Or via some other mechanism. Email works fine for long messages and really doesn't need yet another incompatible reinvention. For short messages, there's IM, SMS, and so on, and Twitter does the broadcast version of these. Facebook tries to be the broadcast version of email and largely exists because email sucks if you want to share photos and setting up a mailing list is hard for non-techy people (even with Yahoo / Google groups).
That's basically how cardholder-present transactions work. You enter your pin, the card produces a hash of the recipient, time, and amount, and a shared secret. The merchant then presents the hash, the time and the amount to your card issuer (via some layers of indirection) and they confirm that the transaction is valid. For Internet transactions, unlike in-person transactions, you can guarantee that the recipient has network connectivity, so it's even easier for them to communicate with the bank and verify the hash.
Some of the schemes for one-time CC numbers actually allow the CC number to be re-used, but it's only valid for one transaction of a specific amount per day. If you want to use it again, you have to correctly guess the amount that it's valid for that day, and put in your fake transaction after the next person to be issued with it requests it, but before they use it.
My mother got a Windows 8 laptop just before Christmas. This holiday has involved a lot of listening to her swear at it. Her desktop ran XP and Win8 is so different that she finds is almost unusable. 8.1 is marginally better, but still does weird things (like pop up the People app and ask her to sign into the Microsoft account that she doesn't have, in the middle of her typing a password in a web page). I hadn't realised quite how bad 8 was until I saw her trying to use it.
I don't need to install from Google Play. In fact, I install software from other marketplaces including F-Droid
That's more or less true, but it's increasingly untrue. For example, there haven't been updates to the open source mail or web apps for a while. All of Google's effort goes to the GMail app and Chrome. The same thing for the music player app. Google abandons development in favour of their proprietary ones. This makes it increasingly difficult to create a Google-free Android. Samsung and Amazon are both interested in a Google-free Android platform though, so there's hope that they'll start collaborating...
Meh, I have a macbook pro too, but I'm more than a little non-plussed by the newest one... no ethernet without dongle is a joke in a pro level product.
It seemed odd to me. I believe that their idea is that you should use an Apple display with the Ethernet provided via Thunderbolt. In this setting, you don't need the GigE connection on the machine, because it appears as soon as you plug in the display, and whenever you're not at your desk you use WiFi. I almost never use the GigE port on mine anywhere other than at work, so it's not completely silly (I don't have the Apple display, but I leave a Thunderbolt GigE adaptor connected to the network cable, so I just plug it in when I arrive).
I have them, so I might as well use them. It's not as if they're the devil incarnate that will rape your babies just because you use them.
No, but they will cost you more money. I replaced 60W incandescents with either 15W or 20W CFLs[1]. Assuming the 20W ones, that's a 40W saving, so to convert into electricity-company units, that's 25 hours to save one kWh. The cost of a new CFL is, for me, around the same cost as 2-3kWh of electricity. If I buy a new CFL today and install it replacing an incandescent that is on for an average of 2 hours a day, then it takes me less than two months to be better off, including the initial outlay.
[1] A lot of the complaints I've seen seem to be from people who replace 100W incandescents with 10W CFLs and then complain that they're dark. Don't trust the 'Watt equivalent' numbers that manufacturers put on the boxes - they're pretty much nonsense.
Let's see, listen to these lying bastards and buy a "lifetime" bulb for $20. each, OR just use the 30cent bulbs that I've been using for 50 years?
The last CFLs I bought were around 50, about 3-4 years ago. In their first six months of use, they saved more electricity than their total purchase cost. The last time I saw a CFL on sale for close to $20 was 12 years ago, and that was for a very bright one that was intended to go outside and so had some extra circuitry for starting when very cold. Last time I looked, the up-front price difference for CFLs and incandescents was less that the energy saving from about 10 hours of use.
I live in a cold country. I don't care about a trivial power cost. I need the heat. Doesn't matter if it comes from a bulb or natural gas. Shit, I heat with computers sometimes!
I suppose it's your money to waste, but given that heat generated from gas tends to cost around 20-50% of heat generated from electricity depending on where you live, you're not making a very convincing argument for anyone else.
You can pry my incandescents from my cold dead hands!
Being willing to defend poor choices with your life is generally a sign of idiocy.
Since that time I'm sure it was passed onto other HOME OWNERS. It was the Home owners seeking the savings rather that Renters or Apt. Dwellers
I switched all of the bulbs in a place I was renting 12 years ago to CFLs. They were expensive back then, but by the time I moved out I'd saved enough on the electricity to offset the cost. I took the CFLs with me to the next flat and replaced the bulbs in the old one with the ones I'd removed. Fast forward to about 4 years ago, when I bought my own place and I left the aging CFLs behind when I moved. I'd replaced about half of them over that period, but I was now paying 30p/bulb (about 50), so it seemed more effort to move them than to buy new ones.
