Firstly it's difficult to prove who is intentionally disrupting traffic and who is just caught up in the disruption. Especially if the disruption strategy is to focus a large number of vehicles on a small area but otherwise drive normally. Secondly if the roads are gridlocked getting the cop cars and tow trucks in and out is going to be difficult.
I remember reading an EPE article that claimed tha tthe difference in sound between valve amps and most modern transistor amps was caused by driving the speaker differently with the valve amps being effectively controlled current sources while the modern transistor amps are effectively controlled voltage sources. Thus causing the speaker to respond differently.
Said article also gave a design for a FET based amp which used the topology of and supposedly sounded like a valve amp.
I suspect they do have a backup soloution (paper based or otherwise) but using it results in a much lower capacity than the main system. After all why would you build a new computer system if their wasn't a significant benefit to doing so?
If you have an incident that reduces capacity significantly in a system that is close to capacity then you have to prioritise. The priority for ATC is going to be to get planes already in the air safely onto the ground before they run out of fuel. Planes on the ground can wait.
I don't see anything in his post saying he was dumb enough to try and power the whole rack with such a power cord.
In my experiance (living in the UK) most IEC cords are 0.75mm. 0.75mm flex is supposed to be good for 6A (and is probablly in practice good for substantially more, the current ratings on the small flex sizes are very conservative). That should be sufficient for most servers.
I do find it very misleading that the moulded IEC connectors seem to always be stamped with 10A even when connected to 0.75mm flex but if his flex melted either it was a very big server or the 0.75mm printed on the flex was a lie.
The JPEG originals from the camera are generally far too large for reasonable web use anyway. Nearly all photos you see on the web will have been downscaled and recompressed since they left the camera.
the difference is there is a virtually unlimited number of function names most of which do not have a pre-existing meaning. There is a very limited number of overloadable operators all of which do have a pre-existing meaning. So the temptation to abuse is much higher and ideas of what constitutes abuse much fuzzier..
Update: now i've managed to get on odriod's site and get some answers to the question. It seems they now offer a $9 shipping option for low value orders and they are using the postal service which tends to have marginally lower brokerage charges than couriurs. That makes the updated estimate.
The final cost will depend on both where you live and what if any distributor arrangements odriod set up. I would be very wary of considering a board cheaper based on headline price alone as the difference between headline price and final cost varies massively.
AIUI ipods are meant to play two classes of media files, non-drm files and files protected with apple drm*.
AIUI circumventing protections on access to a copyrighted work is illegal in the US but that isn't what real did. They found a way to put their own DRM media on ipods without undrming it by convincing the ipod that the file was using apple drm.
* Yes I know apple no longer uses drm for music sold on itunes but afaict they still use it for videos, previously purchased music that was not upgraded and some audiobooks. At one stage they were also still using it for music giveaways after they stopped using it for music sales, not sure if that is still the case.
The big question with this is what will the actual cost be? last time I bought an odriod product I had to pay the price of a fairly steep delivery charge to odriod and the dreaded courier VAT collection fee and of course the VAT itself. I don't remember exactly what it was in my case but according to http://forum.odroid.com/viewto... the final cost of of an "$89" U2 plus a "$9" HDMI cable shipped to the UK was £108.83.
If we assume the same delivery charge for the new board as for the U2* and that it still ships from outside the EU on a regular courier service**, and that they tell the truth on the customs form then the final price will be about $35 (board) + $30 (shipping) + ~£10.5 (brokerage) + ~£9.5 VAT = ~£68.
Whereas with a rasperry pi I pay the distributors listed price for the Pi itself (which is marginally higher than the raspberry pi foundations nominal price) and the VAT (delivery is "free") final cost of a "$35" raspberry pi B+ is £27.44.
Of course i'm in the UK which skews things a bit in the Pi's favour, other places the calculation may work out different but still comparing nominal prices is only a very rough way to compare the cost of SBCs.
* I can't check easilly if odriod's delivery charges have changed or even if this board is actually available yet because odriod's main site seems to have been/.ed ** Some couriers are now offering all-inclusive services where the seller handles collecting the VAT and brokerage costs are consolidated across multiple shipments, but in my experiance only a handful of vendors use them.
Banana Pi is only dual core 1GHz, This is quad core 1.5GHz (Though A5 is slower than A7 clock for clock but at least if we go by arm's DMIPs figures a 1.5GHz A5 core should outperform a 1GHz A7 core), Also note that while the banana Pi looks sorta like a Pi model B it is actually a bit larger, so it won't fit in Pi cases.
