However deaths that are actually alowed to happen are only one side of the story, the other thing that keeps getting brought up with nuclear disasters (see this article for instance) is the areas of land that are rendered unsuitable for their previous use (be that habitation, farming or whatever) because the way we prevent deaths is to avoid consuming food from contaminated areas and removing people completely from the most contaminated ones or (in the case of an ongoing incident) ones that may suddenly become highly contaminated if the incident gets worse.
My suspicion is in terms of overall "land area rendered unusable for it's previous purpose" nuclear power is fairly low down the scale but it would be nice to actually see the comparison with other accidents on that basis done properly and in a place the public will see it.
It's been set back because the usual suspects spent a couple of years trying to lock everybody into some proprietary, patented system but the light is starting to break through now.
Afaict for video the stalemate is still ongoing with one side refusing to implement theora or vp8 and the other side refusing to implement h.264 support (even on systems where there is already a usable h.264 decoder available).
Given this and given backwards compatibility (afaict internet explorer on winxp will never support the video tag) I would expect most sites to implement a video tag based player so they can support idevices but keep the flash based player for browsers that refuse to implement h.264 support in the video tag.
Just because some archivists were worried (rightly) about archiving MSOFFICE formats doesn't mean that ODF is the right solution. The impression I get is that ODF is mostly a red herring and was brought in by groups with a vested interest in unseating MS who want to use it as a lever to replace MSOFFICE with OOo.
IMO the real issue is that WYSIYG office formats blur the line between input and output and in the case of spreasheets blur the line between data input and processing rules input. The files store the input but the user is always (or nearly always) looking at the output so they think of the file as representing the output. That works fine as long as everyone is using the exact same applications but if an office file is archived and later read with a newer version of the office suite or with a different office suite the document may change.
They can try to specify things so tightly that any conforming implementation gives the same results but it is extremely difficult not to leave corner cases. It only takes a tiny difference in rounding to change the document. From a quick read of the ODF 1.2 spreadsheet formula spec it is nowhere near tight enough to guarantee that every spreadsheet app will give the same results. Heck it explicitly says that some things are implementation defined.
When archiving the first question should always be what you are trying to preserve? Are you trying to preserve the formulas your accountant used or are you trying to preserve the figures he produced? Are you trying to preserve the formatting commands you user entered into the word processor or are you trying to preserve the set of laid out pages the word processor produced from that input?
As I understand it these figures come from stripping a device down, making a list of all the significant parts (and maybe counting but not fully identifying the less significant ones like resistors and small capacitors), assigning a price to each one and then adding up a total.
The trouble is noone really knows what amazon pay for each part. They have to estimate based on general market prices and their estimates may be higher or lower than what amazon actually pays.
The question I always have with these large scale false positives is why aren't they caught in testing? I can understand a false positive on some obscure piece of software but a false positive on a core OS file or an app as popular as chrome would seem to speak of either extreme sloppiness or a more complex issue (such as MS and google releasing updates at about the same time).
Let us start with the premise that receiving a call on a mobile is a lot more expensive for the receiving telco than receiving one for a landline since they may have to route it round the country and it uses scarce radio bandwidth. This was certainly true in the early days of mobile phones though I suspect it's no so true nowadays. Afaict there are two main systems for charging for the service of getting an incoming call to a cellphone.
In europe they put mobile phones in special numbering blocks which typically cost more to call than normal landline numbers (at least when calling from a landline or an out of network mobile).
In north america mobile phones have normal phone numbers and calls to them are the same price as calls to a normal landline. So the receiver has to pay for the extra cost of getting the call to his mobile phone.
Really with modern technology driving down costs I'm not convinced either is justifiable anymore but good luck getting rid of said charges.
Though our system has it's own issues, for example on many plans calling someone on another mobile network is FAR more expensive than calling either a landline or a mobile on the same network. IIRC calls from landlines to mobiles are also often expensive.
