Actually you get quite a bit of advantage. The instruction set for 64 bit mode is quite a bit more efficient even when you're not doing 64 bit arithmetic, not least because there are twice as many registers. Code that can be optimised using the SIMD instructions also gets twice as many SIMD registers, which can make a big difference for many graphics, signal processing and mathematical tasks. Most floating point code will be more efficient in 64 bit mode too.
Furthermore the 64 bit address space is useful even with only 1GB of RAM since it allows you to mmap files that are bigger than 4GB. Programs like MongoDB mmap their database files and when running on 32 bit processors MongoDB actually limits processes to a total of 2GB of data in the database, irrespective of how small the working set is. With a 64 bit address space the limit goes from below typical storage sizes to far beyond any plausible storage you'll ever connect to a Pi.
The idea that "the fact that only companies and labs have access to this technology can represent a threat" is patently absurd. Theano, Caffe and Torch are all open source and even Google has open-sourced its Tensor Flow platform which makes it easy to build new tools and run then, fast, on all the GPUs you can find. If you need to do this at scale and you're not the size of Google or IBM you can use Amazon's Machine Learning for AWS. There are many, many higher level toolkits out there that are available under licenses that are much less restrictive than GPLv3.
How about nVidia's Jetson TK1 board? It has a great 192-core Kepler GPU, a nice quad-core ARM CPU, on-board gigabit ethernet, all the ports you're likely to need and comes pre-installed with Ubuntu Linux.
Even if you never make use of the new programming language it is almost always worth getting your head around a new way of thinking about problems. You may not ever need to write code in Lisp but understanding what a functional language is and isn't good for is helpful in other languages. If you're building flight control systems then Python might not be the language to do it in but getting deep into it and start understanding why you'd want a metaclass in the first place can help you stricture your code better. There aren't a ton of job openings for Erlang programmers but there are lots for people who understand High Availability and Disaster Recovery and the background knowledge will stand you in good stead.
In short, yes, it's worth learning all the languages your brain can handle. Even if you forget the details the concepts will stick around and help you later.
Ubiquiti Networks make some decent cameras with a whole bunch of decent features including fetching a snapshot with an HTTP request. They are designed to send their video output to their DVR software (which is actively supported on Linux) but in practice if all you need is to access still images over HTTP and video over RTSP then you can set up the control software on your laptop, fire it up once to configure the camera and then switch it off and the camera will continue to run without the DVR.
Of course, as other posters have pointed out, the right answer is to brew your own with a Raspberry Pi and a Pi Camera.
Firstly, if you can facilitate multi-way calling then it is clearly technically feasible to support a wire tap. Secondly, unlike many other snooping regulations, CALEA explicitly obliges telecommunications companies to modify their systems and equipment in order to facilitate "lawful access" (sic). Verizon are a telco, not an app company, so they are bound by CALEA in ways that people like Silent Circle or CellTrust are not.
If I was publishing an article talking about how huge numbers of eyeballs solves security problem I'm not sure that I'd choose to publish it the day after it was announced that the X window server code has had some serious security bugs for 25 years that have only just been discovered. Clearly open source code can have serious security holes that go unnoticed for a very long time.
Presumably when they OP author wrote "a relational database that is as capable as 'proprietary database engines at 1/10 the cost,' " what (s)he really meant was "a relational database (that is as capable as proprietary database engines) at 1/10 the cost".
This begs the question whether mining for BitCoins is more damaging to the environment than mining for precious metals, for a given value of return. The EPA emissions factor for electricity is about 0.69 tons of CO2 per megawatt hour, so producing the electricity used by this datacenter is, on average, dumping into the atmosphere 331 tons of CO2 per day or about 120,000 tons of CO2 per year. While there are many other forms of environmental damage from gold mining, a quick search suggest that the greenhouse emissions from gold extraction run to about 11.5 tons of CO2 equivalent per kg of Gold. At this rate 120,000 tons of CO2 yields of 10.5 tons of gold, worth nearly $500 million at today's price. Will this datacentre yield more than half a billion dollars worth of bit coins each year?
I have four RPi boards. One monitors my UPSs, cleanly suspending my server when the power goes out and sending wake-on-LAN massages to it when the power comes back up so that the UPS only needs to drive my switch and AP, one has a camera board and does motion detection to spot people coming into my office, one is currently operating as a Bluetooth LE beacon for testing the new iOS iBeacon stuff and one is just for tinkering. Most of these have a few other services running on them too (two have I2C thermometers on them).
