The problem with DIY/custom laptops is that they usually end up substantially more expensive overall than similarly configured serially-produced Windows laptops.
I also put together my own desktops because I like having absolute control over the box's content and the bottom-line cost is comparable to similar pre-assembled boxes. For laptops though, a similar custom would have cost me over $300 more... so, I saved $300 by paying the "Windows tax" and I got a sort-of-free copy of XP Home in the process.
Current DRAM chips were there years ago: current DRAM chips are around 1Gbit per square centimeter. On a ~5" wafer, this means ~40 potentially working chips per wafer and 40Gbits/wafer, four times as much.
And as far as downtime reduction goes, NRAM would be no good unless the server has time to suspend-to-RAM... so you would still need an UPS or ultra-capacitors to cover this.
I am particularly impressed by the way some 'upgrades' cost more than the equivalent after-market retail parts.
Back when I was shopping for a laptop on dell.ca, "upgrades" from 2x256MB to 1x512MB cost $150, pretty much the same price as an after-market 512MB SODIMM... and $400 for a 2x512MB upgrade when the after-market price for a 1GB SODIMM was $300. Same goes with HDDs... the 60GB to 80GB upgrade cost $200 but an after-market 80GB drive costs $150.
The USA site's prices make more sense but the Canadian site's prices (at least when I checked early this year) need a grossly needed reality check. I am surprised there have not been class-action suit to get refunds for what I would brand fraudulent pricing. (Paying more for an upgrade than the fair market value of the same/equivalent/superior after-market solution.)
I did a quick check about Linux and C++... it seems like even switching from gcc to g++ front-end caused headaches in the past by triggering incorrect code generation compiler bugs - but those were from the 1992-1996 period, GCC must have improved a little since then.
Finally, while Linus maintains the development kernel, he is the one who makes the final call. In case there are any doubts on what his opinion is, here is what he said in 2004:
In fact, in Linux we did try C++ once already, back in 1992.
It sucks. Trust me - writing kernel code in C++ is a BLOODY STUPID IDEA.
There is such a thing as 'dialects': variations of a language to suit specific needs, like machine-binary, embedded and real-time Javas. Vanilla Java is all bytecode within a contained runtime environment, it cannot have any direct binary interface with the outside world. The other Java dialects can do things that may be technically impossible or simply forbidden by plain Java.
As for Linux having C++ in it, with the number of times I have seen Linus quoted saying he would not put any C++ in the kernel, it seems unlikely that the kernel would contain actual C++. Since mixing plain C with C++ often causes maintenance nightmares, any C++ code in Linux should be in some isolated sub-module, well outside everything else's way.
How many 'close to the hardware' aspects of C could not be implemented in machine-binary Java?
Give machine-binary Java inline-ASM and that takes care of the 1% of situations where absolute control is necessary or justifiable. Or just link with external libraries like we already do in C/Pascal/etc.
Actually, since Java has some machine binary compilers, nothing stops people from rewriting most of the JRE in Java.
Before people could use text editors, someone had to input them as direct machine language... same thing goes even for ASM compilers. The first compiler and runtime environment generation has to be created from existing technologies but nothing (other than motivation or justification) is preventing it from coming full-circle.
How many pre-approved credit cards have you received over the last few years?
In many cases, all you need to activate a pre-approved card is to call a number and enter the activation code. Anyone who might go through your mailbox could potentially go away with it and use it. Most of the time, these come from super-stores promoting their branded cards but phishers could also get card issued by using collected information.
When activating a credit card that you may not have ever requested is this trivial, I see making card issuers fully liable as a perfect solution to unsollicited credit offers. At least three people I know got bills for credit cards they never had and had to pay the $500 maximum liability.
4) phishers use your SSN and other info to open bank account and apply for credit cards at other institutions where you have not registered your SecureID yet. 5) Repeat from #1.
Last time I read about region-coding, the DVD Forum said that Macrovision-free or region-less DVD players were not allowed to use the "DVD-Video" logos on their devices.
Maybe you are right and they do sell more expensive licenses that remove some of these requirements... but the licenses I read about require both. It would be possible for chinese/taiwanese/etc. manufacturers to male logo-less, macrovision-free and region-less players but these may not be legal for import (due to studio lobbying / DMCA and equivalents) in some key markets, like the USA.