Do injections in the eyball cause sigificant damage that would prevent them being used on a regular basis or are you just talking about the "what the patient will put up with" factor?
I would think that if the alternative was blindness patients would put up with weekly injections into the eyeball. From some googling it seems there are already eyeball injected drugs that are given every two months.
register size == memory address == unit of memory that can be operated on.
Every modern processor i'm aware of can operate on memory in 8-bit units despite having 32-bit or 64-bit register sizes and 32-bit or 64-bit memory addresses. Older processors with 8 and 16 bit register sizes typically had memory address sizes larger than their register size.
If multiplication of your objects doesn't make sense, then why on earth would you overload the * operator?
If bit shifting your object doesn't make sense then why on earth would you overload the << operator
I can't really make up my mind on operator overloading. On the one hand it can make custom types far more pleasant to use. On the other hand because operators are more concise than function calls and you can't add new ones it's very tempting to abuse operators to mean something other than their original meaning. Heck it's so tempting that even the authors of the standard library did it.
On the mac while "classic" mode is gone "carbon" is still there and was explicitly intended to allow porting of code from classic macos. I'd be surprised if there wasn't some code that had been written for classic macos still in there somewhere.
Similarly win32 was designed as a 32-bit variant of win16 and i'd be very surprised if there wasn't still some old code hanging arround somewhere.
While compression may help prevent some types of attack (in particular known plaintext attacks) it also creates new avenues of attack. In particular it makes the message length dependent on the message content.
For something like human written pgp emails this is not a massive deal,the length variability of human written messages and the low message rate mean that the chances of useful information being leaked this way are negligable.
For things like VOIP this can cause information leakage even to a passive attacker as what people say will impact the compressibility of the data. http://www.esecurityplanet.com...
For things like https the message rate is too low and the content too variable for a passive attacker to stand much chance of exploiting the messsage length side channel but an active attacker can stimulate traffic with known characteristics and partially known content to effectively exploit the channels.
I would argue that it is better to learn to write decent code than to let the language protect you from the effects of bad code.
I would argue that there is a middle ground. C++ still lets you do the low level stuff when you need to but also provides higher level structures that when used responsiblly make code clearer and less error prone.
I tried using a VGA cable and also a DVI to HDMI cable.
On VGA I couldn't get it to go above 1024x768, when I tried forcing a 1366x768 output on the computer it refused to display it (which I found very surprising as there really shouldn't be any significant differences in the sync timings between the two). I can't remember exactly what the results with the DVI to HDMI cable were but I remember having even more trouble..
I admittedly didn't have a video card to hand with an actual HDMI output to try.
If you get a 1080p TV that can do 60Hz, then you've got something that's pretty much the same as a monitor
Unfortunately in some cases they then go and ruin it by putting inappropriate signal processing in the path which can't be turned off. I have a hanspree TV which has the "FULL HD" badge on it and supposedly has a native resoloution of 1920x1200 but trying to use it as a computer monitor (regardless of input resoloution) results in a blurry mess.
My parents also have a smaller TV (think it's a sharp but I can't remember for sure) that supposedly has a native resoloution of 1366x768 but I never managed to get it to accept a signal at that resoloution.
Maybe I've just been unlucky or buying too cheap but I don't think I would buy a TV for use as a monitor again unless I could try before buying.
thousands of dollars per kilowatt hour in some cases
Reassembling a raid from drives taken at multiple different times will likely be harder than reassembling from disks taken at the same time because most raid implementations are designed to stop you doing it by accident but with sufficient determination it should be possible.
Which would be fine if they stuck to doing it in their own languages, unfortunately they cause ambiguity and confusion by writing english text but failling to follow english numeric conventions.
Not sure I would agree with "much of the world" either, from what I can tell it seems to be a mainland european thing.
if these open source projects are going to accept corporate sponsorship, they must do that corporation's bidding.
The people and entities who signed the sponsorship contract must do what they contractually agreed to do (which may be virtually nothing or it may be very specific depends what was in the contract).
Other people aren't bound by that though. Most contributors to open source projects do not have any contract with or obligation to the operators of the project. If they (or their employers if relavent) decide they would rather put their effort into a fork then they are perfectly entitled to do so.
Slashdot Mirror
Easier said than done.
Firstly it's difficult to prove who is intentionally disrupting traffic and who is just caught up in the disruption. Especially if the disruption strategy is to focus a large number of vehicles on a small area but otherwise drive normally. Secondly if the roads are gridlocked getting the cop cars and tow trucks in and out is going to be difficult.
