Isn't that exactly the type of wasteful behavior which attributes to higher costs? If for instance classrooms were at 50% utilization for two hours between 8-5, just because everyone is doing meetings at the same time, you could reduce the number of classrooms by 10% if you simply spread meetings throughout the day.
It doesn't work that way. The reality is that students are used to being in school from about 8 to 3. They tend to resist taking classes much past that time, and by college, they tend to resist taking classes before 10 as well. Realistically, you get about five good hours during which you can teach classes, and the more classes you schedule outside those core hours, the more students will cram into the classes within those hours, so you just end up with very imbalanced sections that make it harder to teach.
And it isn't just momentum, either. Lots of students commute to their university, which means early and late classes don't work. Parents (both college students and faculty) have to pick their kids up from school. Students have part-time jobs to pay the bills. And so on.
Finally, it isn't practical to just say, "We're going to spread classes evenly throughout the day", because students need time to actually work on their homework. And that time needs to be during the day so that they can use campus facilities such as computer labs, tutoring centers, etc. It simply isn't practical for the entire day to be used for instruction, because it costs money to operate those other facilities, too, and you'd end up having to cover the cost of extending their hours dramatically if you extend the core hours for classes, which means significantly increased staffing, which ends up costing more over the long run than adding one or two extra rooms to a building.
Although true, I would argue that what's really needed are standard, third-party cab hailing apps that know about all the cab companies in an area and find you a cab, rather than having to have an app for each cab company in each locality where you might need a cab. It isn't really reasonable to expect each cab company to solve the problem themselves, and it can be tricky for competitors to work together.
I worked as I traveling consultant for 10 years, 80 to 100 flight segments per year, in major cities across the US, with the accompanying cab/uber rides to go with them, and I can unequivocally say that taxi/limo service before Uber was terrible. It was caused by cities artificially limiting supply/bullshit regulation/catering to special interests, all of which Uber/Lyft/etc need to continue to kill, for the good of all.
Taxi services are terrible because it is hopelessly expensive to drive a vehicle point-to-point and the amount of money that the government allows them to charge is not enough to actually pay for repairs and improvements to the vehicles. Uber only "works" because:
They've externalized the vehicle costs by forcing the drivers to pay them without allowing the drivers to actually charge fees high enough to cover those costs.
There are plenty of people who haven't figured out how much money they're going to end up spending on vehicle maintenance as a result of all that extra driving.
They're taking advantage of huge subsidies and burning through their cash reserves to dump their services on the market even further below cost.
The unionization threats are happening because a large enough percentage of the drivers are recognizing Uber for the complete scam that it is. By many estimates, the minimum price at which Uber will be profitable while providing the current level of service is about 4x their current prices. That makes taxis look downright cheap. Increased competition can't ever reduce the cost below a floor set by certain unavoidable costs for things like gasoline, brakes, etc. Well, I guess technically you could have a taxi service with no brakes, but I wouldn't recommend it.
My basic land line at home, with caller ID as the only service, costs... I think $50 per month. Prepaid cellular plans without data start at $3 per month. I could get a separate cellular phone on a separate prepaid plan in every room in the house that currently has a land-line phone, and it would still cost less than a third of what my land line costs.
At this point, just about the only people that have them are businesses and the elderly—the former because it's easier to manage assets that don't move around, and the latter because picking up a phone is natural enough that people with degenerative diseases are likely to remember how to do it, whereas hitting an answer button isn't.
I mentioned to her "You do realize that pan & scan hides about half the scene from your view, right?" She didn't want to hear it.
Of course, that's not necessarily true. Some P&S movies were originally shot on a larger frame, but with physical mattes on the eyepiece so the camera operator could see what would actually show up on a widescreen setup. In those cases, sometimes you got more content in the P&S version of some scenes, because they went back to the original negatives. And in other cases, they shot two different versions of certain shots, one for P&S and one for widescreen, such as doing a two shot on widescreen, alternating between two over-the-shoulder shots for narrowscreen. Usually P&S was just a cropped version of the full movie, but in some cases, it was an entirely different visual experience. And that's what they're talking about doing here, I think.
