The thing with "VGA" is there really isn't too much to it, three analog video signals and two sync signals with some loose agreements on timings.
That means that there is very little theoretical limit on resolution* but it also means that.
1: All components in the chain have to actually have sufficient analog bandwidth. Lack of strong standards and gradual failure (rather than the brick wall failure you get with digital systems) if the analog circuitry is skimped on encourages skimping on the analog components. This is particually bad with TVs (monitors seem to make an effort to give acceptable performance on VGA at their native resoloution). 2: When driving a screen with discrete pixels the receiver has to guess where each line starts and ends. They are generally pretty good at it but again poor implementations, unhelpful content (completely black screen, screen with black bars from the source) or just plain bad luck can cause mis-locks which are annoying. 3: The individual pixels will inevitably not be completely isolated from each other.
* The connector probably imposes some limit but using the rule of thumb that structures less than a tenth of a wavelength can be regarded as of negligable size it should be usable up to a few gigahertz with careful termination..
With computer monitors, you're generally paying a premium for better input latencies, refresh rates, color reproduction, and ghosting.
And not fucking with the input signal.
I've learnt the hard way that some TVs are incapable of taking an input signal in their advertised native resolution and displaying it without fucking it up though application of inappropriate processing that smears single pixel lines making the desktop a blurry mess.
I've also learnt the hard way that some TVs have terrible VGA inputs that take ages to lock, can't lock properly to the low resoloutions seen during bootup/bios, sometimes mis-lock even at their native resooution (though to be fair i've seen monitors mis-lock from time to time too) and degrade the signal even more than the aforementioned inappropriate processing does.
I would NOT buy a TV for use as a monitor without confirming it doesn't suffer from these problems.
* And before anyone gives me any BS VGA being unsuitable for "full HD" resoloutions i've been running plenty of monitors at 1920x1080x60 and 1920x1200x60 over VGA without any noticeable issues.
Games rarely have to specifically support a resolution. Most will query the system to see what resolutions are possible - they may have to upscale UI elements that are normally 1:1 or downscaled
Theres also some games that don't scale the UI elements at all.
I remember experiencing this years back with the original quake (dos version), It was designed for something like 320x200 and if you cranked it up to 800x600 or worse 1024x760 the UI became unusablly small (I presume this has been fixed by now either by ID or by third parties).
A lot of "builder" and "rts" type games have also had UIs and sometimes content too (though that is rarer now that contents is 3D rendered) in fixed pixel sizes. In recent years it hasn't been so much of a problem because monitors (when running at their native resoloution) have all been roughly the same pixel density but with these new high DPI screens UIs designed for a set number of pixels may become a problem again (it's a problem on the desktop too and MS and apple have both had to implement workarounds there which I can see may or may not play nice with games).
Almost all payment methods have some risk that a payment you thought you had received doesn't actually go through
How long does it take for a credit card transaction to become irreversable? (indeed do they EVER become irreversable?) Some grocery stores apparently even take swiped payments without signature (in america) or contactless payments (in europe) making them even more vulnerable to reversal!
Cash can contain forgeries that are good enough to only be detected when the cash is banked.
Is the risk of accepting unconfirmed bitcoin transactions greater or less than the above risks?
I was a security researcher In a previous life. You say "...in practice there are several attack vectors" without references or listing any, and I wonder if you are making things up. We'd all like to hear what the "several attack vectors" are for BitCoin - the best I've seen is one computationally unfeasible possibility. Brute force doesn't count. Can you tell us more?
The main attack vector is getting more hashing power than the rest of the bitcoin network put together. This is expensive* but not unfeasible. Once you do that you can make it so that all new blocks in the "strongest blockchain" are ones generated by you.
Once you do that you can stop transactions from ever confirming unless you approve of them, You also get to keep all the mining rewards and transaction fees from the transactions you allow.
So for example you could institute the following rules * transactions between two registered addresses registered to the same person require no fee * transactions between two registered addresses registered to different people require a 0.1% free * transactions between a registered address and an unregistered address see a 10% fee * transactions between two unregistered addresses see a 20% fee
Anyone who refused to comply with these new rules would have their transactions fail to ever confirm. This would create a very strong motivation for all legitimate users of bitcoin to register with the attacker so they see much lower fees. Those who refused to register with the attacker would have to chose between either having their transactions remain unconfirmed or gradually have their bitcoin wealth transferred to you.
