This thing has more than enough power to run X, so don't bother with the text-mode distro.
The problem you'll run into is not processing power, but connectivity. No SATA connector, for example. As long as you're OK with serving files from an external USB 2.0 drive or the SD card, you'll be OK.
Considering the low-cost target, I doubt many features will be added. Most improvements will simply be implemented via USB or extension boards anyway.
That said, here's my wish list of 2nd generation improvements, which are mostly just incremental improvements and rely heavily on Broadcom:
1) More RAM 2) Upgraded CPU speed (a given, really) 3) More hardware decoding options for video, audio, etc. 4) Even smaller form factor
And although these could technically be done as an external extension, I would like to see these features built-in for convenience: 5) Wi-fi (For the Model A especially) 6) Optional power-over-ethernet (Model B only, of course) 7) Optional battery power (on-board power management IC and a simple connector would be enough)
Every other feature I can think of can be done as an external board or USB. That said, I would imagine that if somebody made a feature that lots of people use and want, you might see it integrated onto the 2nd gen Raspberry Pi.
Any time you see a CPU speed listed for a computer, you're actually seeing information about a quartz crystal. For example, the Raspberry Pi has a 700 MHz processor, which really means that there's a stable clock source timed to oscillate 700 million times per second. At the core of this clock source is a quartz crystal to serve as a time reference.
It's rare to see the oscillation frequency of a crystal to be used directly as a clock frequency nowadays, so don't assume that the crystal in the RP is 700 MHz.
There's already plenty of replies to explain this, but I thought I'd add my thoughts anyway.
I recently worked with code that had this exact problem.
The first problem was that keeping the system running had an incredibly high priority, as specified by the contract with our customer. The second problem was sprinkling catch(...) throughout the code whenever there was a crash (this was C++), which then logged a message and rethrew the exception, rarely trying to do anything about the problem. The third problem was with Microsoft's compiler, which will turn segmentation faults into null pointer exceptions if you want.
Thus, we had null pointers in our code, and instead of crashing where these were found, we threw exceptions that were only caught and logged in random places. Finding the cause of a segmentation fault is hard enough if you just core dump on the spot -- at least you know what pointer had the null value and where it was accessed, but you may not know how that pointer became null in the first place -- but imagine trying to find the cause when your only indication is misdirection and redirection in log files, which at most give a partial stack trace.
So, how did we solve this dilemma? We added more catch(...) statements and log messages everywhere, of course! Because "we can't let the system crash" said my managers.
And, in the meantime, the system keeps chugging along, getting more an more flaky over time, with very bizarre behaviors and inexplicable data values. Until it finally crashes anyway, that is.
No, you have that backwards. Apple did nothing to make video chat "just work."
The real "innovation" was having a platform so ubiquitous that you can make video chat and other features easy to use.
That's about the only thing that everybody on/. can agree on: regardless of what you think about Apple and its devices, you can't deny that having a ubiquitous platform is key to Apple's continuing success, and let's them recreate old ideas into things that people actually use.
What gets really annoying is to see the power that Apple has simply because it already has power. Kinda like looking at the "1%" and realizing that they have wealth simply because they have wealth, as most of their money comes from investing their money, not from hard work and innovation like "the rest of us." That perception of both Apple and the 1% probably isn't true (Apple really does make good products and the 1% usually work in positions using the talents that got them into the top 1% to begin with), but there you go.
Indeed, the limits in small claims court seem to be in the range of $2,500 to $25,000 (depending on the state). Then again, the RIAA has been awarded roughly this amount as the per-song damages in cases they bring against individuals... that makes it feel about right in my mind.
So, it's not enough to bankrupt UMG by any means, but it's still just and appropriate, given the actions they would take if the tables were turned.
Technically, you're right. The actual charge of impeachment was for perjury and obstruction of justice. But there's more to the story than that.
What got the public up in arms to begin with was the affair. Gingrich took that fact and ran with it, leading the charge against Clinton. When Clinton was caught lying, that's when the charges were brought against him.
Mostly, the whole thing was a political maneuver of Republicans (lead by Gingrich) against Clinton. The action was so unpopular that Gingrich eventually resigned. But underlying it all was the affair, which makes Gingrich a hypocrite by any measure.