There are "untrusted applets" which are sandboxed preventing access to loal resources and limiting access to network resources. Theese run without any warning prompts from the java plugin (IIRC some browsers will put up a warning before launching the plugin but firefox isn't one of them).
Then there are "trusted applets", theese pop up a warning with digital signature information when launched but afaict after that they can do pretty much anything they like to your system! "Trusted applets" are as dangerous as activex controls but for some reason they don't seem to get anything like the ammount of bad PR.
If you have a recent machine 8GB is indeed cheap. If you have a slightly older machine that takes DDR2 and supports 8GB (not all DDR2 based systems do) it's about twice as expensive. If your machine is any older than that you are most likely looking at replacing the motherboard and quite possiblly the CPU as well to get 8GB.
The whole POINT of a cellphone network is to get the maximum data moved while using only a small ammount of RF bandwidth. They do this by
1: moving the data a short distance using radio and moving it most of the way over wired networks. As the user density increases networks add more base stations so that the signal travels less distance by radio and the power and reuse distances are reduced. 2: carefully planning the use of frequency space first by allocating it to towers and then having the towers allocate it to users.
multi hop routing (what I presume you meant by P2P) with end devices acting as repeaters is a nice idea if you are trying to maximise coverage and have low capacity needs, it's a crappy idea if you are trying to provide maximum capacity.
Those who want to dual boot and want to pay as little for windows as possible without resorting to outright piracy (the big vendors get it cheaper than anyone else)? Those who want a decent laptop (there are a few barebones laptops out there)?
Also don't just think of the immediate impact. Think of the impact a few years down the line. Afaict many people get their initial linux experiance with a box that started life as a windows desktop and is being repurposed.
There is no requirement even right now that states a company must allow you the right to install any OS you choose.
As I understand it
Pre secure boot a manufacturer would have to go out of their way to precent the user installing any operating system they choose With secure boot a manufacturer would have to go out of their way to allow booting an operating system that isn't on their approved list.
We will have to see what vendors do in practice but I bet it will be a negative for those who don't want windows 8.
You don't need to sniff the user agent. The html5 video tag has provisions for sites to supply multiple formats and allow the browser to select the compatible one.
Some of us use our machines for more than just web browsing. If your browser is using most of your ram and you want to perform some task that also requires most of your ram then you have to wait while your browser is swapped out so the other task can proceed and then wait again for large chunks of your browser to be swapped back in (remember swapping happens in blocks, so if anything in a block is needed the entire block has to be swapped back in) when you switch back. All because your browser decided to cache a load of stuff that most likely won't be used again.
It's a simple rule of intelligence, the more people know something and the less well vetted those people are the greater the chance that one of those people is working for either a current enemy or at least a potential future enemy.
Your subcontractor is not meeting your quality standards and can't give a satisfactory explanation as to why.
OR
The system you supplied to a customer with a service contract is not working properly and you are getting nowhere with trying to fix it remotely or guide the customer through fixing it.
OR
Someone you need to deal with never answers there phone and is so slow in responding to emails you can't get anything finalised.
OR
You want to discuss something and you want to make SURE it stays off the record.
For all our modern technology sometimes there is no good substitute for actually being there to do something or sort something out.
mmm, games are a tricky one. On the one hand some games (particularly those with excruciating load times like RCT3) benefit big time from being on a SSD. On the other hand modern games are pretty big.
Still unless you are someone who hops madly between games a 120GB SSD (not that big by todays standards) should be enough to keep all the games you play frequently on the SSD.
The problem is if you want people to use public transport links you need to make them more attractive than driving. In a rich country where private cars are affordable to buy and run that means you need to make the public transport option at least comparable in time terms.