The bitcoin guys dismiss this as "probablly not a problem" but don't really give any evidence for that position.
* current bitcoin network hashrate is of the order of 100 terahashes per second. FPGAs are achiving about 1 megahashes per second per dollar. Making the cost of the attack arround 100 million dollars. Some asic miner vendors are advertising 50 megahash per second per dollar making the attack cost only 2 million dollars but may be difficult to source quickly in sufficient qualitity since the chips are specific to bitcoin mining. It may well be the way to go with such an attack is to arrange your own asic production run.
Also with some software it's trivial to get the license key from an installed copy of the software. I've even seen at least one program that showed it on the splash screwn at startup.
The fact that they are VM's really only makes two differences practical differences that matter, fist is that is easy to roll them back and second is that your aren't running on bare metal.
The third is that the VM soloution is essentially an operating system in it's own right. Like any other operating system it can suffer from privilage escalation exploits.
Humans are very good at quickly getting objects in and out of awkward spaces but only if those objects are fairly light. Your AA batteries are no problem for even a small child to handle. The starter batteries for petrol powered cars are getting towards the limit of what one person can easily and safely handle.
Afaict an electric car battery is of the order of half a ton. Getting something that weight in and out quickly while also keeping it in a place that is sheilded from crashes and doesn't mess with the praciticality or aerodynamics of the car is a much trickier proposition than dealing with a few AAs.
Some quick estimation (read: looking at online vendors but not shopping arround carefully for best prices nor carefullly checking compatibility) puts the cost of a basic system built round that board and with all slots filled with 16 core processors at the order of $5K.
I guess that may be cheap to you, it certainly isn't to me.
Sure you can find a few dev boards or industrial embedded computing boards where the vendor happened to use the mini-itx form factor rather than something custom which is handy if you want to slap one of them in an off the shelf case for whatever reason. Such boards have been arround for some time.
But they lack the features one takes for granted on atom mini-itx boards like multiple SATA ports, a regular PCI or PCIe slot*, support for more ram (most arm boards have 1GB with the occasional board with 2GB just coming onto the market while most atom boards support 4GB with newer boards supporting 8GB). I also noticed that none of your links mentioned price. Shaving even a few tens of watts is not worth it if the board is double the price of an equivilent atom soloution.
Interestingly the Marvell armada XP platform can in principle offer most of what you find on atom mini-itx boards but I haven't seen anyone build such a board arround it. The only armada XP hardware i've seen is the crazily overpriced and somewhat strangely set-up (lots of network ports, hardly any SATA) openblocks stuff, the even more overpriced baserock slab and the vapourware dell copper..
*one of them has miniPCIe but that has a far smaller selection of cards and while in theory you can use adaptors in practice you'll never get it into a standard case if you do which kinda defeats the object of standard form factors
I think that his point is that, with CAD, even a trained monkey can tell the software "Just iterate through all the pretty little pictures we drew, multiply their volume by their density, and then add it all up" and arrive at a final weight.
Sure they can tell the software to do that.
But the CAD software is fundamentally only as good as the information put into it (the density of each material, what material is being used for what, how thick it will be, the weight of assemblies designed elsewhere and incorporated as black boxes and so-on), if that information is inaccurate then it can throw the total by a lot regardless of whether the design is being on paper or by software.
Automating things means less chance of arithmetic errors but it can also increase the chance that the calculations are applied blindly without anyone asking whether what is being calculated actually represents reality.
High speed DSL won't go far enough to get to the telephone exchange but it will go far enough to get to the cabinet in most urban settings at least.
The plan in the UK seems to be to do a mixture with the existing ADSL system continuing to serve the lower tier broadband users. FTTC+VDSL for the middle tier broadband users and FTTH for the top tier broadband users.
mmm, buisness "broadband" is only a small step up (if any) from home "broadband", mostly just a lifting of restrictions and a higher pricetag.
If you want a dedicated connection to the ISPs core network and then agreements that they will as far as reasonably practical run an uncongested core network and uncongested links to their upstreams and peering points then you can get that but expect to pay a hell of a lot more for it than you would pay for buisness broadband.