Thank you! I was browsing through Linus's patch and couldn't make heads or tails of it.
That and I couldn't help noticing that he got rid of a bunch of goto statements while he was at it. At least these gotos were actually used for error handling...
(I'm going to make big generalizations, assumptions, and approximations here. Bear with me.)
Let's say that your stopping distance at 40 mph is 75 ft, and your reaction time is 1.5 sec, or about 90 ft. So, if anything happens within 165 ft, you won't be able to come to a complete stop.
However, let's say your car can go 0-60 in 6 seconds. That's an average acceleration of around 10mph/s, or about 15 ft/s^2. Instead of trying to brake, let's say you hit the accelerator. By the time you react, you're within 75 ft of the obstacle. Had you done nothing, you would have hit that obstacle in 1.28 s, but because you accelerate, you'll be 12 feet further than if you had done nothing. That's about one car length.
Keep in mind that this is the best-case scenario: braking is just barely unable to stop you before you reach the intersection. Any more reaction time and braking would be a clear winner. Even then, the most you can hope for by accelerating instead is to advance your car about one car length. Less reaction time makes things worse.
Is one car length enough? Yes, that's barely enough to avoid an accident. But braking has twice the ability to change velocity than acceleration, and has the advantage of giving you more time to react and maneuver. And then, if you still can't avoid a collision, you're going slower, instead of faster.
I'm fully prepared to admit that accelerating might be the right choice to avoid an accident. The problem is that most any scenario where it might be better requires faster thinking than most people are capable of.
(p.s. This being Slashdot, someone will come along and point out some terrible mistake in my math. Never do math on Slashdot!)
The iPhone 4 battery is 5.25 Whr at 3.7V. The Raspberry Pi draws 300mA peak. Let's be pessimistic and assume that's the constant draw for XBMC video decoding. At 5V, that's 1.5W, which will give about 3.5 hrs of battery life. I'd bet you'd actually see closer to 4 hours in real-life tests. SD or 720p video would probably see even lower power consumption.
So, how does the iPhone do? Real-world examinations of Apple's claim of "up to 10 hours" for playing video are hard to find. Apple's tests were done with a video from iTunes: 640×480 resolution, so this is hardly a fair comparison. PCWorld found the life to be about 6 hours for 720p video, but that includes the power from the display (at full brightness) and wifi. (The iPhone has had battery life issues because of an OS problem, just to complicate things a bit).
So, the Raspberry Pi compares pretty well. I would love to see someone make a fair test here: play an HD video over HDMI for both devices and measure the power consumption. The Raspberry Pi will draw more current, probably, but not nearly as much as you might think.
One thing we do know is the force from any net charge on the Earth is completely overwhelmed by gravity, the weakest of the fundamental forces. Remember, electromagnetism is about 10^36 times stronger than gravity. This would make detection very hard, indeed, but also suggests that any net charge is very small.
Now, I'm no cosmologist, but my understanding is that current theories require that the universe itself be electrically neutral (but I don't know why this has to be true, personally). So, for the Earth to have any net charge, you have to postulate some mechanism for the charges to get separated by cosmological distances.
I presume similar provisions have been made for boundaries defined by rivers? At first, a river seems to be a natural border... this side is ours, that side is yours. But rivers have a tendency to change over time, sometimes very dramatically.
Just to be clear: my point wasn't concerning the availability of games or the ability to play games on the device. My point was more on the "feel" of the device.
If you were to put a simple B&W, text-only display on a modern blackberry... no, they won't sell well, even if 100% of their functionality is still there.
That said, I was probably a bit too absolute in my original post, and blackberries are an example of that -- instead of creating a "game-like" device, RIM was a business tool that many professionals found to be invaluable. Still is, for that matter.
I expect Apple to keep taking the hits for the next few years while Microsoft continues playing "nice" until a smaller company, with arrogant stupidity, decides to poke the sleeping lion. After which Microsoft will slaughter them and hang them out to dry before going back to playing nice with everybody else
And that smaller company would be Barnes and Noble. B&N is standing up to Microsoft quite well so far, but things are still in the early stages.