I see a lot of negative comments about the Pi being underpowered. Perhaps if what you want to do is run FPS games or you are trying to run Big Data analytics then this is true but it's plenty powerful enough for a whole host of service tasks. It's not that many years ago that the Pi's level of power would have been considered a high-end desktop configuration. The purpose of the device is to give kids a low-cost entry into programming and it does just that. On top, at $25 for a Model A its fine to put in 'dangerous' places where something bad might happen to it (like outdoors, driving the sensors and servos for my Halloween decorations). No, I don't have my MongoDB server on a Raspberry Pi, but for many many projects they are just about perfect.
"More than an actual technology, SDDN is the culmination of many other efforts at abstracting, consolidating, managing, provisioning, load balancing and distributing datacenter assets." Which is a fancy way of saying it's a bunch of commodity PCs running Xen, attached to some f***ing big Juniper QFabric switch with some PHP scripts to let middle-managers bring up servers without knowing where they are. It's just that right now our stage in the hype cycle is the Peak of Inflated Expectations.
A number of smartphone providers have been talking about adding fingerprint readers to phones to make the security stronger. Over 40% of serious crime involves smart devices and half of those crimes are violent in some way, many at knife-point. Does anyone else worry that it won't take long for muggers to work out that if they take the phone they need to take your index finger too?
Given that much of the rise in crime in New York last year was due to people having the iOS devices stolen, how long will it be before muggings at knife-point typically also involve the thief stealing the owner's index finger too?
Those Porsches also use circular wheels rotating around a central axle. I mean geez, that's so 4th Millennium BC; even Mesopotamians sports carts were using those. Can't the Germans come up with something that's actually new?
The test that the competition regulators apply is "Will this reduce competition and consumer choice?" When Google bought Motorola Motorola was already a maker of Android phones and the immediate effect on the market was small. If Apple bought Nokia it would almost certainly want to kill Nokia's Windows phones, which would largely kill Windows Mobile, which would significantly reduce choice. There is no way that the EU would allow this and it seems unlikely that the US would allow it either (although that would be moot if the EU nixed it).
If you are going to check something at a checkpoint then it makes sense to stochastically sample with secondary checks to test your error rate. Apparently the TSA believe that there is a reason to limit the liquids through airport checkpoints and screen those liquids that they do allow through. Irrespective of if this is itself a rational position, if you believe that it is then it is also rational to check randomly sample liquids after the checkpoint.
It's important the appreciate that the Swedish arrest warrant for Svartholm isn't for file-sharing, it's for skipping bail and fleeing after his last round of appeals failed. Irrespective of the soundness of the original trail, the guy is a fugitive with a current conviction who's sentence has not been served. The charges he will face if he is caught now are far more serious than the ones he faced with Pirate Bay.
As someone who doesn't have US citizenship but who lives and works in the US, creating businesses that have hired hundreds of people (including plenty of H1-B holders) I have an alternate approach; I shall simply be avoiding Arizona as much as possible. I shall not be holding any group meetings there, I'll see what I can do to avoid conventions there or transfers through PHX and they can kiss goodbye to any prospect of my opening offices there. I'm probably too white to actually be harassed under this law but that doesn't make it any less disgusting to me.
Not only did Bruce demonstrate how useless the security at MSP is, but it's his home airport. My guess is that the TSA want all this extra 'security' to keep tabs on him.
Ultimately decisions about email encryption come down to what threats you think you might be protecting yourself against. I have a PGP key, and on occasion I use it to sign and decrypt emails when I think it matters. The rest of the time I send mail, over SSL, through my own mail server, which will use SMTP's 'startTLS' command whenever possible. Most people I know read their mail either using SSH on the machine that runs the mail server or over some SSL-protected IMAP or webmail interface. Thus, for most cases, the mail is encrypted in transit but never encrypted on the servers. If the threat is one of people eavesdropping then this keeps me safe; if the threat is one of hackers targeting one of the mail servers then it doesn't. Most of my mail doesn't warrant any more effort to achieve any more security.
Slashdot Mirror
Alternatively, you could just ditch the TV altogether and go read book.