No, Sony confirmed that "PS4 games will be playable without an Internet connection". That pretty much nails it: this used game policy can only be enforced with an online DRM.
I don't see why.
You could have a writable area on the media that "bound" the copy to the first console it was used on. A game shop reselling the game would "unbind" the copy (and pay a fee for doing so) before reselling it. If you bought a copy privately you'd have to either take it into a game shop for unbinding or connect your console to the internet and pay the unbinding fee online. Or you could sell "unbinding cards" with little NFC chips in them that the user held up to their console to unbind the game. Or you could supply unbinding codes over the phone like how windows activation is handled without an internet connection.
In general the expectation is that internet client use will be bursty. When I want to move data I want it as fast as I can but that doesn't mean i'm going to be anywhere close to maxing out my internet connection 24/7. Having to buy connections priced on the assumption that they would be heavilly loaded 24/7 would make fast connections crazy expensive.
"unlimited" home internet connections are kinda like "all you can eat" buffets.
At an all you can eat buffet there is no direct limit on how much food you are allowed to eat but there are other rules that practically limit what you can eat. For example there is usually a maximum ammount of time you can stay and a rule that you can't take food away, a rule that everyone in the group has to pay and a rule that taking food out is forbidden. If they think you are consuming an ammount of food that would be impractical if you are following the rules they are going to investigate further.
Similarly an unlimited home internet connection, particaully one with a very high peak bandwidth will typically have rules that are designed to restrict users to normal client usage and if they think you are using an ammount of traffic that would be impractical with normal client usage they are going to investigate further.
A buisness connection has less rules on how you can use it and as a result they expect you to use more traffic and hence charge you a higher price and/or put explicit limits in place.
They could throw $50 billion at it and not destroy the network
If they want to destroy bitcoin then all they need to do is get more hasing power than the rest of the bitcoin network put together. I'm fairly sure that with the current level of hashing power that is within the reach of most governments.
Once they do that they can create arbitary rules for refusing transactions they don't like. For example they could require that all bitcoin addresses used as transaction destinations are registered with the government before use. They would enforce these rules by refusing to accept transactions that violate them into blocks they generate and refusing to use blocks that contain such transactions as a base for blocks they generate. Occasionally a non-governement miner may generate a block with your non-government approved transaction in it but the government miners would fork it off before it got any significant number of confirmations.
Bitcoin is and always has been psudononymous. It's trivial to track the flow of coins between bitcoin addresses (which are essentially psudonyms). So how much anonymity there is depends on how difficult it is to associate those bitcoin addresses with real people.
How difficult it is to associate those addresses with real people depends hugely on both what records are kept and how many significant players there are in the bitcoin market (it's much easier to force a small number of large entities to hand over records to the cops than it is to force a large number of small entities to do it)
You could potentially supply power to something using a male to male cord
There is a reason such cords are reffered to as widowmakers.
If you really want multiple inlets then I guess you could use connectors where both sides are touchproof but even then there would be a risk if someone hooked up too cords to the same peice of furniture. At the very least i'd want to see padlocks and warning notices in such a scenario.
Both apple and MS now have smartphone and locked down arm based tablet platforms where they force people into app stores from which they take a substantial cut. Both apple and MS still have traditional desktop/laptop (and in MS's case bulky x86 based tablet) platforms where for the moment they still let you install apps in the traditional way. However even on those platforms afaict they limit some features of the platform to store apps only.
The difference is of course that with apple nowadays the desktops/laptops are a sideline and the phones/locked down arm based tablets are the main thrust of the buisness. With MS the OS for phones/locked down arm based tablets is a sideline and the desktop/laptop OS is the main thrust of the buisness.
Computers typically use air cooling, the exhaust temperature of a computer is not very far above the intake temperature and the intake temperature is typically arround normal room temperature or lower. So the exhaust temperature is likely to be barely above normal room temperature making moving the heat arround difficult.
You could raise the intake temperature to the computers but doing so would have significant disadvantages. Firstly it would reduce the ammount of time you had between cooling equipment failure and the temperature rising beyond the maximum safe level for the equipment. Secondly it may cause equipment that isn't designed to work in those temperatures to fail or at least reduce it's life. It would also make things rather uncomforable for people working in the datacenter.