Unless I'm missing something, the summary omits some pretty important details and in so doing, kind of misses the point.
The protein in question is believed to be one cause of frontotemporal dementia (and possibly ALS). If I understand the UniProtKB page (skimming), the gene codes for a protein that is supposed to degrade over time, but in some people, it doesn't, and as a result, it builds up in brain cells. This, coupled with other genes that don't suppress production sufficiently, is believed to speed the deterioration of spindle neurons in the frontal and/or temporal lobes.
More to the point, this is almost certainly unrelated to Alzheimer's, as spindle neurons are typically unaffected by that disease except perhaps in the final stages of the disease. I have no idea why that disease was mentioned in the summary, except that it is a common brain disorder that everybody would like to see cured.
On the other hand, other forms of dementia are often misdiagnosed as Alzheimer's disease, so there's that....
... Even if someone can review 1,000 pictures a day facebook would have to hire 300,000 people to ensure none of the pictures posted are 'kiddie porn'
And computer scientists know there is no automated way to screen these photos without generating false positives. Even an algorithm that was 99% accurate, would mean 1% of 300 million pictures, or 3 MILLION pictures would get falsely reported as child pornography and taken down every single day. And let me tell you, our image recognition algorithms are nowhere near 99% accurate.
That's why you do a multi-tier system, with AI doing the first couple of passes, humans doing a quick glance as the third pass, and serious scrutiny only for things that make it through the first three passes. You configure the algorithms to err on the side of detection. If that results in a 5% hit rate instead of a fraction of a percent, then you have maybe 15 million pictures for a human to review and decide whether to take them down. Then you remove all the pictures that are duplicates or near duplicates of photos that have been posted before (e.g. ignoring the text part of image macros), and you probably have closer to 1–2 million pictures for a human to review. At that point, if a person can review 1,000 in a day, you now need 1,000 reviewers.
Also, I suspect that 1,000 photos per person per day is a low estimate. Most of the time you can glance at it for a quarter of a second and tap the "It's safe" button. Some will take longer, obviously. I'd expect it to average only a couple of seconds unless something gets flagged for secondary review. That's on the order of 10–15,000 photos per day per reviewer, so we're probably talking about a hundred reviewers, plus a half dozen tier 2 reviewers that do secondary screening for stuff flagged by the front-line screeners.
That will be very, very interesting given that in some states, for them to do the paper temp license they need you to present photo ID.
I assume that's only for getting a paper temp license after surrendering a license from another state. If you're in your own state, they should have your photo in the system already.
That said, if you're traveling, it's a good idea to keep your passport separate from your license, and keep an old, expired passport separate from that.
Mind you if someone is incapable of adjusting their sleep schedule so they don't lose an hour of sleep, let them suffer.
Most people are incapable of adjusting their sleep schedules at the drop of a hat. What I find is that about a week before the time change used to be, my body clock starts shifting. This means not only does the actual time change cause problems, but also the lack of the old time change. And instead of gaining an hour on the other end, somehow my body clock shifts the start time so that I'm getting an hour less sleep for the week prior to the time change, and then the time changes and I still wake up at the same time for another week and can't get back to sleep even though it is an hour earlier.
It has also been proven that at least at the high school level, students do better when they start school later. Why, in a post-agrarian society, are we still starting school at 8:00 (or worse, 7:00 for some classes)? Just start school at 9:00 instead of 8:00 and end it at 3:30 instead of 2:30, and the problem basically goes away.
Most? I think it's not most are terrible. Some are terrible, some are bad, some are so-so, some are pretty fair, some are good and some are excellent.
... at any given moment. The problem is that every driver is terrible at least some of the time. People get distracted (both externally and internally), fatigued, can look in only one direction at a time, can't see in certain directions from the driver's seat, etc., all of which can effectively mimic impaired reaction time just as easily as drunk driving, albeit for a shorter period of time. Fortunately, most of the time, you don't need fast reaction time to drive safely.