I stopped buying into entertainment material when medias and studios started making asses of themselves with copyright infringement, region-coding, DRM and the rest. This is generally unfair for pretty much everyone beyond the studios. I'd rather be a 'pirate' than buy into lock-in/out policies.
At least, we are seeing increasingly more people get frustrated with incompatible DRM schemes and artificially non-interoperable systems/devices. This should help media-happy politicians realize that accepting media money for campaign financing can potentially ruin their careers in the long run.
Inductive charging itself is nothing new, is very simple and fairly efficient. All it requires is close range and decent alignment. IIRC, inductive chargers are already in use for some artificial hearts but they still need to change the battery every year or so due to premature capacity degradation from continuous high-drain operation. Without inductive chargers, people with artificial hearts would have to be opened up to replace batteries at least twice a week.
As far as diode 'losses' go, this is not about resistance. In any semiconductor junction (like diodes), there is a minimum energy (voltage) that electrons must have to jump across the junction. For ordinary silicon diodes are around 0.7V, schottky diodes can go as low as 0.2V (sometimes a little less under very low currents) and the lowest I have seen were germanium that went down to 0.12V.
Since batteries degrade even when they are in cold storage, even the best batteries need to be replaced or serviced every ~10 years. Because stimulators are sufficiently low-power to be limited by the battery's shelf life, it makes no sense to make these rechargeable. (Rechargeable batteries are far more likely to fail/leak and they have lower power densities.)
One could work around the battery issue by using capacitors but ultra-capacitors have far lower energy densities than batteries: a D-cell can provide up to 160kJ while a similarly sized ultra-capacitor (2.5V, 2600F) could provide about 15kJ at best. One problem with these is that they have relatively high leakage (self-discharge) current, somewhat like rechargeable batteries.
First, RF energy is fairly weak. Second, it is readily absorbed by tissues and water. Third, even the lowest forward voltage rectifiers still need more than 0.1V to conduct while antenna voltages are typically under 0.01V.
It would be possible to use a high-Q resonnant circuit to boost the voltage to a rectifiable level but such circuits are effective only over a very small frequency band which pretty much requires a tuned source, effectively ruling out most random external sources.
Since cell. phones transmit at most 1W while talking and are likely to be more than 1m away from such a coil and not directly in front of them, the power density would be under 1mW per square inch. Add skin absorbption (assuming the coil is directly under the skin) and you are left with less than 500uW, assuming there is always at least one phone operating at its maximum output power within 1m in front of your coiled side.
Possible? Yes, if the typical implant requires less than 100uW and there is always someone talking on his/her cell. phone within ~2m of you 24 hours per day, 365 days per year. Practical? No.
BTW, your North-America 1900MHz RF-charged pace-maker would not work in 1800MHz Europe and areas that either have no cell. service or have already upgraded to G3.
"Should the customer decide to cancel the service before the this contract's expiration, the company shall collect a $200 termination fee and recover the phone. Should the customer have lost the phone and be unable to return it within 30 days from termination, the company shall invoice an additional amount up to the cash discount given when the service was contracted."
When you use customized OSS running on some company's servers, how does having the source code as a user helping you fix bugs for yourself?
Even if you go through the code and fix bugs for yourself, this does not magically update the company's servers and there is no guarantee that the company would accept your patch. And unless you actually have a personnal interest in the modifications themselves, the customized code provides no advantages to you. If those mods are to interface proprietary hardware and network protocols, being forced to release the code could be a major liability both ways.
I personally despise the GPL's anti-integration nature. My ideal license would be betweem GPL and BSD: one where only contributions (changes, fixes, enhancements, etc.) to so-licensed code would have to be published. This way, open code stays open without 'infecting' everything that uses it.
Thankfully, most of the most useful libraries I have ever needed were either public-domain, copyleft or BSD-ish.
In eMule, there is an option to save logs to disk but it is not enabled by default.
If the RIAA followed due process, they would have filed a complaint with the RCMP or local police, the police would have investigated the lead and verified the *AA's claims, then they would have requested a warrant and raided the house.