You could also redesign the computers to use liquid cooling, since liquid cooling is far more efficient than air cooling you could run the loop at a significantly higher temperature than typical datacenter air temperatures while keeping the core temperature the same. The downside is of course you'd need to redesign the cooling systems in all your computers and come up with a system for safely adding and removing computers to/from the liquid cooling system.
Sun released openoffice under the GPL (I don't remember the exact version offhand) and in the wake of the oracle takeover a number of developers decided to fork and produce libreoffice. While the oracle takeover was apparently the immediate trigger I get the impression that there were deep tensions within the project already.
Normally when such a fork happens one of a few things results 1: The two forks continue in paralell with code flowing from the original to the fork and possiblly back again (it would be difficult for it to flow back in this case though because of contributor agreement requirements). 2: The operators of the forked project ignores the fork but is either abandoned or at least development slows to a cral 3: The fork is blessed by the forked project to some greater or lesser extent.
In this case though none of those things happened. Instead oracle decided to give the rights for openoffice to apache who proceeded to release it under the apache license (a non-copyleft license) and afaict the version given to apache was not the same as the one libreoffice was based on.
Libreoffice in turn have decided that while they want copyleft they want one weaker than the GPL, so they are rebasing their work on top of apache openoffice. Afaict this means that code will be able to flow from apache openoffice to libreoffice but not vice-versa.
I thought, "send them off for a few years of hard labour", then my brain started thinking though how such a sentance could work in the modern world. Forbidding their friends and family from contact completely would be too cruel IMO, OTOH having rich friends and family jetting in all the time and bring expensive gifts wouldn't seem like much of a punishment.
I don't belive conventional prison is appropriate for most white collar criminals but I do belive they need to have their luxuries taken away and be made to live a hard life for a while.
Slashdot Mirror
The thing with "VGA" is there really isn't too much to it, three analog video signals and two sync signals with some loose agreements on timings.
That means that there is very little theoretical limit on resolution* but it also means that.
1: All components in the chain have to actually have sufficient analog bandwidth. Lack of strong standards and gradual failure (rather than the brick wall failure you get with digital systems) if the analog circuitry is skimped on encourages skimping on the analog components. This is particually bad with TVs (monitors seem to make an effort to give acceptable performance on VGA at their native resoloution).
2: When driving a screen with discrete pixels the receiver has to guess where each line starts and ends. They are generally pretty good at it but again poor implementations, unhelpful content (completely black screen, screen with black bars from the source) or just plain bad luck can cause mis-locks which are annoying.
3: The individual pixels will inevitably not be completely isolated from each other.
* The connector probably imposes some limit but using the rule of thumb that structures less than a tenth of a wavelength can be regarded as of negligable size it should be usable up to a few gigahertz with careful termination..
With computer monitors, you're generally paying a premium for better input latencies, refresh rates, color reproduction, and ghosting.
And not fucking with the input signal.
I've learnt the hard way that some TVs are incapable of taking an input signal in their advertised native resolution and displaying it without fucking it up though application of inappropriate processing that smears single pixel lines making the desktop a blurry mess.
I've also learnt the hard way that some TVs have terrible VGA inputs that take ages to lock, can't lock properly to the low resoloutions seen during bootup/bios, sometimes mis-lock even at their native resooution (though to be fair i've seen monitors mis-lock from time to time too) and degrade the signal even more than the aforementioned inappropriate processing does.
I would NOT buy a TV for use as a monitor without confirming it doesn't suffer from these problems.
* And before anyone gives me any BS VGA being unsuitable for "full HD" resoloutions i've been running plenty of monitors at 1920x1080x60 and 1920x1200x60 over VGA without any noticeable issues.
Games rarely have to specifically support a resolution. Most will query the system to see what resolutions are possible - they may have to upscale UI elements that are normally 1:1 or downscaled
Theres also some games that don't scale the UI elements at all.
I remember experiencing this years back with the original quake (dos version), It was designed for something like 320x200 and if you cranked it up to 800x600 or worse 1024x760 the UI became unusablly small (I presume this has been fixed by now either by ID or by third parties).