At the time you need to find the random answers to these questions is when you need to know your password - if you have to look up the answers you may as well look up the password! The times when the questions are very useful is when you're not at home, you're stranded in a strange city, your wallet was stolen, etc.
... when you're stealing the identity of the guy whose wallet you just stole in a strange city...
Security questions are fundamentally useless. If asked in addition to your password, the only thing they do is give you one more thing to write down so that you remember it. If they are asked as an alternative to your password, the only thing they do is effectively lower the complexity of an attack on the password by a factor of at least [n] where [n] is the number of security questions + 1, and that's the best-case scenario where you use random values.
The worst-case scenario, where you answer the question truthfully, is that if they are asked in addition to your password, they just increase the odds of getting accidentally locked out of your account by making a typo, and if asked as an alternative, they lower the complexity of an attack to the amount of time required to look up your public profile on Facebook and copy and paste the requested information.
To recover access to a bank account, bring your government-issued photo ID in with you and request a new password. Any other approach does more harm than good. If your photo ID got stolen, too, go to the DMV and request a reissue. They'll give you a paper temp license on the spot, which you can bring with you to the bank, and they'll mail you a permanent replacement.
It's also the primary argument for having a tightly curated, closed ecosystem App Store model like Apple has.
No, not really. You're conflating "someone" with "everyone" here. The closed ecosystem provides a benefit because the odds are high that someone will do something bad. Bans on hot-patching provides a benefit only if you assume that everyone will do something bad. This difference is subtle, but critical.
Apple has a number of protections to prevent malicious apps from causing harm—blacklisting an app so that it won't even launch, removing the app from the store, banning the developer from submitting new apps, etc., all of which are made possible by that closed ecosystem. These allow Apple to provide oversight that prevents bad people from doing bad things, and are necessary because it isn't absurd to believe that some people will try to do so. What's absurd is assuming that all developers (or even a large percentage of developers) will risk destroying their reputation and livelihood to do bad things unless Apple nit-picks every single submission into the ground.
More to the point, curation is about minimizing the risk of getting complete junk apps, not about preventing bad people from doing bad things in app that only become visible after the fact. There's nothing Apple can realistically do in an app review that could detect malicious code, because it is entirely trivial for an app to ask a server what to do and then either behave normally or maliciously depending on the response. That behavior could be hard-coded into an app, and Apple would never realistically be able to detect it. The only way you could prevent a malicious developer from doing that would be to ban apps that make Internet requests. Thus, banning hot patching cannot possibly have any effect on whether malicious developers can create apps that cause harm, because it isn't necessary for apps to use hot patching to cause harm. It isn't really even all that helpful.
What hot patching can do is allow developers to flagrantly ignore app store policies. And it makes sense for Apple to crack down on developers who use it in that way, in much the same way that it makes sense for police to arrest people who commit crimes. It doesn't make sense to ban the technology under the theory that if Apple doesn't crack down constantly, the developers will all run amok, in much the same way that it doesn't make sense for police to arrest everyone because a few people might commit crimes.
How is that false equivalence? The original argument is that the only thing preventing everybody from doing bad things is oversight by other people. That's prima facie absurd. Most people will do the right thing even without being watched by other members of their peer group, police, etc. In much the same way, most developers will not abuse the ability to hot patch their code merely because they have that ability.
"Could" is the operative word. By this same logic, every adult male has the tools required to commit rape. Therefore, any time a single adult male is out on the streets after dark without being part of a large group, we must incarcerate him to ensure that he doesn't use those tools for that purpose. See how silly that reasoning sounds when shifted into a different problem space?
Rationally, Apple should penalize developers that abuse this capability to deliberately ship features that they know would be rejected by Apple, such as using private SPIs, egregiously violating app store rules, etc. Anything short of that line is probably a reasonable thing to do (e.g. patching functions to fix a serious crash more quickly). Punish the app when the developer uses the feature in a way that harms users, and not before.
Who's going to pay for that? Google? All of their money is tied up in offshore accounts.
You speak of this as though I'm talking about a huge cost. We're talking about twenty or thirty feet of asphalt, some new road striping, a handful of temporary concrete barriers to close off some roads, and minor traffic light changes. The average Google employee could pay for this.