If transport links only run occasionally passengers lose a lot of time waiting (both to set off initially and at change points) but to justify running them frequently you need a LOT of passengers trying to go the same way. The more people you have living within reasonable walking distance of a pick up point (or line of pick up points) the more passengers you are likely to have going the same way and therefore you can justify a more frequent service. Also in high density areas congestion slows down car traffic while public transport that uses segregated routes (rail, subway, sometimes bus and tram) keeps moving quickly.
Further the only reasonable way to provide even a tolerable service frequency in lower density areas is to have LOTS of pick up points but the more pick up points you have on a service the slower that service becomes.
Sure but each office can only be across the street from ONE exchange.
Currencies, commodities and some shares are traded on multiple markets. If you own the first computer to spot a difference in price between the same item in different markets then you can make money by making a buy in one market and an equivilent sell in another. The faster you can exchange data the lower the risk that someone else will get there first.
Sure if it's 16x4GB modules of unregistered non-ecc DRR3. Thing is though that isn't a very realistic configuration because it would require EIGHT ram channels to support it (unregistered DDR3 has a maximum of two modules per channel). Single socket boards and most dual socket boards simply don't have that many ram channels.
8x8GB modules of registered ECC DDR3 (a more realistic configuration) will set you back $608. Even then there is a good chance you will be looking at a CPU/MB replacement to support the ram.
But note that even that board supports a max of 24GB of RAM, so you couldn't necessarily even get 32GB, even with 6 slots.
Indeed single socket LGA1366 CPUs/motherboards max out at 24GB (which is a little weird because dual LGA1366 will support many times that). Being a high end platform though it's pretty expensive ($100+ for the MB, $300+ for the CPU).
LGA1155 systems will do 32GB but to do so you need 32GB sticks of unregistered non-ecc memory which are insanely expensive at the moment.
If you want 32GB and don't care about single threaded CPU performance your cheapest option is probably to go with an AMD G34 system. This platform has four ram channels and at least according to supermicro it supports desktop (unregistered non-ECC) ram. So you can load it up with 8x4GB of cheap desktop ram. Price for CPU/MB is in the same ballpark as LGA1366 but you get 8 slow cores instead of 4 fast cores.
I just bought 8 GB high-end (1600Mhz DDR3) RAM for 30 quid.
I wouldn't call 4GB DIMMs (I very much doubt you got an 8GB stick at that price) of desktop ram "high end" even if they happen to be overclocked slightly.
That makes 240 for 64 GB.
Unfortunately it doesn't (I know the GP said ignoring the cost of the motherboard but the GGP didn't). To do 64GB with 4GB DIMMs would require 16 sticks. Given that desktop DDR3 only supports 2 DIMMs per channel and than a typical system has 2-3 channels this becomes a problem. To fit 64GB total using 4GB desktop modules would require a board with 8 ram channels. The most ram channels you get on a single CPU socket is 4.
Unless it's already a high end server platform "Upgrading" an existing system to 64GB means ripping out the CPU and motherboard and replacing them with a server platform that can take all that memory. Further it most likely means use of more expensive ram
Lets look at how much it would actually cost to do that first using 8GB DIMMs of registered ECC memory*.
Slashdot Mirror
However deaths that are actually alowed to happen are only one side of the story, the other thing that keeps getting brought up with nuclear disasters (see this article for instance) is the areas of land that are rendered unsuitable for their previous use (be that habitation, farming or whatever) because the way we prevent deaths is to avoid consuming food from contaminated areas and removing people completely from the most contaminated ones or (in the case of an ongoing incident) ones that may suddenly become highly contaminated if the incident gets worse.
My suspicion is in terms of overall "land area rendered unusable for it's previous purpose" nuclear power is fairly low down the scale but it would be nice to actually see the comparison with other accidents on that basis done properly and in a place the public will see it.
It's been set back because the usual suspects spent a couple of years trying to lock everybody into some proprietary, patented system but the light is starting to break through now.
Afaict for video the stalemate is still ongoing with one side refusing to implement theora or vp8 and the other side refusing to implement h.264 support (even on systems where there is already a usable h.264 decoder available).