A lot of "builder" and "rts" type games have also had UIs and sometimes content too (though that is rarer now that contents is 3D rendered) in fixed pixel sizes. In recent years it hasn't been so much of a problem because monitors (when running at their native resoloution) have all been roughly the same pixel density but with these new high DPI screens UIs designed for a set number of pixels may become a problem again (it's a problem on the desktop too and MS and apple have both had to implement workarounds there which I can see may or may not play nice with games).
Almost all payment methods have some risk that a payment you thought you had received doesn't actually go through
How long does it take for a credit card transaction to become irreversable? (indeed do they EVER become irreversable?) Some grocery stores apparently even take swiped payments without signature (in america) or contactless payments (in europe) making them even more vulnerable to reversal!
Cash can contain forgeries that are good enough to only be detected when the cash is banked.
Is the risk of accepting unconfirmed bitcoin transactions greater or less than the above risks?
I was a security researcher In a previous life. You say "...in practice there are several attack vectors" without references or listing any, and I wonder if you are making things up. We'd all like to hear what the "several attack vectors" are for BitCoin - the best I've seen is one computationally unfeasible possibility. Brute force doesn't count. Can you tell us more?
The main attack vector is getting more hashing power than the rest of the bitcoin network put together. This is expensive* but not unfeasible. Once you do that you can make it so that all new blocks in the "strongest blockchain" are ones generated by you.
Once you do that you can stop transactions from ever confirming unless you approve of them, You also get to keep all the mining rewards and transaction fees from the transactions you allow.
So for example you could institute the following rules
* transactions between two registered addresses registered to the same person require no fee
* transactions between two registered addresses registered to different people require a 0.1% free
* transactions between a registered address and an unregistered address see a 10% fee
* transactions between two unregistered addresses see a 20% fee
Anyone who refused to comply with these new rules would have their transactions fail to ever confirm. This would create a very strong motivation for all legitimate users of bitcoin to register with the attacker so they see much lower fees. Those who refused to register with the attacker would have to chose between either having their transactions remain unconfirmed or gradually have their bitcoin wealth transferred to you.
The bitcoin guys dismiss this as "probablly not a problem" but don't really give any evidence for that position.
* current bitcoin network hashrate is of the order of 100 terahashes per second. FPGAs are achiving about 1 megahashes per second per dollar. Making the cost of the attack arround 100 million dollars. Some asic miner vendors are advertising 50 megahash per second per dollar making the attack cost only 2 million dollars but may be difficult to source quickly in sufficient qualitity since the chips are specific to bitcoin mining. It may well be the way to go with such an attack is to arrange your own asic production run.
Also with some software it's trivial to get the license key from an installed copy of the software. I've even seen at least one program that showed it on the splash screwn at startup.
The fact that they are VM's really only makes two differences practical differences that matter, fist is that is easy to roll them back and second is that your aren't running on bare metal.
The third is that the VM soloution is essentially an operating system in it's own right. Like any other operating system it can suffer from privilage escalation exploits.
Humans are very good at quickly getting objects in and out of awkward spaces but only if those objects are fairly light. Your AA batteries are no problem for even a small child to handle. The starter batteries for petrol powered cars are getting towards the limit of what one person can easily and safely handle.
Afaict an electric car battery is of the order of half a ton. Getting something that weight in and out quickly while also keeping it in a place that is sheilded from crashes and doesn't mess with the praciticality or aerodynamics of the car is a much trickier proposition than dealing with a few AAs.
Some quick estimation (read: looking at online vendors but not shopping arround carefully for best prices nor carefullly checking compatibility) puts the cost of a basic system built round that board and with all slots filled with 16 core processors at the order of $5K.
I guess that may be cheap to you, it certainly isn't to me.
Sure you can find a few dev boards or industrial embedded computing boards where the vendor happened to use the mini-itx form factor rather than something custom which is handy if you want to slap one of them in an off the shelf case for whatever reason. Such boards have been arround for some time.
But they lack the features one takes for granted on atom mini-itx boards like multiple SATA ports, a regular PCI or PCIe slot*, support for more ram (most arm boards have 1GB with the occasional board with 2GB just coming onto the market while most atom boards support 4GB with newer boards supporting 8GB). I also noticed that none of your links mentioned price. Shaving even a few tens of watts is not worth it if the board is double the price of an equivilent atom soloution.