Motorola also stood up to MS... and only one of the patents in the lawsuit actually stuck (and it was a pretty minor one at that), the rest were shown to be the rubbish that they were. Once Motorola officially becomes part of Google I would expect things to get even more interesting.
Yeah, the refresh rate of the Mirasol still limits it to the ebook reader market. Current-generation is 15 fps, which is much better than current e-ink, but still not sufficient for general purpose use. I believe that a 30 fps version has been demonstrated, so the next generation devices might have a wider market. 60+ fps is probably only theoretically possible, at the moment.
I think the Kindle Fire would have been a prime candidate for the Pixel Qi, except for the $200 price point. People were already comparing it to the iPad, so what difference would it make? It might have even helped on the comparison, to solidify the Fire as an e-book reader. But I don't think there's any way they would be selling it for $200, even if they expect the hardware to be a loss leader.
Oh well... maybe the Fire 2 will have a Mirasol or Pixel Qi display.
The screens themselves can probably be paired with any touchscreen and controller. See the Notion Ink Adam for an example of a capacitive touchscreen matched with the Pixel Qi.
Re: the Dynabook, purely utilitarian devices will not sell... as someone once put it, the more game-like a product is, the more people will want it. (Paraphrasing, sorry). Relevantly, this was in reference to the introduction of color computer monitors in the '80s. Likewise, a Dynabook-like tablet will only be successful if it at least feels like a game, which means a color display.
I actually took an elective called "The Future of Technology" for my undergrad degree.
The whole field is a bunch of BS and guesswork. It makes Psychoanalysis look like a hard science. I would sooner believe someone who says the world will end on December 12, 2012, than any "futurologist."
And the reason is exactly as you said: all they want is for people to believe what they say, and pay them money to say more.
From my own observations, there seem to be two types of "futurologists:" 1) Those who make grandiose and sensational predictions, from the perspective of "wouldn't it be cool if..." or even "the worst thing I can think of is...." Their audience is anybody who will listen and pay them to write an article. 2) Those who try to predict the next short-term trend before it happens. Their audience generally consists of technology and investment companies, who pay them directly for their insight as a consultant.
My favorite example of the first type is a committee who claimed the USA would wake up one day in the near future, unable to afford to feed itself. Europe would have to step in and bail them out, delivering subsidized food. In reality, recent events have shown that the global economy is so tied together that if the USA ever got in that much trouble, there's no way that Europe would have the resources to help on that scale.
The second type reminds me of a Pinky and the Brain episode where the Brain predicts the next modern art craze will be paintings of doughnuts. Someone really does stand a chance at making a fortune if they can correctly predict the near-term trends in technology. (Or if they can drive the trend how they want, see also: Steve Jobs). Predicting those trends is very risky and expensive, however, and I don't even want to think about the cash that's been lost trying to catch the next big wave before it even forms, only to have it never materialize.
Most predictions on the future of technology are absolutely and amazingly incorrect... the rest are just lucky guesses.
Speaking from current experience (porting a large codebase which includes several languages, including Java and C++), this is generally true. But that doesn't stop people from doing really stupid things.
For example, previous developers on the code I'm working with decided to use Runtime.exec() everywhere. Need to copy a file? Try running 'cp'... and be sure to hard-code the path where you think it will exist. Need a directory listing? Call 'ls'... and be sure to rely on a specific output format for error messages. I have many more examples.
This isn't a Java problem, because Java provides a native mechanism that isn't being used. And, to give credit to Java, at least it's generally pretty obvious what will break on other platforms, and most of that will be encapsulated into JNI or the like.
Java is "write once, run anywhere" if you think about cross-platform compatibility as you're coding. But, the same could be said for most languages, including C and C++ (with Java being a bit better).
Slashdot Mirror
Perfect for compound fractures, then. Heals both the skin and the bone.
/dev/urandom
This thing has more than enough power to run X, so don't bother with the text-mode distro.