I remember reading an EPE article that claimed tha tthe difference in sound between valve amps and most modern transistor amps was caused by driving the speaker differently with the valve amps being effectively controlled current sources while the modern transistor amps are effectively controlled voltage sources. Thus causing the speaker to respond differently.
Said article also gave a design for a FET based amp which used the topology of and supposedly sounded like a valve amp.
I suspect they do have a backup soloution (paper based or otherwise) but using it results in a much lower capacity than the main system. After all why would you build a new computer system if their wasn't a significant benefit to doing so?
If you have an incident that reduces capacity significantly in a system that is close to capacity then you have to prioritise. The priority for ATC is going to be to get planes already in the air safely onto the ground before they run out of fuel. Planes on the ground can wait.
I don't see anything in his post saying he was dumb enough to try and power the whole rack with such a power cord.
In my experiance (living in the UK) most IEC cords are 0.75mm. 0.75mm flex is supposed to be good for 6A (and is probablly in practice good for substantially more, the current ratings on the small flex sizes are very conservative). That should be sufficient for most servers.
I do find it very misleading that the moulded IEC connectors seem to always be stamped with 10A even when connected to 0.75mm flex but if his flex melted either it was a very big server or the 0.75mm printed on the flex was a lie.
The JPEG originals from the camera are generally far too large for reasonable web use anyway. Nearly all photos you see on the web will have been downscaled and recompressed since they left the camera.
the difference is there is a virtually unlimited number of function names most of which do not have a pre-existing meaning. There is a very limited number of overloadable operators all of which do have a pre-existing meaning. So the temptation to abuse is much higher and ideas of what constitutes abuse much fuzzier..
Update: now i've managed to get on odriod's site and get some answers to the question. It seems they now offer a $9 shipping option for low value orders and they are using the postal service which tends to have marginally lower brokerage charges than couriurs. That makes the updated estimate.
$35 (board) + $9 (shipping) + ~£10 (brokerage) + ~£5.60 VAT = ~£43.6.
They also link to a store in germany that sells their products but said store doesn't have the C1 listed yet.
2. You can fake a USB device over the GPIO ports on both the A and B through various bitbang techniques.
Do you know anyone who has actually achived this on a Pi or are you just speculating.
The Pi is at least 2/3 more expensive.
The final cost will depend on both where you live and what if any distributor arrangements odriod set up. I would be very wary of considering a board cheaper based on headline price alone as the difference between headline price and final cost varies massively.
AIUI ipods are meant to play two classes of media files, non-drm files and files protected with apple drm*.
AIUI circumventing protections on access to a copyrighted work is illegal in the US but that isn't what real did. They found a way to put their own DRM media on ipods without undrming it by convincing the ipod that the file was using apple drm.
* Yes I know apple no longer uses drm for music sold on itunes but afaict they still use it for videos, previously purchased music that was not upgraded and some audiobooks. At one stage they were also still using it for music giveaways after they stopped using it for music sales, not sure if that is still the case.
The big question with this is what will the actual cost be? last time I bought an odriod product I had to pay the price of a fairly steep delivery charge to odriod and the dreaded courier VAT collection fee and of course the VAT itself. I don't remember exactly what it was in my case but according to http://forum.odroid.com/viewto... the final cost of of an "$89" U2 plus a "$9" HDMI cable shipped to the UK was £108.83.
If we assume the same delivery charge for the new board as for the U2* and that it still ships from outside the EU on a regular courier service**, and that they tell the truth on the customs form then the final price will be about $35 (board) + $30 (shipping) + ~£10.5 (brokerage) + ~£9.5 VAT = ~£68.
Whereas with a rasperry pi I pay the distributors listed price for the Pi itself (which is marginally higher than the raspberry pi foundations nominal price) and the VAT (delivery is "free") final cost of a "$35" raspberry pi B+ is £27.44.
Of course i'm in the UK which skews things a bit in the Pi's favour, other places the calculation may work out different but still comparing nominal prices is only a very rough way to compare the cost of SBCs.
* I can't check easilly if odriod's delivery charges have changed or even if this board is actually available yet because odriod's main site seems to have been /.ed
** Some couriers are now offering all-inclusive services where the seller handles collecting the VAT and brokerage costs are consolidated across multiple shipments, but in my experiance only a handful of vendors use them.
Banana Pi is only dual core 1GHz, This is quad core 1.5GHz (Though A5 is slower than A7 clock for clock but at least if we go by arm's DMIPs figures a 1.5GHz A5 core should outperform a 1GHz A7 core), Also note that while the banana Pi looks sorta like a Pi model B it is actually a bit larger, so it won't fit in Pi cases.