Interestingly the Marvell armada XP platform can in principle offer most of what you find on atom mini-itx boards but I haven't seen anyone build such a board arround it. The only armada XP hardware i've seen is the crazily overpriced and somewhat strangely set-up (lots of network ports, hardly any SATA) openblocks stuff, the even more overpriced baserock slab and the vapourware dell copper..
*one of them has miniPCIe but that has a far smaller selection of cards and while in theory you can use adaptors in practice you'll never get it into a standard case if you do which kinda defeats the object of standard form factors
I think that his point is that, with CAD, even a trained monkey can tell the software "Just iterate through all the pretty little pictures we drew, multiply their volume by their density, and then add it all up" and arrive at a final weight.
Sure they can tell the software to do that.
But the CAD software is fundamentally only as good as the information put into it (the density of each material, what material is being used for what, how thick it will be, the weight of assemblies designed elsewhere and incorporated as black boxes and so-on), if that information is inaccurate then it can throw the total by a lot regardless of whether the design is being on paper or by software.
Automating things means less chance of arithmetic errors but it can also increase the chance that the calculations are applied blindly without anyone asking whether what is being calculated actually represents reality.
High speed DSL won't go far enough to get to the telephone exchange but it will go far enough to get to the cabinet in most urban settings at least.
The plan in the UK seems to be to do a mixture with the existing ADSL system continuing to serve the lower tier broadband users. FTTC+VDSL for the middle tier broadband users and FTTH for the top tier broadband users.
3: Those that have decided to charge per traffic charge get greedy and charge FAR more than a cent a gigabyte.
mmm, buisness "broadband" is only a small step up (if any) from home "broadband", mostly just a lifting of restrictions and a higher pricetag.
If you want a dedicated connection to the ISPs core network and then agreements that they will as far as reasonably practical run an uncongested core network and uncongested links to their upstreams and peering points then you can get that but expect to pay a hell of a lot more for it than you would pay for buisness broadband.
No, Sony confirmed that "PS4 games will be playable without an Internet connection". That pretty much nails it: this used game policy can only be enforced with an online DRM.
I don't see why.
You could have a writable area on the media that "bound" the copy to the first console it was used on. A game shop reselling the game would "unbind" the copy (and pay a fee for doing so) before reselling it. If you bought a copy privately you'd have to either take it into a game shop for unbinding or connect your console to the internet and pay the unbinding fee online. Or you could sell "unbinding cards" with little NFC chips in them that the user held up to their console to unbind the game. Or you could supply unbinding codes over the phone like how windows activation is handled without an internet connection.
In general the expectation is that internet client use will be bursty. When I want to move data I want it as fast as I can but that doesn't mean i'm going to be anywhere close to maxing out my internet connection 24/7. Having to buy connections priced on the assumption that they would be heavilly loaded 24/7 would make fast connections crazy expensive.
"unlimited" home internet connections are kinda like "all you can eat" buffets.
At an all you can eat buffet there is no direct limit on how much food you are allowed to eat but there are other rules that practically limit what you can eat. For example there is usually a maximum ammount of time you can stay and a rule that you can't take food away, a rule that everyone in the group has to pay and a rule that taking food out is forbidden. If they think you are consuming an ammount of food that would be impractical if you are following the rules they are going to investigate further.
Similarly an unlimited home internet connection, particaully one with a very high peak bandwidth will typically have rules that are designed to restrict users to normal client usage and if they think you are using an ammount of traffic that would be impractical with normal client usage they are going to investigate further.
A buisness connection has less rules on how you can use it and as a result they expect you to use more traffic and hence charge you a higher price and/or put explicit limits in place.
They could throw $50 billion at it and not destroy the network
If they want to destroy bitcoin then all they need to do is get more hasing power than the rest of the bitcoin network put together. I'm fairly sure that with the current level of hashing power that is within the reach of most governments.
Once they do that they can create arbitary rules for refusing transactions they don't like. For example they could require that all bitcoin addresses used as transaction destinations are registered with the government before use. They would enforce these rules by refusing to accept transactions that violate them into blocks they generate and refusing to use blocks that contain such transactions as a base for blocks they generate. Occasionally a non-governement miner may generate a block with your non-government approved transaction in it but the government miners would fork it off before it got any significant number of confirmations.