The problem you'll run into is not processing power, but connectivity. No SATA connector, for example. As long as you're OK with serving files from an external USB 2.0 drive or the SD card, you'll be OK.
http://elinux.org/Rpi_Hardware
Considering the low-cost target, I doubt many features will be added. Most improvements will simply be implemented via USB or extension boards anyway.
That said, here's my wish list of 2nd generation improvements, which are mostly just incremental improvements and rely heavily on Broadcom:
1) More RAM
2) Upgraded CPU speed (a given, really)
3) More hardware decoding options for video, audio, etc.
4) Even smaller form factor
And although these could technically be done as an external extension, I would like to see these features built-in for convenience:
5) Wi-fi (For the Model A especially)
6) Optional power-over-ethernet (Model B only, of course)
7) Optional battery power (on-board power management IC and a simple connector would be enough)
Every other feature I can think of can be done as an external board or USB. That said, I would imagine that if somebody made a feature that lots of people use and want, you might see it integrated onto the 2nd gen Raspberry Pi.
Any time you see a CPU speed listed for a computer, you're actually seeing information about a quartz crystal. For example, the Raspberry Pi has a 700 MHz processor, which really means that there's a stable clock source timed to oscillate 700 million times per second. At the core of this clock source is a quartz crystal to serve as a time reference.
It's rare to see the oscillation frequency of a crystal to be used directly as a clock frequency nowadays, so don't assume that the crystal in the RP is 700 MHz.
There's already plenty of replies to explain this, but I thought I'd add my thoughts anyway.
I recently worked with code that had this exact problem.
The first problem was that keeping the system running had an incredibly high priority, as specified by the contract with our customer. The second problem was sprinkling catch(...) throughout the code whenever there was a crash (this was C++), which then logged a message and rethrew the exception, rarely trying to do anything about the problem. The third problem was with Microsoft's compiler, which will turn segmentation faults into null pointer exceptions if you want.
Thus, we had null pointers in our code, and instead of crashing where these were found, we threw exceptions that were only caught and logged in random places. Finding the cause of a segmentation fault is hard enough if you just core dump on the spot -- at least you know what pointer had the null value and where it was accessed, but you may not know how that pointer became null in the first place -- but imagine trying to find the cause when your only indication is misdirection and redirection in log files, which at most give a partial stack trace.
So, how did we solve this dilemma? We added more catch(...) statements and log messages everywhere, of course! Because "we can't let the system crash" said my managers.
And, in the meantime, the system keeps chugging along, getting more an more flaky over time, with very bizarre behaviors and inexplicable data values. Until it finally crashes anyway, that is.
The picture that came to my mind was a scene from The Incredibles:
"Edna. Mode. .... And guest."
No, you have that backwards. Apple did nothing to make video chat "just work."
The real "innovation" was having a platform so ubiquitous that you can make video chat and other features easy to use.
That's about the only thing that everybody on /. can agree on: regardless of what you think about Apple and its devices, you can't deny that having a ubiquitous platform is key to Apple's continuing success, and let's them recreate old ideas into things that people actually use.
What gets really annoying is to see the power that Apple has simply because it already has power. Kinda like looking at the "1%" and realizing that they have wealth simply because they have wealth, as most of their money comes from investing their money, not from hard work and innovation like "the rest of us." That perception of both Apple and the 1% probably isn't true (Apple really does make good products and the 1% usually work in positions using the talents that got them into the top 1% to begin with), but there you go.
Indeed, the limits in small claims court seem to be in the range of $2,500 to $25,000 (depending on the state). Then again, the RIAA has been awarded roughly this amount as the per-song damages in cases they bring against individuals ... that makes it feel about right in my mind.
So, it's not enough to bankrupt UMG by any means, but it's still just and appropriate, given the actions they would take if the tables were turned.
Technically, you're right. The actual charge of impeachment was for perjury and obstruction of justice. But there's more to the story than that.
What got the public up in arms to begin with was the affair. Gingrich took that fact and ran with it, leading the charge against Clinton. When Clinton was caught lying, that's when the charges were brought against him.
Mostly, the whole thing was a political maneuver of Republicans (lead by Gingrich) against Clinton. The action was so unpopular that Gingrich eventually resigned. But underlying it all was the affair, which makes Gingrich a hypocrite by any measure.