Actually you get quite a bit of advantage. The instruction set for 64 bit mode is quite a bit more efficient even when you're not doing 64 bit arithmetic, not least because there are twice as many registers. Code that can be optimised using the SIMD instructions also gets twice as many SIMD registers, which can make a big difference for many graphics, signal processing and mathematical tasks. Most floating point code will be more efficient in 64 bit mode too.
Furthermore the 64 bit address space is useful even with only 1GB of RAM since it allows you to mmap files that are bigger than 4GB. Programs like MongoDB mmap their database files and when running on 32 bit processors MongoDB actually limits processes to a total of 2GB of data in the database, irrespective of how small the working set is. With a 64 bit address space the limit goes from below typical storage sizes to far beyond any plausible storage you'll ever connect to a Pi.
The idea that "the fact that only companies and labs have access to this technology can represent a threat" is patently absurd. Theano, Caffe and Torch are all open source and even Google has open-sourced its Tensor Flow platform which makes it easy to build new tools and run then, fast, on all the GPUs you can find. If you need to do this at scale and you're not the size of Google or IBM you can use Amazon's Machine Learning for AWS. There are many, many higher level toolkits out there that are available under licenses that are much less restrictive than GPLv3.
They didn't need to give him a non-exec position. All they needed to do was give him and his son a nice plane ride and some Superb Owl tickets.
Note that you can get the whole kit for $99 using discount code from Make Magazine.
How about nVidia's Jetson TK1 board? It has a great 192-core Kepler GPU, a nice quad-core ARM CPU, on-board gigabit ethernet, all the ports you're likely to need and comes pre-installed with Ubuntu Linux.
Even if you never make use of the new programming language it is almost always worth getting your head around a new way of thinking about problems. You may not ever need to write code in Lisp but understanding what a functional language is and isn't good for is helpful in other languages. If you're building flight control systems then Python might not be the language to do it in but getting deep into it and start understanding why you'd want a metaclass in the first place can help you stricture your code better. There aren't a ton of job openings for Erlang programmers but there are lots for people who understand High Availability and Disaster Recovery and the background knowledge will stand you in good stead.
In short, yes, it's worth learning all the languages your brain can handle. Even if you forget the details the concepts will stick around and help you later.
Ubiquiti Networks make some decent cameras with a whole bunch of decent features including fetching a snapshot with an HTTP request. They are designed to send their video output to their DVR software (which is actively supported on Linux) but in practice if all you need is to access still images over HTTP and video over RTSP then you can set up the control software on your laptop, fire it up once to configure the camera and then switch it off and the camera will continue to run without the DVR.
Of course, as other posters have pointed out, the right answer is to brew your own with a Raspberry Pi and a Pi Camera.
Firstly, if you can facilitate multi-way calling then it is clearly technically feasible to support a wire tap. Secondly, unlike many other snooping regulations, CALEA explicitly obliges telecommunications companies to modify their systems and equipment in order to facilitate "lawful access" (sic). Verizon are a telco, not an app company, so they are bound by CALEA in ways that people like Silent Circle or CellTrust are not.
If I was publishing an article talking about how huge numbers of eyeballs solves security problem I'm not sure that I'd choose to publish it the day after it was announced that the X window server code has had some serious security bugs for 25 years that have only just been discovered. Clearly open source code can have serious security holes that go unnoticed for a very long time.
Presumably when they OP author wrote "a relational database that is as capable as 'proprietary database engines at 1/10 the cost,' " what (s)he really meant was "a relational database (that is as capable as proprietary database engines) at 1/10 the cost".
That is of course "acre feet", not "ache feet". Bloody auto-correct!
It's used by the same sort of people who measure their irrigation water in ache feet.
This begs the question whether mining for BitCoins is more damaging to the environment than mining for precious metals, for a given value of return. The EPA emissions factor for electricity is about 0.69 tons of CO2 per megawatt hour, so producing the electricity used by this datacenter is, on average, dumping into the atmosphere 331 tons of CO2 per day or about 120,000 tons of CO2 per year. While there are many other forms of environmental damage from gold mining, a quick search suggest that the greenhouse emissions from gold extraction run to about 11.5 tons of CO2 equivalent per kg of Gold. At this rate 120,000 tons of CO2 yields of 10.5 tons of gold, worth nearly $500 million at today's price. Will this datacentre yield more than half a billion dollars worth of bit coins each year?