It still has it, they just moved from two seperate connectors for analog video and audio to one 4-pole connector.
Do injections in the eyball cause sigificant damage that would prevent them being used on a regular basis or are you just talking about the "what the patient will put up with" factor?
I would think that if the alternative was blindness patients would put up with weekly injections into the eyeball. From some googling it seems there are already eyeball injected drugs that are given every two months.
register size == memory address == unit of memory that can be operated on.
Every modern processor i'm aware of can operate on memory in 8-bit units despite having 32-bit or 64-bit register sizes and 32-bit or 64-bit memory addresses. Older processors with 8 and 16 bit register sizes typically had memory address sizes larger than their register size.
If multiplication of your objects doesn't make sense, then why on earth would you overload the * operator?
If bit shifting your object doesn't make sense then why on earth would you overload the << operator
I can't really make up my mind on operator overloading. On the one hand it can make custom types far more pleasant to use. On the other hand because operators are more concise than function calls and you can't add new ones it's very tempting to abuse operators to mean something other than their original meaning. Heck it's so tempting that even the authors of the standard library did it.
I wouldn't be so sure about that.
On the mac while "classic" mode is gone "carbon" is still there and was explicitly intended to allow porting of code from classic macos. I'd be surprised if there wasn't some code that had been written for classic macos still in there somewhere.
Similarly win32 was designed as a 32-bit variant of win16 and i'd be very surprised if there wasn't still some old code hanging arround somewhere.
While compression may help prevent some types of attack (in particular known plaintext attacks) it also creates new avenues of attack. In particular it makes the message length dependent on the message content.
For something like human written pgp emails this is not a massive deal,the length variability of human written messages and the low message rate mean that the chances of useful information being leaked this way are negligable.
For things like VOIP this can cause information leakage even to a passive attacker as what people say will impact the compressibility of the data. http://www.esecurityplanet.com...
For things like https the message rate is too low and the content too variable for a passive attacker to stand much chance of exploiting the messsage length side channel but an active attacker can stimulate traffic with known characteristics and partially known content to effectively exploit the channels.
I would argue that it is better to learn to write decent code than to let the language protect you from the effects of bad code.
I would argue that there is a middle ground. C++ still lets you do the low level stuff when you need to but also provides higher level structures that when used responsiblly make code clearer and less error prone.
I tried using a VGA cable and also a DVI to HDMI cable.
On VGA I couldn't get it to go above 1024x768, when I tried forcing a 1366x768 output on the computer it refused to display it (which I found very surprising as there really shouldn't be any significant differences in the sync timings between the two). I can't remember exactly what the results with the DVI to HDMI cable were but I remember having even more trouble..
I admittedly didn't have a video card to hand with an actual HDMI output to try.
If you get a 1080p TV that can do 60Hz, then you've got something that's pretty much the same as a monitor
Unfortunately in some cases they then go and ruin it by putting inappropriate signal processing in the path which can't be turned off. I have a hanspree TV which has the "FULL HD" badge on it and supposedly has a native resoloution of 1920x1200 but trying to use it as a computer monitor (regardless of input resoloution) results in a blurry mess.
My parents also have a smaller TV (think it's a sharp but I can't remember for sure) that supposedly has a native resoloution of 1366x768 but I never managed to get it to accept a signal at that resoloution.
Maybe I've just been unlucky or buying too cheap but I don't think I would buy a TV for use as a monitor again unless I could try before buying.
thousands of dollars per kilowatt hour in some cases
Reassembling a raid from drives taken at multiple different times will likely be harder than reassembling from disks taken at the same time because most raid implementations are designed to stop you doing it by accident but with sufficient determination it should be possible.
thousands of dollars per kilowatt hour in some cases
Do you have a source for that claim? are you sure you have your units right?
Thousands of dollars per megawatt hour I can just about belive, thousands of dollars per kilowatt hour is into the range of the absurd.
Which would be fine if they stuck to doing it in their own languages, unfortunately they cause ambiguity and confusion by writing english text but failling to follow english numeric conventions.
Not sure I would agree with "much of the world" either, from what I can tell it seems to be a mainland european thing.
if these open source projects are going to accept corporate sponsorship, they must do that corporation's bidding.
The people and entities who signed the sponsorship contract must do what they contractually agreed to do (which may be virtually nothing or it may be very specific depends what was in the contract).
Other people aren't bound by that though. Most contributors to open source projects do not have any contract with or obligation to the operators of the project. If they (or their employers if relavent) decide they would rather put their effort into a fork then they are perfectly entitled to do so.