BTW, I just noticed that I left half a sentence in my original post. Where I sad "Use traffic cones and", that was an aborted attempt at writing the last bullet point about adding safety cones to turn those roads temporarily back into two-way streets to allow more inbound traffic on event days if it proves necessary to do so.
Not at all. The Bay Area isn't even remotely overpopulated compared with other cities. In fact, the most dense city in the Bay Area, San Francisco, isn't even in the top 20 cities in the U.S. density-wise. It actually ranks below a suburb of Louisville, KY. The population density is basically a non-issue. The problem is not the number of people, but rather how far they are driving, how often they are driving, and how efficiently they use the roads. Specifically:
We have terrible public transit because the density is too low on average.
Ill-conceived property tax laws in the 1970s make it too expensive to move closer to your job, resulting in some of the longest average commutes in non-rural America.
Businesses are not sufficiently encouraged through tax incentives to spread out beyond the peninsula, resulting in businesses locating in areas where the vast majority of workers can't afford to actually live.
Traffic lights are deliberately programmed to slow the flow of traffic to encourage cars to use the freeway, effectively wasting the vast majority of our road capacity.
Poorly designed lanes force unnecessary merges.
Sudden lane shifts cause traffic to enter other lanes accidentally, causing a lot of wrecks.
Inadequate "business" routes fail to provide high-volume secondary highways that parallel the major freeways.
I'm sure there are many other issues that contribute, but these are the problems that I've noticed causing the largest number of wrecks and significant backups, at least in the South Bay.
Jumping on and off the freeway, only taking up the right lane, is a very minimal impact on traffic.
You couldn't be more wrong. In practice, there is effectively a maximum difference of just a few miles per hour between adjacent lanes, dictated by the way traffic flows. Every time a vehicle in the next lane over needs to exit, it has to merge into that right lane. If the right lane is moving more slowly, that merge slows down the second lane for a good mile or more behind that vehicle, bringing it down to just faster than the rightmost lane. Then this repeats for the next lane over, and so on.
Incidentally, this is why every study that has ever taken a serious look at carpool lanes has concluded that they are almost completely useless unless they have their own dedicated exits on the opposite side of the road. I can count the number of times that I've seen the carpool lane going more than about 5 MPH faster than the next lane over on one hand in almost two decades of driving in the Bay Area. As long as those cars have to exit through a glorified parking lot, how can they possibly move significantly faster than the parking lot? Everything affects everything. And in aggregate, all those cars that get onto the highway when they could have gotten there in the same time on a city street with better traffic light timing cause a significant impact on traffic.
My experience (as someone originally from a rural area) was that RS's selection wasn't significantly better than Walmart, and its prices were generally significantly higher.
The real problem is that the freeways are designed to be "hurry up and wait" systems. Designers bring in thousands and thousands of cars on 3 lanes at 60 MPH, then ask them all to take a single lane exit at 45MPH, then they scratch their heads and wonder why things go to shit.
A single exit lane with a traffic light at the top. Approximately 90% of the problems on 101 are caused by the city streets near Google not accommodating the flow of incoming traffic, causing major backups onto the highway. Want to fix it? Do this:
Close the La Avenida entrance onto North Shoreline.
Connect the right lanes of the exit ramp at North Shoreline directly to La Avenida with no traffic light.
Make North Shoreline be one-way southbound from the intersection at Amphitheatre Parkway to 101. Use traffic cones and
Make Amphitheater Parkway be one-way northbound from the intersection at North Shoreline to 101.
Make Charleston Road be one one-way westbound from North Shoreline to Amphitheatre Parkway.
Convert the traffic light at Charleston and Amphitheatre Parkway to a series of green and flashing yellow turn arrows.
Convert the traffic light at Charleston and North Shoreline to a series of green and flashing yellow turn arrows.
Tear down walls and make it possible to cut through parking lots to get from from Pear Ave. to Space Park Way, Shorebird Way, and Charleston Rd.