I would expect people who are renting to be more keen to save money than people who own their own home, but perhaps it's crazy non-American logic to expect people with less capital to be more interested in saving...
16GB of RAM? What is this, 2000? My laptop has 16GB of RAM, the machines in the rack have 256GB, and they're a year old (and due an upgrade). If you're running a database on something with 16GB of RAM, you either don't have much data or you're seriously skimping on hardware.
No, I'm not shouting at you, I'm shouting at the fucking bogus pseudo-academics who wanted to bullshit with micro-optimisation rather than making actual advancements in the field of databases.
Any paper that does X on a GPU generally fits into this category. It's not science to run an existing algorithm on an existing Turing-complete processor. At most it's engineering. But it's a fairly easy way to churn out papers. Doing X 'in the cloud' or 'with big data' have a similar strategy. It's usually safe to ignore them.
The C standard does not guarantee that sizeof(long) is as big as sizeof(void*). The type that you want is intptr_t (or ptrdiff_t for differences between pointers). If you've gone through replacing everything with long, then good luck getting your code to run on win64 (where long is 4 bytes).
He's right. If you mix x32 and amd64 binaries on the same system, then you need two copies of every shared library that they use to be mapped at the same time. And this means that every context switch between them is going to be pulling things into the i-cache that would already be present (assuming a physically-mapped cache, which is a pretty safe assumption these days) because the other process is using them.
This is why x32 doesn't make sense on a consumer platform like Ubuntu unless the entire system is compiled to use it, making the entire article a 'well, duh'. The real advantage of x32 is on custom deployments and embedded systems where you can build everything in x32 mode.
Oh, and on the subject of caches, x86 chips typically have 64 byte cache lines. If you make pointers 4 bytes instead of 8, then you can fit twice as many in a cache line, which is usually nice. It can be a problem for multithreaded applications though, because you may now end up with more contention in the cache coherency protocol.
It depends on the reason. If it can't take HTML/CSS/JavaScript and display it correctly, then it's simply a buggy device. If, however, it has a very small screen then there's a good reason for asking for a custom version. Browsing the web from my tablet, I typically want the same UI as on a desktop. The screen is big enough and it's fine. On my phone, I really hate the 3-column layout that lots of sites love because it's a pain to zoom just enough to see the middle one (where the content goes) and to scroll keeping it inline.
And some things, like mouse-over pop-up menus are just a bad idea on any device. At best they're lacking in discoverability, at worst they're impossible to use.
Prior to FB's IPO, Goldman Sachs invested $500M for 1% ownership. This implied a $50B valuation of the company that earned less than $1B in revenue. Any other company out there would be around $15-$20B with $1B in revenue, not $50.
All that really means is that GS expected that they could offload 1% of FB stock for more than $500M. Their mechanism for doing this was to create a fund backed by the FB shares that they could then trade. Their favoured customers were allowed to buy it early. They then hyped it and let those customers sell it on again. Then, at the Facebook IPO, they liquidated the fund and turned it into public shared. The people who bought it at the highest level of hype made a loss, GS made a profit, and GS's favoured customers made a profit.
It didn't matter to GS what the fund was actually backed by, just that they could hype it and dump it.
Snapchat doesn't delete content (as their recent disclosures about law enforcement access showed). They just stop it being accessible to the recipient. But people assume this is the same as deleting it, so they keep using it to send embarrassing or sensitive pictures. This means that Snapchat is a platform for building the world's largest database of blackmail material, and even better the victims, uh, users, contribute the material themselves. Anyone who can't see a revenue stream there is extremely shortsighted.
Not the original poster, but yes, I'm sure. Well, not quite. He meant to say 1/3 of the traffic on the Internet in the USA, but this is Slashdot and so the Internet and the USA get mixed up quite often. Netflix runs their own CDN, which runs FreeBSD. Scott Long has given a few talks about their infrastructure. They do some quite crazy things with aio to pull 1MB chunks of a load of films into RAM and then dump them out over the network (it turns out people seek quite a lot, so being too aggressive in the prefetching isn't always a win, and you've got a lot of concurrent clients so it's easy to run out of RAM if you pull too much in).
Games are written using OpenGL (or Direct3D), which doesn't expose the notion of a screen pixel as a programmer abstraction at all, beyond creating the original context. OpenGL programmers only need to care about the aspect ration once the initial configuration is done. Everything else is done in terms of a floating point coordinate space.
