector and the public sector. Private enterprises can become uncompetitive when management is unwilling to spend the money it needs to compete.