On days with events at Shoreline, temporarily restore two-way traffic flow on both roads with safety cones. Use La Avenida and Inigo Way to route traffic from the freeway onto North Shoreline.
And with only a little bit of minor roadwork and traffic light reprogramming, you've just changed 101 from a horrible nightmare to a free-flowing road from roughly Menlo park through Mountain View on the southbound side in the evening and from Santa Clara to Mountain View on the northbound side in the morning... and all without actually touching 101.
I would argue that to some degree, this is a normal side effect of living on a street that parallels a major highway, and to the extent that it isn't, it is actually caused by attempts to keep people from using secondary streets. One of the biggest reasons that our highways are such a nightmare in the Bay Area is that people use them too much. People hop on the freeway to go one or two exits when they could just as easily take city streets, largely because of broken traffic light timing that discourages people from using those streets instead. The result is that the freeway system becomes completely clogged by people entering and exiting unnecessarily often, and everybody suffers.
Ironically, the best way to fix this problem is to rip out all of the traffic "calming" silliness, program the traffic lights on major city streets to maximize the flow of traffic in the direction that parallels any traffic backup, and encourage people to use those larger city streets for their local commutes instead of getting on the freeway in the first place. The more cars you get using (at least) major city streets, the more the highways will improve. Just reducing the number of cars entering a highway by a small margin can have a big impact on the speed of that highway.
Well, that's not entirely true. The ADA requires that content be accessible if students request it, too. The difference is that students who request it can get assistance with the non-accessibility-friendly content on a one-off basis in exchange for their tuition. It isn't that it is too expensive to do the captioning, but rather that it is too expensive to do it without compensation.
I had the same problem while trying to call in a wreck on California's SR 17. I gave up trying to call 911 when a volunteer firefighter happened to come upon the scene and after verifying that everybody was okay, called it in on his radio. If you can't rely on critical safety systems to actually work in a real emergency, then what's the point of even having them? From an outsider's perspective, our 911 system appears to be a train wreck and should probably be scrapped outright and replaced with something entirely different.
Slashdot Mirror
It doesn't work that way. The reality is that students are used to being in school from about 8 to 3. They tend to resist taking classes much past that time, and by college, they tend to resist taking classes before 10 as well. Realistically, you get about five good hours during which you can teach classes, and the more classes you schedule outside those core hours, the more students will cram into the classes within those hours, so you just end up with very imbalanced sections that make it harder to teach.
And it isn't just momentum, either. Lots of students commute to their university, which means early and late classes don't work. Parents (both college students and faculty) have to pick their kids up from school. Students have part-time jobs to pay the bills. And so on.
Finally, it isn't practical to just say, "We're going to spread classes evenly throughout the day", because students need time to actually work on their homework. And that time needs to be during the day so that they can use campus facilities such as computer labs, tutoring centers, etc. It simply isn't practical for the entire day to be used for instruction, because it costs money to operate those other facilities, too, and you'd end up having to cover the cost of extending their hours dramatically if you extend the core hours for classes, which means significantly increased staffing, which ends up costing more over the long run than adding one or two extra rooms to a building.
Although true, I would argue that what's really needed are standard, third-party cab hailing apps that know about all the cab companies in an area and find you a cab, rather than having to have an app for each cab company in each locality where you might need a cab. It isn't really reasonable to expect each cab company to solve the problem themselves, and it can be tricky for competitors to work together.
Actually, the state should not set a statewide maximum, but rather a minimum maximum.
Taxi services are terrible because it is hopelessly expensive to drive a vehicle point-to-point and the amount of money that the government allows them to charge is not enough to actually pay for repairs and improvements to the vehicles. Uber only "works" because:
The unionization threats are happening because a large enough percentage of the drivers are recognizing Uber for the complete scam that it is. By many estimates, the minimum price at which Uber will be profitable while providing the current level of service is about 4x their current prices. That makes taxis look downright cheap. Increased competition can't ever reduce the cost below a floor set by certain unavoidable costs for things like gasoline, brakes, etc. Well, I guess technically you could have a taxi service with no brakes, but I wouldn't recommend it.
