Deploy a pico-cell that only carries 911 calls and no data. In terms of an escalation technical arms race there are plenty of things that can be done. I sure as heck hope this backfires and people simply walk out of the store to check the price and (maybe) walk back in. The downside to the retailer is that they may have the best price or at least one that's good enough that people would be a premium for instant gratification but since they are now already outside of the store, they may decide it's not worth going back in.
Well, to be devil's advocate, if I were implementing this inside of my store, you can be pretty sure that I'd make sure that no 4G signals reached inside so my WiFi would be your only choice. Not that I'm advocating for this, but there's no point of deploying a solution when there's a trivial workaround.
Almost every IDE can format code. And *good* IDEs will do it perfectly. Unfortunately what they don't do is preserve the original formatting mechanism for future checkins. There's a menu item in Eclipse but I don't know what it is. The key combination is Shift-Ctrl-F to reformat the code. For Java especially, it's super-easy for Eclipse to do this since Eclipse knows what the parse tree looks like. Some formatting utilities can be very weak in they rely on things like regexes and counting brackets which could be error prone. But if you format code by first parsing into the AST and then regenerating the code from that, you'll get an accurate formatting every time.
The answer is that the IDE should allow you to view the code with your own preferences, but what goes into revision control should get automatically formatted with whatever standard is set for a project.
What is shocking to me is that we even have this discussion. There ought to be a check-in hook that properly formats the code. Would also reduce defects where a brace/bracket is missing.
No but every *person* may need at least 1 IP address that that's will 6B people and an address range of 2B. When IPv4 was designed, there was no NAT so it was easy to anticipate that even if everybody in the world owned an IBM XT, there weren't enough addresses. In fairness, what many people were doing at the time was running a non-TCP network locally and having just one "gateway" device. But even then, you don't get past the one IP address per person problem. The way that addresses are assigned has certain exacerbated the problem.
Or, more likely, they migrated the server to 64-bit when they hit 1B games and also the 64-bit clients but didn't test the 32-bit clients. As BronsCon has pointed out multiple times, they probably got lucky on the 64-bit version in that they happened to have used a size that was platform dependent. I don't know the chess.com protocol, but if it returns XML or JSON or the like, the number ends up in a "string" representation and then converted back to an actual integer type. If they were using RPC or something that had strong typing this would be harder to do. If somebody explicitly cast 64 to 32-bit (or did an assignment that lopped off the high-order bits), that would be a bit more brain-dead.
I think your previous post is pretty valuable (and I would mod it up if I could). These two were addressing slightly different points but I do see why there is an inconsistency in my answer. (Calling me "unstable" is a bit over the top but hopefully meant in good humor.)
In 2005 nobody was thinking about 64-bit iOS (It came about in 2013) and so there are two problems that one would be facing here.
The first would be what the correct range of the unique id should be and the second is forward-compatibility when 64-bit mobile operating systems came about. In 2005, I didn't know anybody who even had a 64-bit desktop OS.
So really there were two things so far in the future as to be unimaginable and, in order to avoid this, one would have had to correctly anticipate *both*
Who says they used an architecture dependent int type. They very well may have declared an int64 and then, when they ported to 64-bit, changed it to int128 but never merged the change back to the 32-bit branch. Not everybody has one branch that builds for 32 and 64-bit. Often the 64-bit transition is considered "one way" and the 32-bit branch isn't maintained.
Well, yes, but the problem here is that how does one know the correct range when first starting to build an application? There was probably a design document (or just written on a white board) that the apps should be able to handle 2M games. Then suddenly they are very successful and get 2B.
Considering the designers of the Internet got this wrong with IP addresses and had the benefit of huge amounts of collaboration, I fail to see how a relatively tiny operation with a small budget can be considered idiots for this. Probably when they started they were just trying to get an app out at all and intended to fix this if they ever got near 2B games. But probably at the time it seemed like an impossible number.
Only Congress can charge the president with a crime. This provision exists to ensure that the governance of the country cannot be interrupted by frivolously charging the president. Neither can Congressmen be "detained" as allowing it would present a means to prevent them from getting to important votes. To charge the president with a crime, you essentially need to follow the impeachment procedures.
I'm with Osgeld (sister post). The fact is that most of us don't need near the amount of available horsepower. Even running compilers and Java server applications (i.e. Tomcat + WebApp) is within the realm of the processors we see in phones. We're seeing Atom-based laptops (i.e. https://www.newegg.com/Product...) getting adopted. These aren't much more powerful than the S8. People are talking about docking stations for phones but I actually think that things will go a different direction. Windows remote desktop is fabulous (No VNC and the like don't come close) and I imagine people will have everything on their phone and apps will run either in tablet or desktop mode. When you get where you're going, plug your phone into a power source and RDP to it from a dumb terminal and pick up right where you left off. Saves the problem of having to have docking stations for different types of devices.