Bitcoin is and always has been psudononymous. It's trivial to track the flow of coins between bitcoin addresses (which are essentially psudonyms). So how much anonymity there is depends on how difficult it is to associate those bitcoin addresses with real people.
How difficult it is to associate those addresses with real people depends hugely on both what records are kept and how many significant players there are in the bitcoin market (it's much easier to force a small number of large entities to hand over records to the cops than it is to force a large number of small entities to do it)
You could potentially supply power to something using a male to male cord
There is a reason such cords are reffered to as widowmakers.
If you really want multiple inlets then I guess you could use connectors where both sides are touchproof but even then there would be a risk if someone hooked up too cords to the same peice of furniture. At the very least i'd want to see padlocks and warning notices in such a scenario.
http://www.osnews.com/story/25547/Apple_Restricts_Certain_APIs_to_Mac_App_Store_Applications
http://techcrunch.com/2012/09/29/windows-store-metro-mode-no-sideloading-psa/
IOS != macos
Both apple and MS now have smartphone and locked down arm based tablet platforms where they force people into app stores from which they take a substantial cut. Both apple and MS still have traditional desktop/laptop (and in MS's case bulky x86 based tablet) platforms where for the moment they still let you install apps in the traditional way. However even on those platforms afaict they limit some features of the platform to store apps only.
The difference is of course that with apple nowadays the desktops/laptops are a sideline and the phones/locked down arm based tablets are the main thrust of the buisness. With MS the OS for phones/locked down arm based tablets is a sideline and the desktop/laptop OS is the main thrust of the buisness.
Someone else in this discussion suggested cutting a notch in the edge of the card to destroy the antenna.
Not easilly.
Computers typically use air cooling, the exhaust temperature of a computer is not very far above the intake temperature and the intake temperature is typically arround normal room temperature or lower. So the exhaust temperature is likely to be barely above normal room temperature making moving the heat arround difficult.
You could raise the intake temperature to the computers but doing so would have significant disadvantages. Firstly it would reduce the ammount of time you had between cooling equipment failure and the temperature rising beyond the maximum safe level for the equipment. Secondly it may cause equipment that isn't designed to work in those temperatures to fail or at least reduce it's life. It would also make things rather uncomforable for people working in the datacenter.
You could also redesign the computers to use liquid cooling, since liquid cooling is far more efficient than air cooling you could run the loop at a significantly higher temperature than typical datacenter air temperatures while keeping the core temperature the same. The downside is of course you'd need to redesign the cooling systems in all your computers and come up with a system for safely adding and removing computers to/from the liquid cooling system.
It's messy :(
Sun released openoffice under the GPL (I don't remember the exact version offhand) and in the wake of the oracle takeover a number of developers decided to fork and produce libreoffice. While the oracle takeover was apparently the immediate trigger I get the impression that there were deep tensions within the project already.
Normally when such a fork happens one of a few things results
1: The two forks continue in paralell with code flowing from the original to the fork and possiblly back again (it would be difficult for it to flow back in this case though because of contributor agreement requirements).
2: The operators of the forked project ignores the fork but is either abandoned or at least development slows to a cral
3: The fork is blessed by the forked project to some greater or lesser extent.
In this case though none of those things happened. Instead oracle decided to give the rights for openoffice to apache who proceeded to release it under the apache license (a non-copyleft license) and afaict the version given to apache was not the same as the one libreoffice was based on.
Libreoffice in turn have decided that while they want copyleft they want one weaker than the GPL, so they are rebasing their work on top of apache openoffice. Afaict this means that code will be able to flow from apache openoffice to libreoffice but not vice-versa.
Heh,
I thought, "send them off for a few years of hard labour", then my brain started thinking though how such a sentance could work in the modern world. Forbidding their friends and family from contact completely would be too cruel IMO, OTOH having rich friends and family jetting in all the time and bring expensive gifts wouldn't seem like much of a punishment.
I don't belive conventional prison is appropriate for most white collar criminals but I do belive they need to have their luxuries taken away and be made to live a hard life for a while.