I can't figure out why anyone would vote for that guy.
Because he's "not Romney." At least, that's what the media seems to think.
See also: G.O. 1984.
Thank you! I was browsing through Linus's patch and couldn't make heads or tails of it.
That and I couldn't help noticing that he got rid of a bunch of goto statements while he was at it. At least these gotos were actually used for error handling...
However, the SD card won't stick out that much on production boards. They had supplier issues for the first batch and used these as a workaround.
That said, you're right that only the PCB fits the "credit card" dimensions.
Math time!
(I'm going to make big generalizations, assumptions, and approximations here. Bear with me.)
Let's say that your stopping distance at 40 mph is 75 ft, and your reaction time is 1.5 sec, or about 90 ft. So, if anything happens within 165 ft, you won't be able to come to a complete stop.
However, let's say your car can go 0-60 in 6 seconds. That's an average acceleration of around 10mph/s, or about 15 ft/s^2. Instead of trying to brake, let's say you hit the accelerator. By the time you react, you're within 75 ft of the obstacle. Had you done nothing, you would have hit that obstacle in 1.28 s, but because you accelerate, you'll be 12 feet further than if you had done nothing. That's about one car length.
Keep in mind that this is the best-case scenario: braking is just barely unable to stop you before you reach the intersection. Any more reaction time and braking would be a clear winner. Even then, the most you can hope for by accelerating instead is to advance your car about one car length. Less reaction time makes things worse.
Is one car length enough? Yes, that's barely enough to avoid an accident. But braking has twice the ability to change velocity than acceleration, and has the advantage of giving you more time to react and maneuver. And then, if you still can't avoid a collision, you're going slower, instead of faster.
I'm fully prepared to admit that accelerating might be the right choice to avoid an accident. The problem is that most any scenario where it might be better requires faster thinking than most people are capable of.
(p.s. This being Slashdot, someone will come along and point out some terrible mistake in my math. Never do math on Slashdot!)
Back of the envelope calculations:
The iPhone 4 battery is 5.25 Whr at 3.7V. The Raspberry Pi draws 300mA peak. Let's be pessimistic and assume that's the constant draw for XBMC video decoding. At 5V, that's 1.5W, which will give about 3.5 hrs of battery life. I'd bet you'd actually see closer to 4 hours in real-life tests. SD or 720p video would probably see even lower power consumption.
So, how does the iPhone do? Real-world examinations of Apple's claim of "up to 10 hours" for playing video are hard to find. Apple's tests were done with a video from iTunes: 640×480 resolution, so this is hardly a fair comparison. PCWorld found the life to be about 6 hours for 720p video, but that includes the power from the display (at full brightness) and wifi. (The iPhone has had battery life issues because of an OS problem, just to complicate things a bit).
So, the Raspberry Pi compares pretty well. I would love to see someone make a fair test here: play an HD video over HDMI for both devices and measure the power consumption. The Raspberry Pi will draw more current, probably, but not nearly as much as you might think.
One thing we do know is the force from any net charge on the Earth is completely overwhelmed by gravity, the weakest of the fundamental forces. Remember, electromagnetism is about 10^36 times stronger than gravity. This would make detection very hard, indeed, but also suggests that any net charge is very small.
Now, I'm no cosmologist, but my understanding is that current theories require that the universe itself be electrically neutral (but I don't know why this has to be true, personally). So, for the Earth to have any net charge, you have to postulate some mechanism for the charges to get separated by cosmological distances.
I presume similar provisions have been made for boundaries defined by rivers? At first, a river seems to be a natural border ... this side is ours, that side is yours. But rivers have a tendency to change over time, sometimes very dramatically.
Just to be clear: my point wasn't concerning the availability of games or the ability to play games on the device. My point was more on the "feel" of the device.
If you were to put a simple B&W, text-only display on a modern blackberry ... no, they won't sell well, even if 100% of their functionality is still there.
That said, I was probably a bit too absolute in my original post, and blackberries are an example of that -- instead of creating a "game-like" device, RIM was a business tool that many professionals found to be invaluable. Still is, for that matter.