Given this and given backwards compatibility (afaict internet explorer on winxp will never support the video tag) I would expect most sites to implement a video tag based player so they can support idevices but keep the flash based player for browsers that refuse to implement h.264 support in the video tag.
Just because some archivists were worried (rightly) about archiving MSOFFICE formats doesn't mean that ODF is the right solution. The impression I get is that ODF is mostly a red herring and was brought in by groups with a vested interest in unseating MS who want to use it as a lever to replace MSOFFICE with OOo.
IMO the real issue is that WYSIYG office formats blur the line between input and output and in the case of spreasheets blur the line between data input and processing rules input. The files store the input but the user is always (or nearly always) looking at the output so they think of the file as representing the output. That works fine as long as everyone is using the exact same applications but if an office file is archived and later read with a newer version of the office suite or with a different office suite the document may change.
They can try to specify things so tightly that any conforming implementation gives the same results but it is extremely difficult not to leave corner cases. It only takes a tiny difference in rounding to change the document. From a quick read of the ODF 1.2 spreadsheet formula spec it is nowhere near tight enough to guarantee that every spreadsheet app will give the same results. Heck it explicitly says that some things are implementation defined.
When archiving the first question should always be what you are trying to preserve? Are you trying to preserve the formulas your accountant used or are you trying to preserve the figures he produced? Are you trying to preserve the formatting commands you user entered into the word processor or are you trying to preserve the set of laid out pages the word processor produced from that input?
Afaict the truth is noone really knows.
As I understand it these figures come from stripping a device down, making a list of all the significant parts (and maybe counting but not fully identifying the less significant ones like resistors and small capacitors), assigning a price to each one and then adding up a total.
The trouble is noone really knows what amazon pay for each part. They have to estimate based on general market prices and their estimates may be higher or lower than what amazon actually pays.
The question I always have with these large scale false positives is why aren't they caught in testing? I can understand a false positive on some obscure piece of software but a false positive on a core OS file or an app as popular as chrome would seem to speak of either extreme sloppiness or a more complex issue (such as MS and google releasing updates at about the same time).
Let us start with the premise that receiving a call on a mobile is a lot more expensive for the receiving telco than receiving one for a landline since they may have to route it round the country and it uses scarce radio bandwidth. This was certainly true in the early days of mobile phones though I suspect it's no so true nowadays. Afaict there are two main systems for charging for the service of getting an incoming call to a cellphone.
In europe they put mobile phones in special numbering blocks which typically cost more to call than normal landline numbers (at least when calling from a landline or an out of network mobile).
In north america mobile phones have normal phone numbers and calls to them are the same price as calls to a normal landline. So the receiver has to pay for the extra cost of getting the call to his mobile phone.
Really with modern technology driving down costs I'm not convinced either is justifiable anymore but good luck getting rid of said charges.
I miss the UK where incoming calls were free.
Though our system has it's own issues, for example on many plans calling someone on another mobile network is FAR more expensive than calling either a landline or a mobile on the same network. IIRC calls from landlines to mobiles are also often expensive.
How can you even compare Flash and Java?
For those trying to develop apps in a web browser that don't fit the traditional page by page model there are essentially 4 choices.
1: AJAX
2: Java applet
3: FLASH
4: Activex control
So of course those choices will get compared. They all have strengths and weaknesses of course but they can be used for much the same tasks.
Afaict java applets come into two categories.
There are "untrusted applets" which are sandboxed preventing access to loal resources and limiting access to network resources. Theese run without any warning prompts from the java plugin (IIRC some browsers will put up a warning before launching the plugin but firefox isn't one of them).
Then there are "trusted applets", theese pop up a warning with digital signature information when launched but afaict after that they can do pretty much anything they like to your system! "Trusted applets" are as dangerous as activex controls but for some reason they don't seem to get anything like the ammount of bad PR.