My basic land line at home, with caller ID as the only service, costs... I think $50 per month. Prepaid cellular plans without data start at $3 per month. I could get a separate cellular phone on a separate prepaid plan in every room in the house that currently has a land-line phone, and it would still cost less than a third of what my land line costs.
At this point, just about the only people that have them are businesses and the elderly—the former because it's easier to manage assets that don't move around, and the latter because picking up a phone is natural enough that people with degenerative diseases are likely to remember how to do it, whereas hitting an answer button isn't.
Of course, that's not necessarily true. Some P&S movies were originally shot on a larger frame, but with physical mattes on the eyepiece so the camera operator could see what would actually show up on a widescreen setup. In those cases, sometimes you got more content in the P&S version of some scenes, because they went back to the original negatives. And in other cases, they shot two different versions of certain shots, one for P&S and one for widescreen, such as doing a two shot on widescreen, alternating between two over-the-shoulder shots for narrowscreen. Usually P&S was just a cropped version of the full movie, but in some cases, it was an entirely different visual experience. And that's what they're talking about doing here, I think.
Unless I'm missing something, the summary omits some pretty important details and in so doing, kind of misses the point.
The protein in question is believed to be one cause of frontotemporal dementia (and possibly ALS). If I understand the UniProtKB page (skimming), the gene codes for a protein that is supposed to degrade over time, but in some people, it doesn't, and as a result, it builds up in brain cells. This, coupled with other genes that don't suppress production sufficiently, is believed to speed the deterioration of spindle neurons in the frontal and/or temporal lobes.
More to the point, this is almost certainly unrelated to Alzheimer's, as spindle neurons are typically unaffected by that disease except perhaps in the final stages of the disease. I have no idea why that disease was mentioned in the summary, except that it is a common brain disorder that everybody would like to see cured.
On the other hand, other forms of dementia are often misdiagnosed as Alzheimer's disease, so there's that....
Reminds me of these:
Bus on a bus
Early camper truck?
That's why you do a multi-tier system, with AI doing the first couple of passes, humans doing a quick glance as the third pass, and serious scrutiny only for things that make it through the first three passes. You configure the algorithms to err on the side of detection. If that results in a 5% hit rate instead of a fraction of a percent, then you have maybe 15 million pictures for a human to review and decide whether to take them down. Then you remove all the pictures that are duplicates or near duplicates of photos that have been posted before (e.g. ignoring the text part of image macros), and you probably have closer to 1–2 million pictures for a human to review. At that point, if a person can review 1,000 in a day, you now need 1,000 reviewers.
Also, I suspect that 1,000 photos per person per day is a low estimate. Most of the time you can glance at it for a quarter of a second and tap the "It's safe" button. Some will take longer, obviously. I'd expect it to average only a couple of seconds unless something gets flagged for secondary review. That's on the order of 10–15,000 photos per day per reviewer, so we're probably talking about a hundred reviewers, plus a half dozen tier 2 reviewers that do secondary screening for stuff flagged by the front-line screeners.
Now for videos... that's another problem.
I assume that's only for getting a paper temp license after surrendering a license from another state. If you're in your own state, they should have your photo in the system already.
That said, if you're traveling, it's a good idea to keep your passport separate from your license, and keep an old, expired passport separate from that.
Most people are incapable of adjusting their sleep schedules at the drop of a hat. What I find is that about a week before the time change used to be, my body clock starts shifting. This means not only does the actual time change cause problems, but also the lack of the old time change. And instead of gaining an hour on the other end, somehow my body clock shifts the start time so that I'm getting an hour less sleep for the week prior to the time change, and then the time changes and I still wake up at the same time for another week and can't get back to sleep even though it is an hour earlier.
It has also been proven that at least at the high school level, students do better when they start school later. Why, in a post-agrarian society, are we still starting school at 8:00 (or worse, 7:00 for some classes)? Just start school at 9:00 instead of 8:00 and end it at 3:30 instead of 2:30, and the problem basically goes away.