I expect Apple to keep taking the hits for the next few years while Microsoft continues playing "nice" until a smaller company, with arrogant stupidity, decides to poke the sleeping lion. After which Microsoft will slaughter them and hang them out to dry before going back to playing nice with everybody else
And that smaller company would be Barnes and Noble. B&N is standing up to Microsoft quite well so far, but things are still in the early stages.
Motorola also stood up to MS ... and only one of the patents in the lawsuit actually stuck (and it was a pretty minor one at that), the rest were shown to be the rubbish that they were. Once Motorola officially becomes part of Google I would expect things to get even more interesting.
Yeah, the refresh rate of the Mirasol still limits it to the ebook reader market. Current-generation is 15 fps, which is much better than current e-ink, but still not sufficient for general purpose use. I believe that a 30 fps version has been demonstrated, so the next generation devices might have a wider market. 60+ fps is probably only theoretically possible, at the moment.
I think the Kindle Fire would have been a prime candidate for the Pixel Qi, except for the $200 price point. People were already comparing it to the iPad, so what difference would it make? It might have even helped on the comparison, to solidify the Fire as an e-book reader. But I don't think there's any way they would be selling it for $200, even if they expect the hardware to be a loss leader.
Oh well ... maybe the Fire 2 will have a Mirasol or Pixel Qi display.
The screens themselves can probably be paired with any touchscreen and controller. See the Notion Ink Adam for an example of a capacitive touchscreen matched with the Pixel Qi.
Re: the Dynabook, purely utilitarian devices will not sell ... as someone once put it, the more game-like a product is, the more people will want it. (Paraphrasing, sorry). Relevantly, this was in reference to the introduction of color computer monitors in the '80s. Likewise, a Dynabook-like tablet will only be successful if it at least feels like a game, which means a color display.
I actually took an elective called "The Future of Technology" for my undergrad degree.
The whole field is a bunch of BS and guesswork. It makes Psychoanalysis look like a hard science. I would sooner believe someone who says the world will end on December 12, 2012, than any "futurologist."
And the reason is exactly as you said: all they want is for people to believe what they say, and pay them money to say more.
From my own observations, there seem to be two types of "futurologists:"
1) Those who make grandiose and sensational predictions, from the perspective of "wouldn't it be cool if..." or even "the worst thing I can think of is...." Their audience is anybody who will listen and pay them to write an article.
2) Those who try to predict the next short-term trend before it happens. Their audience generally consists of technology and investment companies, who pay them directly for their insight as a consultant.
My favorite example of the first type is a committee who claimed the USA would wake up one day in the near future, unable to afford to feed itself. Europe would have to step in and bail them out, delivering subsidized food. In reality, recent events have shown that the global economy is so tied together that if the USA ever got in that much trouble, there's no way that Europe would have the resources to help on that scale.
The second type reminds me of a Pinky and the Brain episode where the Brain predicts the next modern art craze will be paintings of doughnuts. Someone really does stand a chance at making a fortune if they can correctly predict the near-term trends in technology. (Or if they can drive the trend how they want, see also: Steve Jobs). Predicting those trends is very risky and expensive, however, and I don't even want to think about the cash that's been lost trying to catch the next big wave before it even forms, only to have it never materialize.
Most predictions on the future of technology are absolutely and amazingly incorrect ... the rest are just lucky guesses.
Speaking from current experience (porting a large codebase which includes several languages, including Java and C++), this is generally true. But that doesn't stop people from doing really stupid things.
For example, previous developers on the code I'm working with decided to use Runtime.exec() everywhere. Need to copy a file? Try running 'cp' ... and be sure to hard-code the path where you think it will exist. Need a directory listing? Call 'ls' ... and be sure to rely on a specific output format for error messages. I have many more examples.
This isn't a Java problem, because Java provides a native mechanism that isn't being used. And, to give credit to Java, at least it's generally pretty obvious what will break on other platforms, and most of that will be encapsulated into JNI or the like.
Java is "write once, run anywhere" if you think about cross-platform compatibility as you're coding. But, the same could be said for most languages, including C and C++ (with Java being a bit better).