If you have a recent machine 8GB is indeed cheap. If you have a slightly older machine that takes DDR2 and supports 8GB (not all DDR2 based systems do) it's about twice as expensive. If your machine is any older than that you are most likely looking at replacing the motherboard and quite possiblly the CPU as well to get 8GB.
The whole POINT of a cellphone network is to get the maximum data moved while using only a small ammount of RF bandwidth. They do this by
1: moving the data a short distance using radio and moving it most of the way over wired networks. As the user density increases networks add more base stations so that the signal travels less distance by radio and the power and reuse distances are reduced.
2: carefully planning the use of frequency space first by allocating it to towers and then having the towers allocate it to users.
multi hop routing (what I presume you meant by P2P) with end devices acting as repeaters is a nice idea if you are trying to maximise coverage and have low capacity needs, it's a crappy idea if you are trying to provide maximum capacity.
Those who want to dual boot and want to pay as little for windows as possible without resorting to outright piracy (the big vendors get it cheaper than anyone else)?
Those who want a decent laptop (there are a few barebones laptops out there)?
Also don't just think of the immediate impact. Think of the impact a few years down the line. Afaict many people get their initial linux experiance with a box that started life as a windows desktop and is being repurposed.
There is no requirement even right now that states a company must allow you the right to install any OS you choose.
As I understand it
Pre secure boot a manufacturer would have to go out of their way to precent the user installing any operating system they choose
With secure boot a manufacturer would have to go out of their way to allow booting an operating system that isn't on their approved list.
We will have to see what vendors do in practice but I bet it will be a negative for those who don't want windows 8.
To be fair this would be two jumpers, since you don't seem to understand how jumpers work.
A 3-pin jumper can be either placed on pins 1-2, placed on pins 2-3 or placed so it does not connect any pins at all. That gives 3 possible options.
You don't need to sniff the user agent. The html5 video tag has provisions for sites to supply multiple formats and allow the browser to select the compatible one.
Some of us use our machines for more than just web browsing. If your browser is using most of your ram and you want to perform some task that also requires most of your ram then you have to wait while your browser is swapped out so the other task can proceed and then wait again for large chunks of your browser to be swapped back in (remember swapping happens in blocks, so if anything in a block is needed the entire block has to be swapped back in) when you switch back. All because your browser decided to cache a load of stuff that most likely won't be used again.
It's a simple rule of intelligence, the more people know something and the less well vetted those people are the greater the chance that one of those people is working for either a current enemy or at least a potential future enemy.
So what do you do if
Your subcontractor is not meeting your quality standards and can't give a satisfactory explanation as to why.
OR
The system you supplied to a customer with a service contract is not working properly and you are getting nowhere with trying to fix it remotely or guide the customer through fixing it.
OR
Someone you need to deal with never answers there phone and is so slow in responding to emails you can't get anything finalised.
OR
You want to discuss something and you want to make SURE it stays off the record.
For all our modern technology sometimes there is no good substitute for actually being there to do something or sort something out.
mmm, games are a tricky one. On the one hand some games (particularly those with excruciating load times like RCT3) benefit big time from being on a SSD. On the other hand modern games are pretty big.
Still unless you are someone who hops madly between games a 120GB SSD (not that big by todays standards) should be enough to keep all the games you play frequently on the SSD.
The problem is if you want people to use public transport links you need to make them more attractive than driving. In a rich country where private cars are affordable to buy and run that means you need to make the public transport option at least comparable in time terms.
If transport links only run occasionally passengers lose a lot of time waiting (both to set off initially and at change points) but to justify running them frequently you need a LOT of passengers trying to go the same way. The more people you have living within reasonable walking distance of a pick up point (or line of pick up points) the more passengers you are likely to have going the same way and therefore you can justify a more frequent service. Also in high density areas congestion slows down car traffic while public transport that uses segregated routes (rail, subway, sometimes bus and tram) keeps moving quickly.