The same is true of any vaguely modern GUI toolkit. Pixels are simply not exposed as a programmer abstraction. In 2D, it matters slightly more, because you often people often stick images on, but if the OS does relatively competent scaling then it doesn't matter too much. It only matters when you do something stupid, like putting text in an image, because then people notice that the text is more blurry than the rest of it.
If you've never written software using mod, then you've never implemented a lockless data structure, or likely written any low-level code. In fact, I'd find it pretty hard to think of what category of software you are writing where you never need to do it. It's not something I use every day, but I'd expect to see it at least once every thousand lines of code. I've no idea what you do, but if you think you've managed to become a competent programmer without ever having to use mod then you're such a statistical outlier that it's not likely to be a problem to accidentally not hire you. If you don't know the basic arithmetic and logical operators in your favourite language, then you are probably not worth hiring.
Someone I was talking to a few months ago claimed that around half of the applicants they get fail to be able to write fizzbuzz in their interview. This is just a function that has to take an input number, print fizz if it's divisible by 3, buzz if it's divisible by 5 (fizzbuzz if it's divisible by both), or the number otherwise. It's a trivial use of simple flow control and if you can't remember the name of a print function in your language of choice (no marks deducted for using printf instead of puts) then there's really no hope for you.
The government does acknowledge the existence of bitcoins. They're just not sure quite how to classify them. They don't have most of the properties of a currency, but they do have some. They have most of the properties of a security, but not all. They have a lot of the properties of a scam, but not all.
It helps less than other mitigation techniques, even on 64-bit, and it breaks a lot of things (e.g. the JVM and other language runtimes). To do it properly you need to provide a good mechanism for programs (and libraries) to opt out of it. Or you can got the OpenBSD route and say 'well, that code must be bad so we won't run it'. We have downstream consumers who consider that to be unacceptable.
For a single user, sure, the text editor is nice and easy. Now how about a multi-user box with a few thousand of them? You can do some grep -v stuff, but now you need to think about concurrency (I bet one of your users will be editing the file at the same time as your automated sweep). And so now you need locking, but your text editor won't lock the file that you open by default, so now what happens? This is part of the reason wrappers like vipw and visudo exist: so that you don't get update races (sanity checking is the other reason).
Slashdot Mirror
Or via some other mechanism. Email works fine for long messages and really doesn't need yet another incompatible reinvention. For short messages, there's IM, SMS, and so on, and Twitter does the broadcast version of these. Facebook tries to be the broadcast version of email and largely exists because email sucks if you want to share photos and setting up a mailing list is hard for non-techy people (even with Yahoo / Google groups).
That's basically how cardholder-present transactions work. You enter your pin, the card produces a hash of the recipient, time, and amount, and a shared secret. The merchant then presents the hash, the time and the amount to your card issuer (via some layers of indirection) and they confirm that the transaction is valid. For Internet transactions, unlike in-person transactions, you can guarantee that the recipient has network connectivity, so it's even easier for them to communicate with the bank and verify the hash.
Some of the schemes for one-time CC numbers actually allow the CC number to be re-used, but it's only valid for one transaction of a specific amount per day. If you want to use it again, you have to correctly guess the amount that it's valid for that day, and put in your fake transaction after the next person to be issued with it requests it, but before they use it.
My mother got a Windows 8 laptop just before Christmas. This holiday has involved a lot of listening to her swear at it. Her desktop ran XP and Win8 is so different that she finds is almost unusable. 8.1 is marginally better, but still does weird things (like pop up the People app and ask her to sign into the Microsoft account that she doesn't have, in the middle of her typing a password in a web page). I hadn't realised quite how bad 8 was until I saw her trying to use it.
I don't need to install from Google Play. In fact, I install software from other marketplaces including F-Droid
That's more or less true, but it's increasingly untrue. For example, there haven't been updates to the open source mail or web apps for a while. All of Google's effort goes to the GMail app and Chrome. The same thing for the music player app. Google abandons development in favour of their proprietary ones. This makes it increasingly difficult to create a Google-free Android. Samsung and Amazon are both interested in a Google-free Android platform though, so there's hope that they'll start collaborating...
American cheese usually doesn't, that's why it tastes of plastic.
Meh, I have a macbook pro too, but I'm more than a little non-plussed by the newest one... no ethernet without dongle is a joke in a pro level product.
It seemed odd to me. I believe that their idea is that you should use an Apple display with the Ethernet provided via Thunderbolt. In this setting, you don't need the GigE connection on the machine, because it appears as soon as you plug in the display, and whenever you're not at your desk you use WiFi. I almost never use the GigE port on mine anywhere other than at work, so it's not completely silly (I don't have the Apple display, but I leave a Thunderbolt GigE adaptor connected to the network cable, so I just plug it in when I arrive).