As I've pointed out earlier, they should have been able to fail-over to another data center. So the fact that they didn't have these procedures and/or hadn't tested them is a management failure. The localized problem, though, should not be blamed on management.
I agree with you that the OPs assumption reads a lot of world-view into a statement that is fairly light on facts. However, there is also probably some truth. A single data center failure shouldn't have caused such an outage. There should be a redundant data center. So, yes, it's reasonable for somebody to accidentally shut off all the power to a data center or for a natural disaster to wipe it out. But it's not reasonable for an operation the size of BA not to have redundant servers somewhere.
"I think the animal and retrospective data is compelling, and I want this treatment to be available to people." - I want to make as much money on this as quickly as I can before it's disproven.
I think you may be proving my point here. Those are pretty significant investments. A thick brick wall is much more expensive than the cinderblock (which is a pretty good insulator) and stick framing (Also not terrible since air is a good insulator) already in use. And their benefit is only marginal. Again air is a good insulator if you don't get convection which, in modern construction, you don't. You have to have a pretty significant heat gradient before insulation matters much. In the winter in Alaska, sure. Summer in Florida? Cooling a 3k sq/ft house is less than a hundred bucks a month. Even if electricity prices doubled, investments in insulation wouldn't make much sense. And in many cases, people have tried to insulate and ended up trapping moisture causing damage. It's cold-weather areas that are carbon-intensive for heating/cooling.
Because air condition is already so efficient and since you typically don't have much of a temperature gradient (76F inside, 96F outside), it's not clear that there is much of a net benefit for insulation in hot parts of the country. (I live in Florida, the air conditioning capital of the world.) If electricity were produced cleanly (solar, wind), air conditioning would be a non-factor. Cars seem to be the hard part because it's not so easy to switch them to cleaner energy sources.
More importantly, I've never talked to a corporate travel agent and told them a cheaper itinerary and had them refuse to book it. If you know what you want it's easy to book through corporate travel.
If the middle man adds value (some do), I don't think that OP applies. Sometimes middle men are high value because they do the things you suggest (aggregate a group of customers and get them a better rate, for example). Many middle-men add no value. The example given was real estate agents who charge 6% to put your house on MLS!
Ed
Slashdot Mirror
Deploy a pico-cell that only carries 911 calls and no data. In terms of an escalation technical arms race there are plenty of things that can be done. I sure as heck hope this backfires and people simply walk out of the store to check the price and (maybe) walk back in. The downside to the retailer is that they may have the best price or at least one that's good enough that people would be a premium for instant gratification but since they are now already outside of the store, they may decide it's not worth going back in.
Well, to be devil's advocate, if I were implementing this inside of my store, you can be pretty sure that I'd make sure that no 4G signals reached inside so my WiFi would be your only choice. Not that I'm advocating for this, but there's no point of deploying a solution when there's a trivial workaround.
Well if you follow my original suggestion of having the code formatted upon checkin, the local formatting changes don't matter!
Almost every IDE can format code. And *good* IDEs will do it perfectly. Unfortunately what they don't do is preserve the original formatting mechanism for future checkins. There's a menu item in Eclipse but I don't know what it is. The key combination is Shift-Ctrl-F to reformat the code. For Java especially, it's super-easy for Eclipse to do this since Eclipse knows what the parse tree looks like. Some formatting utilities can be very weak in they rely on things like regexes and counting brackets which could be error prone. But if you format code by first parsing into the AST and then regenerating the code from that, you'll get an accurate formatting every time.
The answer is that the IDE should allow you to view the code with your own preferences, but what goes into revision control should get automatically formatted with whatever standard is set for a project.
What is shocking to me is that we even have this discussion. There ought to be a check-in hook that properly formats the code. Would also reduce defects where a brace/bracket is missing.
No but every *person* may need at least 1 IP address that that's will 6B people and an address range of 2B. When IPv4 was designed, there was no NAT so it was easy to anticipate that even if everybody in the world owned an IBM XT, there weren't enough addresses. In fairness, what many people were doing at the time was running a non-TCP network locally and having just one "gateway" device. But even then, you don't get past the one IP address per person problem. The way that addresses are assigned has certain exacerbated the problem.
Or, more likely, they migrated the server to 64-bit when they hit 1B games and also the 64-bit clients but didn't test the 32-bit clients. As BronsCon has pointed out multiple times, they probably got lucky on the 64-bit version in that they happened to have used a size that was platform dependent. I don't know the chess.com protocol, but if it returns XML or JSON or the like, the number ends up in a "string" representation and then converted back to an actual integer type. If they were using RPC or something that had strong typing this would be harder to do. If somebody explicitly cast 64 to 32-bit (or did an assignment that lopped off the high-order bits), that would be a bit more brain-dead.