... at any given moment. The problem is that every driver is terrible at least some of the time. People get distracted (both externally and internally), fatigued, can look in only one direction at a time, can't see in certain directions from the driver's seat, etc., all of which can effectively mimic impaired reaction time just as easily as drunk driving, albeit for a shorter period of time. Fortunately, most of the time, you don't need fast reaction time to drive safely.
... when you're stealing the identity of the guy whose wallet you just stole in a strange city...
Security questions are fundamentally useless. If asked in addition to your password, the only thing they do is give you one more thing to write down so that you remember it. If they are asked as an alternative to your password, the only thing they do is effectively lower the complexity of an attack on the password by a factor of at least [n] where [n] is the number of security questions + 1, and that's the best-case scenario where you use random values.
The worst-case scenario, where you answer the question truthfully, is that if they are asked in addition to your password, they just increase the odds of getting accidentally locked out of your account by making a typo, and if asked as an alternative, they lower the complexity of an attack to the amount of time required to look up your public profile on Facebook and copy and paste the requested information.
To recover access to a bank account, bring your government-issued photo ID in with you and request a new password. Any other approach does more harm than good. If your photo ID got stolen, too, go to the DMV and request a reissue. They'll give you a paper temp license on the spot, which you can bring with you to the bank, and they'll mail you a permanent replacement.
No, not really. You're conflating "someone" with "everyone" here. The closed ecosystem provides a benefit because the odds are high that someone will do something bad. Bans on hot-patching provides a benefit only if you assume that everyone will do something bad. This difference is subtle, but critical.
Apple has a number of protections to prevent malicious apps from causing harm—blacklisting an app so that it won't even launch, removing the app from the store, banning the developer from submitting new apps, etc., all of which are made possible by that closed ecosystem. These allow Apple to provide oversight that prevents bad people from doing bad things, and are necessary because it isn't absurd to believe that some people will try to do so. What's absurd is assuming that all developers (or even a large percentage of developers) will risk destroying their reputation and livelihood to do bad things unless Apple nit-picks every single submission into the ground.
More to the point, curation is about minimizing the risk of getting complete junk apps, not about preventing bad people from doing bad things in app that only become visible after the fact. There's nothing Apple can realistically do in an app review that could detect malicious code, because it is entirely trivial for an app to ask a server what to do and then either behave normally or maliciously depending on the response. That behavior could be hard-coded into an app, and Apple would never realistically be able to detect it. The only way you could prevent a malicious developer from doing that would be to ban apps that make Internet requests. Thus, banning hot patching cannot possibly have any effect on whether malicious developers can create apps that cause harm, because it isn't necessary for apps to use hot patching to cause harm. It isn't really even all that helpful.
What hot patching can do is allow developers to flagrantly ignore app store policies. And it makes sense for Apple to crack down on developers who use it in that way, in much the same way that it makes sense for police to arrest people who commit crimes. It doesn't make sense to ban the technology under the theory that if Apple doesn't crack down constantly, the developers will all run amok, in much the same way that it doesn't make sense for police to arrest everyone because a few people might commit crimes.
How is that false equivalence? The original argument is that the only thing preventing everybody from doing bad things is oversight by other people. That's prima facie absurd. Most people will do the right thing even without being watched by other members of their peer group, police, etc. In much the same way, most developers will not abuse the ability to hot patch their code merely because they have that ability.
"Could" is the operative word. By this same logic, every adult male has the tools required to commit rape. Therefore, any time a single adult male is out on the streets after dark without being part of a large group, we must incarcerate him to ensure that he doesn't use those tools for that purpose. See how silly that reasoning sounds when shifted into a different problem space?
Rationally, Apple should penalize developers that abuse this capability to deliberately ship features that they know would be rejected by Apple, such as using private SPIs, egregiously violating app store rules, etc. Anything short of that line is probably a reasonable thing to do (e.g. patching functions to fix a serious crash more quickly). Punish the app when the developer uses the feature in a way that harms users, and not before.
You speak of this as though I'm talking about a huge cost. We're talking about twenty or thirty feet of asphalt, some new road striping, a handful of temporary concrete barriers to close off some roads, and minor traffic light changes. The average Google employee could pay for this.