I have them, so I might as well use them. It's not as if they're the devil incarnate that will rape your babies just because you use them.
No, but they will cost you more money. I replaced 60W incandescents with either 15W or 20W CFLs[1]. Assuming the 20W ones, that's a 40W saving, so to convert into electricity-company units, that's 25 hours to save one kWh. The cost of a new CFL is, for me, around the same cost as 2-3kWh of electricity. If I buy a new CFL today and install it replacing an incandescent that is on for an average of 2 hours a day, then it takes me less than two months to be better off, including the initial outlay.
[1] A lot of the complaints I've seen seem to be from people who replace 100W incandescents with 10W CFLs and then complain that they're dark. Don't trust the 'Watt equivalent' numbers that manufacturers put on the boxes - they're pretty much nonsense.
Let's see, listen to these lying bastards and buy a "lifetime" bulb for $20. each, OR just use the 30cent bulbs that I've been using for 50 years?
The last CFLs I bought were around 50, about 3-4 years ago. In their first six months of use, they saved more electricity than their total purchase cost. The last time I saw a CFL on sale for close to $20 was 12 years ago, and that was for a very bright one that was intended to go outside and so had some extra circuitry for starting when very cold. Last time I looked, the up-front price difference for CFLs and incandescents was less that the energy saving from about 10 hours of use.
I live in a cold country. I don't care about a trivial power cost. I need the heat. Doesn't matter if it comes from a bulb or natural gas. Shit, I heat with computers sometimes!
I suppose it's your money to waste, but given that heat generated from gas tends to cost around 20-50% of heat generated from electricity depending on where you live, you're not making a very convincing argument for anyone else.
You can pry my incandescents from my cold dead hands!
Being willing to defend poor choices with your life is generally a sign of idiocy.
Since that time I'm sure it was passed onto other HOME OWNERS. It was the Home owners seeking the savings rather that Renters or Apt. Dwellers
I switched all of the bulbs in a place I was renting 12 years ago to CFLs. They were expensive back then, but by the time I moved out I'd saved enough on the electricity to offset the cost. I took the CFLs with me to the next flat and replaced the bulbs in the old one with the ones I'd removed. Fast forward to about 4 years ago, when I bought my own place and I left the aging CFLs behind when I moved. I'd replaced about half of them over that period, but I was now paying 30p/bulb (about 50), so it seemed more effort to move them than to buy new ones.
I would expect people who are renting to be more keen to save money than people who own their own home, but perhaps it's crazy non-American logic to expect people with less capital to be more interested in saving...
16GB of RAM? What is this, 2000? My laptop has 16GB of RAM, the machines in the rack have 256GB, and they're a year old (and due an upgrade). If you're running a database on something with 16GB of RAM, you either don't have much data or you're seriously skimping on hardware.
No, I'm not shouting at you, I'm shouting at the fucking bogus pseudo-academics who wanted to bullshit with micro-optimisation rather than making actual advancements in the field of databases.
Any paper that does X on a GPU generally fits into this category. It's not science to run an existing algorithm on an existing Turing-complete processor. At most it's engineering. But it's a fairly easy way to churn out papers. Doing X 'in the cloud' or 'with big data' have a similar strategy. It's usually safe to ignore them.
The C standard does not guarantee that sizeof(long) is as big as sizeof(void*). The type that you want is intptr_t (or ptrdiff_t for differences between pointers). If you've gone through replacing everything with long, then good luck getting your code to run on win64 (where long is 4 bytes).
He's right. If you mix x32 and amd64 binaries on the same system, then you need two copies of every shared library that they use to be mapped at the same time. And this means that every context switch between them is going to be pulling things into the i-cache that would already be present (assuming a physically-mapped cache, which is a pretty safe assumption these days) because the other process is using them.
This is why x32 doesn't make sense on a consumer platform like Ubuntu unless the entire system is compiled to use it, making the entire article a 'well, duh'. The real advantage of x32 is on custom deployments and embedded systems where you can build everything in x32 mode.
Oh, and on the subject of caches, x86 chips typically have 64 byte cache lines. If you make pointers 4 bytes instead of 8, then you can fit twice as many in a cache line, which is usually nice. It can be a problem for multithreaded applications though, because you may now end up with more contention in the cache coherency protocol.
It depends on the reason. If it can't take HTML/CSS/JavaScript and display it correctly, then it's simply a buggy device. If, however, it has a very small screen then there's a good reason for asking for a custom version. Browsing the web from my tablet, I typically want the same UI as on a desktop. The screen is big enough and it's fine. On my phone, I really hate the 3-column layout that lots of sites love because it's a pain to zoom just enough to see the middle one (where the content goes) and to scroll keeping it inline.