Further the only reasonable way to provide even a tolerable service frequency in lower density areas is to have LOTS of pick up points but the more pick up points you have on a service the slower that service becomes.
C is speed of light in vacuum, fiber is not vacuum, hence speed of light in this case is not necessarily
It looks like this GP used wolfram alpha's "light in fiber" travel time so he covered that one.
Sure but each office can only be across the street from ONE exchange.
Currencies, commodities and some shares are traded on multiple markets. If you own the first computer to spot a difference in price between the same item in different markets then you can make money by making a buy in one market and an equivilent sell in another. The faster you can exchange data the lower the risk that someone else will get there first.
64GB DDR3 = $352
Sure if it's 16x4GB modules of unregistered non-ecc DRR3. Thing is though that isn't a very realistic configuration because it would require EIGHT ram channels to support it (unregistered DDR3 has a maximum of two modules per channel). Single socket boards and most dual socket boards simply don't have that many ram channels.
8x8GB modules of registered ECC DDR3 (a more realistic configuration) will set you back $608. Even then there is a good chance you will be looking at a CPU/MB replacement to support the ram.
But note that even that board supports a max of 24GB of RAM, so you couldn't necessarily even get 32GB, even with 6 slots.
Indeed single socket LGA1366 CPUs/motherboards max out at 24GB (which is a little weird because dual LGA1366 will support many times that). Being a high end platform though it's pretty expensive ($100+ for the MB, $300+ for the CPU).
LGA1155 systems will do 32GB but to do so you need 32GB sticks of unregistered non-ecc memory which are insanely expensive at the moment.
If you want 32GB and don't care about single threaded CPU performance your cheapest option is probably to go with an AMD G34 system. This platform has four ram channels and at least according to supermicro it supports desktop (unregistered non-ECC) ram. So you can load it up with 8x4GB of cheap desktop ram. Price for CPU/MB is in the same ballpark as LGA1366 but you get 8 slow cores instead of 4 fast cores.
I just bought 8 GB high-end (1600Mhz DDR3) RAM for 30 quid.
I wouldn't call 4GB DIMMs (I very much doubt you got an 8GB stick at that price) of desktop ram "high end" even if they happen to be overclocked slightly.
That makes 240 for 64 GB.
Unfortunately it doesn't (I know the GP said ignoring the cost of the motherboard but the GGP didn't). To do 64GB with 4GB DIMMs would require 16 sticks. Given that desktop DDR3 only supports 2 DIMMs per channel and than a typical system has 2-3 channels this becomes a problem. To fit 64GB total using 4GB desktop modules would require a board with 8 ram channels. The most ram channels you get on a single CPU socket is 4.
Unless it's already a high end server platform "Upgrading" an existing system to 64GB means ripping out the CPU and motherboard and replacing them with a server platform that can take all that memory. Further it most likely means use of more expensive ram
Lets look at how much it would actually cost to do that first using 8GB DIMMs of registered ECC memory*.
Motherboard: 1x http://www.newegg.com/Product/Product.aspx?Item=N82E16813182240 : $225
CPU: 1x http://www.newegg.com/Product/Product.aspx?Item=N82E16819105266 : $250
Ram: 8x http://www.newegg.com/Product/Product.aspx?Item=N82E16820139280 : $76 each -> $608
total: $1083
Now lets consider using 4GB DIMMs of desktop memory.
Motherboard: 1x http://www.provantage.com/supermicro-h8dg6-f~7SUPM3F1.htm : $578
CPU: 2x http://www.newegg.com/Product/Product.aspx?Item=N82E16819105266 : $250 each -> $500
RAM: 8x http://www.newegg.com/Product/Product.aspx?Item=N82E16820313123 : $40 each -> $320
total: $1398
* 8GB DIMMs of desktop memory are like hens teeth at the moment and cost more than 8GB sticks of registered ECC memory.