I think your previous post is pretty valuable (and I would mod it up if I could). These two were addressing slightly different points but I do see why there is an inconsistency in my answer. (Calling me "unstable" is a bit over the top but hopefully meant in good humor.) In 2005 nobody was thinking about 64-bit iOS (It came about in 2013) and so there are two problems that one would be facing here. The first would be what the correct range of the unique id should be and the second is forward-compatibility when 64-bit mobile operating systems came about. In 2005, I didn't know anybody who even had a 64-bit desktop OS. So really there were two things so far in the future as to be unimaginable and, in order to avoid this, one would have had to correctly anticipate *both*
Does it really take more work to declare something int128?
Who says they used an architecture dependent int type. They very well may have declared an int64 and then, when they ported to 64-bit, changed it to int128 but never merged the change back to the 32-bit branch. Not everybody has one branch that builds for 32 and 64-bit. Often the 64-bit transition is considered "one way" and the 32-bit branch isn't maintained.
Well, yes, but the problem here is that how does one know the correct range when first starting to build an application? There was probably a design document (or just written on a white board) that the apps should be able to handle 2M games. Then suddenly they are very successful and get 2B.
Considering the designers of the Internet got this wrong with IP addresses and had the benefit of huge amounts of collaboration, I fail to see how a relatively tiny operation with a small budget can be considered idiots for this. Probably when they started they were just trying to get an app out at all and intended to fix this if they ever got near 2B games. But probably at the time it seemed like an impossible number.
Only Congress can charge the president with a crime. This provision exists to ensure that the governance of the country cannot be interrupted by frivolously charging the president. Neither can Congressmen be "detained" as allowing it would present a means to prevent them from getting to important votes. To charge the president with a crime, you essentially need to follow the impeachment procedures.
I'm with Osgeld (sister post). The fact is that most of us don't need near the amount of available horsepower. Even running compilers and Java server applications (i.e. Tomcat + WebApp) is within the realm of the processors we see in phones. We're seeing Atom-based laptops (i.e. https://www.newegg.com/Product...) getting adopted. These aren't much more powerful than the S8. People are talking about docking stations for phones but I actually think that things will go a different direction. Windows remote desktop is fabulous (No VNC and the like don't come close) and I imagine people will have everything on their phone and apps will run either in tablet or desktop mode. When you get where you're going, plug your phone into a power source and RDP to it from a dumb terminal and pick up right where you left off. Saves the problem of having to have docking stations for different types of devices.
I'm actually very happy with my Kindle Fire, though!
As I've pointed out earlier, they should have been able to fail-over to another data center. So the fact that they didn't have these procedures and/or hadn't tested them is a management failure. The localized problem, though, should not be blamed on management.
I agree with you that the OPs assumption reads a lot of world-view into a statement that is fairly light on facts. However, there is also probably some truth. A single data center failure shouldn't have caused such an outage. There should be a redundant data center. So, yes, it's reasonable for somebody to accidentally shut off all the power to a data center or for a natural disaster to wipe it out. But it's not reasonable for an operation the size of BA not to have redundant servers somewhere.
"I think the animal and retrospective data is compelling, and I want this treatment to be available to people." - I want to make as much money on this as quickly as I can before it's disproven.
I think you may be proving my point here. Those are pretty significant investments. A thick brick wall is much more expensive than the cinderblock (which is a pretty good insulator) and stick framing (Also not terrible since air is a good insulator) already in use. And their benefit is only marginal. Again air is a good insulator if you don't get convection which, in modern construction, you don't. You have to have a pretty significant heat gradient before insulation matters much. In the winter in Alaska, sure. Summer in Florida? Cooling a 3k sq/ft house is less than a hundred bucks a month. Even if electricity prices doubled, investments in insulation wouldn't make much sense. And in many cases, people have tried to insulate and ended up trapping moisture causing damage. It's cold-weather areas that are carbon-intensive for heating/cooling.
In many parts of the country, the carbon tax would have to be huge. In places where insulation matters it's already well deployed.
Because air condition is already so efficient and since you typically don't have much of a temperature gradient (76F inside, 96F outside), it's not clear that there is much of a net benefit for insulation in hot parts of the country. (I live in Florida, the air conditioning capital of the world.) If electricity were produced cleanly (solar, wind), air conditioning would be a non-factor. Cars seem to be the hard part because it's not so easy to switch them to cleaner energy sources.
Yes I remember that chapter quite well. Thanks for reminding me.
More importantly, I've never talked to a corporate travel agent and told them a cheaper itinerary and had them refuse to book it. If you know what you want it's easy to book through corporate travel.
If the middle man adds value (some do), I don't think that OP applies. Sometimes middle men are high value because they do the things you suggest (aggregate a group of customers and get them a better rate, for example). Many middle-men add no value. The example given was real estate agents who charge 6% to put your house on MLS! Ed