BTW, I just noticed that I left half a sentence in my original post. Where I sad "Use traffic cones and", that was an aborted attempt at writing the last bullet point about adding safety cones to turn those roads temporarily back into two-way streets to allow more inbound traffic on event days if it proves necessary to do so.
Not at all. The Bay Area isn't even remotely overpopulated compared with other cities. In fact, the most dense city in the Bay Area, San Francisco, isn't even in the top 20 cities in the U.S. density-wise. It actually ranks below a suburb of Louisville, KY. The population density is basically a non-issue. The problem is not the number of people, but rather how far they are driving, how often they are driving, and how efficiently they use the roads. Specifically:
I'm sure there are many other issues that contribute, but these are the problems that I've noticed causing the largest number of wrecks and significant backups, at least in the South Bay.
You couldn't be more wrong. In practice, there is effectively a maximum difference of just a few miles per hour between adjacent lanes, dictated by the way traffic flows. Every time a vehicle in the next lane over needs to exit, it has to merge into that right lane. If the right lane is moving more slowly, that merge slows down the second lane for a good mile or more behind that vehicle, bringing it down to just faster than the rightmost lane. Then this repeats for the next lane over, and so on.
Incidentally, this is why every study that has ever taken a serious look at carpool lanes has concluded that they are almost completely useless unless they have their own dedicated exits on the opposite side of the road. I can count the number of times that I've seen the carpool lane going more than about 5 MPH faster than the next lane over on one hand in almost two decades of driving in the Bay Area. As long as those cars have to exit through a glorified parking lot, how can they possibly move significantly faster than the parking lot? Everything affects everything. And in aggregate, all those cars that get onto the highway when they could have gotten there in the same time on a city street with better traffic light timing cause a significant impact on traffic.
My experience (as someone originally from a rural area) was that RS's selection wasn't significantly better than Walmart, and its prices were generally significantly higher.
A single exit lane with a traffic light at the top. Approximately 90% of the problems on 101 are caused by the city streets near Google not accommodating the flow of incoming traffic, causing major backups onto the highway. Want to fix it? Do this:
And with only a little bit of minor roadwork and traffic light reprogramming, you've just changed 101 from a horrible nightmare to a free-flowing road from roughly Menlo park through Mountain View on the southbound side in the evening and from Santa Clara to Mountain View on the northbound side in the morning... and all without actually touching 101.
I would argue that to some degree, this is a normal side effect of living on a street that parallels a major highway, and to the extent that it isn't, it is actually caused by attempts to keep people from using secondary streets. One of the biggest reasons that our highways are such a nightmare in the Bay Area is that people use them too much. People hop on the freeway to go one or two exits when they could just as easily take city streets, largely because of broken traffic light timing that discourages people from using those streets instead. The result is that the freeway system becomes completely clogged by people entering and exiting unnecessarily often, and everybody suffers.
Ironically, the best way to fix this problem is to rip out all of the traffic "calming" silliness, program the traffic lights on major city streets to maximize the flow of traffic in the direction that parallels any traffic backup, and encourage people to use those larger city streets for their local commutes instead of getting on the freeway in the first place. The more cars you get using (at least) major city streets, the more the highways will improve. Just reducing the number of cars entering a highway by a small margin can have a big impact on the speed of that highway.
I'm sure there are many things I've forgotten. Most generalizations are like that. :-)
Well, that's not entirely true. The ADA requires that content be accessible if students request it, too. The difference is that students who request it can get assistance with the non-accessibility-friendly content on a one-off basis in exchange for their tuition. It isn't that it is too expensive to do the captioning, but rather that it is too expensive to do it without compensation.
I had the same problem while trying to call in a wreck on California's SR 17. I gave up trying to call 911 when a volunteer firefighter happened to come upon the scene and after verifying that everybody was okay, called it in on his radio. If you can't rely on critical safety systems to actually work in a real emergency, then what's the point of even having them? From an outsider's perspective, our 911 system appears to be a train wreck and should probably be scrapped outright and replaced with something entirely different.