I have four RPi boards. One monitors my UPSs, cleanly suspending my server when the power goes out and sending wake-on-LAN massages to it when the power comes back up so that the UPS only needs to drive my switch and AP, one has a camera board and does motion detection to spot people coming into my office, one is currently operating as a Bluetooth LE beacon for testing the new iOS iBeacon stuff and one is just for tinkering. Most of these have a few other services running on them too (two have I2C thermometers on them).
I see a lot of negative comments about the Pi being underpowered. Perhaps if what you want to do is run FPS games or you are trying to run Big Data analytics then this is true but it's plenty powerful enough for a whole host of service tasks. It's not that many years ago that the Pi's level of power would have been considered a high-end desktop configuration. The purpose of the device is to give kids a low-cost entry into programming and it does just that. On top, at $25 for a Model A its fine to put in 'dangerous' places where something bad might happen to it (like outdoors, driving the sensors and servos for my Halloween decorations). No, I don't have my MongoDB server on a Raspberry Pi, but for many many projects they are just about perfect.
"More than an actual technology, SDDN is the culmination of many other efforts at abstracting, consolidating, managing, provisioning, load balancing and distributing datacenter assets." Which is a fancy way of saying it's a bunch of commodity PCs running Xen, attached to some f***ing big Juniper QFabric switch with some PHP scripts to let middle-managers bring up servers without knowing where they are. It's just that right now our stage in the hype cycle is the Peak of Inflated Expectations.
A number of smartphone providers have been talking about adding fingerprint readers to phones to make the security stronger. Over 40% of serious crime involves smart devices and half of those crimes are violent in some way, many at knife-point. Does anyone else worry that it won't take long for muggers to work out that if they take the phone they need to take your index finger too?
Given that much of the rise in crime in New York last year was due to people having the iOS devices stolen, how long will it be before muggings at knife-point typically also involve the thief stealing the owner's index finger too?
Those Porsches also use circular wheels rotating around a central axle. I mean geez, that's so 4th Millennium BC; even Mesopotamians sports carts were using those. Can't the Germans come up with something that's actually new?
The test that the competition regulators apply is "Will this reduce competition and consumer choice?" When Google bought Motorola Motorola was already a maker of Android phones and the immediate effect on the market was small. If Apple bought Nokia it would almost certainly want to kill Nokia's Windows phones, which would largely kill Windows Mobile, which would significantly reduce choice. There is no way that the EU would allow this and it seems unlikely that the US would allow it either (although that would be moot if the EU nixed it).
If you are going to check something at a checkpoint then it makes sense to stochastically sample with secondary checks to test your error rate. Apparently the TSA believe that there is a reason to limit the liquids through airport checkpoints and screen those liquids that they do allow through. Irrespective of if this is itself a rational position, if you believe that it is then it is also rational to check randomly sample liquids after the checkpoint.
It's important the appreciate that the Swedish arrest warrant for Svartholm isn't for file-sharing, it's for skipping bail and fleeing after his last round of appeals failed. Irrespective of the soundness of the original trail, the guy is a fugitive with a current conviction who's sentence has not been served. The charges he will face if he is caught now are far more serious than the ones he faced with Pirate Bay.
As someone who doesn't have US citizenship but who lives and works in the US, creating businesses that have hired hundreds of people (including plenty of H1-B holders) I have an alternate approach; I shall simply be avoiding Arizona as much as possible. I shall not be holding any group meetings there, I'll see what I can do to avoid conventions there or transfers through PHX and they can kiss goodbye to any prospect of my opening offices there. I'm probably too white to actually be harassed under this law but that doesn't make it any less disgusting to me.
Not only did Bruce demonstrate how useless the security at MSP is, but it's his home airport. My guess is that the TSA want all this extra 'security' to keep tabs on him.
Ultimately decisions about email encryption come down to what threats you think you might be protecting yourself against. I have a PGP key, and on occasion I use it to sign and decrypt emails when I think it matters. The rest of the time I send mail, over SSL, through my own mail server, which will use SMTP's 'startTLS' command whenever possible. Most people I know read their mail either using SSH on the machine that runs the mail server or over some SSL-protected IMAP or webmail interface. Thus, for most cases, the mail is encrypted in transit but never encrypted on the servers. If the threat is one of people eavesdropping then this keeps me safe; if the threat is one of hackers targeting one of the mail servers then it doesn't. Most of my mail doesn't warrant any more effort to achieve any more security.