And some things, like mouse-over pop-up menus are just a bad idea on any device. At best they're lacking in discoverability, at worst they're impossible to use.
Prior to FB's IPO, Goldman Sachs invested $500M for 1% ownership. This implied a $50B valuation of the company that earned less than $1B in revenue. Any other company out there would be around $15-$20B with $1B in revenue, not $50.
All that really means is that GS expected that they could offload 1% of FB stock for more than $500M. Their mechanism for doing this was to create a fund backed by the FB shares that they could then trade. Their favoured customers were allowed to buy it early. They then hyped it and let those customers sell it on again. Then, at the Facebook IPO, they liquidated the fund and turned it into public shared. The people who bought it at the highest level of hype made a loss, GS made a profit, and GS's favoured customers made a profit.
It didn't matter to GS what the fund was actually backed by, just that they could hype it and dump it.
Snapchat doesn't delete content (as their recent disclosures about law enforcement access showed). They just stop it being accessible to the recipient. But people assume this is the same as deleting it, so they keep using it to send embarrassing or sensitive pictures. This means that Snapchat is a platform for building the world's largest database of blackmail material, and even better the victims, uh, users, contribute the material themselves. Anyone who can't see a revenue stream there is extremely shortsighted.
More to the point, they also support an Android app, which means that they support most consumer Linux installs.
Not the original poster, but yes, I'm sure. Well, not quite. He meant to say 1/3 of the traffic on the Internet in the USA, but this is Slashdot and so the Internet and the USA get mixed up quite often. Netflix runs their own CDN, which runs FreeBSD. Scott Long has given a few talks about their infrastructure. They do some quite crazy things with aio to pull 1MB chunks of a load of films into RAM and then dump them out over the network (it turns out people seek quite a lot, so being too aggressive in the prefetching isn't always a win, and you've got a lot of concurrent clients so it's easy to run out of RAM if you pull too much in).
Ironically while Netflix's infrastructure runs on Linux
And even Phoronix managed to get that one right...
Games are written using OpenGL (or Direct3D), which doesn't expose the notion of a screen pixel as a programmer abstraction at all, beyond creating the original context. OpenGL programmers only need to care about the aspect ration once the initial configuration is done. Everything else is done in terms of a floating point coordinate space.
The same is true of any vaguely modern GUI toolkit. Pixels are simply not exposed as a programmer abstraction. In 2D, it matters slightly more, because you often people often stick images on, but if the OS does relatively competent scaling then it doesn't matter too much. It only matters when you do something stupid, like putting text in an image, because then people notice that the text is more blurry than the rest of it.
If you've never written software using mod, then you've never implemented a lockless data structure, or likely written any low-level code. In fact, I'd find it pretty hard to think of what category of software you are writing where you never need to do it. It's not something I use every day, but I'd expect to see it at least once every thousand lines of code. I've no idea what you do, but if you think you've managed to become a competent programmer without ever having to use mod then you're such a statistical outlier that it's not likely to be a problem to accidentally not hire you. If you don't know the basic arithmetic and logical operators in your favourite language, then you are probably not worth hiring.
Someone I was talking to a few months ago claimed that around half of the applicants they get fail to be able to write fizzbuzz in their interview. This is just a function that has to take an input number, print fizz if it's divisible by 3, buzz if it's divisible by 5 (fizzbuzz if it's divisible by both), or the number otherwise. It's a trivial use of simple flow control and if you can't remember the name of a print function in your language of choice (no marks deducted for using printf instead of puts) then there's really no hope for you.
The government does acknowledge the existence of bitcoins. They're just not sure quite how to classify them. They don't have most of the properties of a currency, but they do have some. They have most of the properties of a security, but not all. They have a lot of the properties of a scam, but not all.
It helps less than other mitigation techniques, even on 64-bit, and it breaks a lot of things (e.g. the JVM and other language runtimes). To do it properly you need to provide a good mechanism for programs (and libraries) to opt out of it. Or you can got the OpenBSD route and say 'well, that code must be bad so we won't run it'. We have downstream consumers who consider that to be unacceptable.
For a single user, sure, the text editor is nice and easy. Now how about a multi-user box with a few thousand of them? You can do some grep -v stuff, but now you need to think about concurrency (I bet one of your users will be editing the file at the same time as your automated sweep). And so now you need locking, but your text editor won't lock the file that you open by default, so now what happens? This is part of the reason wrappers like vipw and visudo exist: so that you don't get update races (sanity checking is the other reason).