The RIAA takes justice in its own hands to extort money from people through out-of-court settlements. Last time I checked, justice self-service and extortion were both illegal in the USA, Canada and most other civilized countries. All we need now is a good definition for the MAFIAA acronym to would accurately represent the MPAA/RIAA/etc.'s self-serving extortionist nature.
If I was planning something big, I would not be using off-the-shelf services... putting together a simple and secure point-to-point protocol is not horribly difficult. I guess the FBI's next step will be to ask for permission to arrest, interrogate and search people who used unknown protocols. (Sort of like the guy who got arrested because someone thought 'lynx' was a hacking tool... guess I should not use elinks too often.)
Supersizing an UPS has three advantages: 1- larger batteries 2- beefier components 3- more efficiency at low to intermediate loads
Of those three, #1 counts double: larger batteries not only have more capacity but they also lose less of their effective capacity for a given load current. So, a 50% larger battery can yield a 75% increase in usable backup power for a given load. The other two factors are relatively minor compared to this.
If you want to power a UPS with solar panels and some dynamo/alternator attached to a bike, you would need a battery or ultracapacitors to help regulate and store the energy. Otherwise, your UPS (and yourself if you implemented the bicycle aspect) would have a really hard time generating steady output and keep the attached equipment running properly.
As far as plugging a laptop in a 1.5kVA UPS for extended battery run, I would be somewhat surprised if 1.5kVA UPSes ever overheated at something like 10% load under otherwise normal operating conditions.
If I modded my BK1500 for extended run with external batteries, I would do a typical-load run-down test to verify that it can take it without overheating then add and test the 'missing' 60mm fan for peace of mind.
I dare you to find one high-power linear regulator in your PC... all but the lightest loads are powered from linear regulators now. With modern integrated SMPS chips, putting switchers everywhere is relatively inexpensive.
Look at CPU voltage regulators... the Prescott VRM must be able to provide over 100A to the socket at as little as 1.4V. This would mean your low-drop-out regulator would be dissipating up to 190W. In the real world, motherboard manufacturers use multi-phase switchers fed from 12V.
Look at video cards... all current boards have onboard switching regulators... they all feature a Maxim, TI or other power modulator switches, some surface-mount MOSFETs all mounted around inductors and filter capacitors.
Look at next-generation Pentium-M, these will have on-board PWM regulator.
Look at recent TI newsletters, there have been lots of promos for chips designed for point-of-load power regulation.
Switchers are also into most voltage-sensitive devices to buck/boost battery voltage so they can be used more effectively and completely.
Switchers are everywhere and much more practical than you appear to think.
If I put a nano in a rear pocket, scratching the display would be the least of my worries and very much in line with cracked (rather than simply scratched) displays.
eDonkey is only a client application, there are plenty of independent eD2k servers out there that will continue to operate without it.
eMule also has it own kademlia network for distributed content indexing but it requires a server to fetch some clients to bootstrap off of - very much like Gnutella. If you do not mind hunting down a bootstrap IP:Port yourself and entering it manually, then you can use Kad without eD2k servers.
Hook up your UPS to a two 70AH car batteries and let it run unplugged at full-load for an hour, it will be hotter. I read many stories about UPSes overheating or even catching fire when run for too long beyond what their stock batteries would normally allow at full-load. UPSes meant for extended runs have fans... look at APC Back-UPS XS and RS series... the XS has no external battery connector and no fan while the equivalent RS has both the external battery connector and a fan.
Laptop bricks only provide 12-20VDC, the laptop itself has a number of extra converters in it to provide all the other voltages (Vcore, Vram, Vio, Vterm, 3.3V, 5V, etc.) from the battery and charge-controller circuitry. The laptops' external brick is there to decouple non-essential AC power circuitry from the mobile components, the laptop still requires local bulk power regulation. There is also the problem that an external PSU would have slower transient response times. With today's systems where the load can change by 10A in microseconds or less, an external PSU would probably need a secondary regulator (at least a large capacitor bank) inside the case. For a laptop, this is not an issue because everything goes through the battery controller. For ITX, this is not as much of an issue because they are mostly low-power systems.
Since technology is moving towards local voltage regulators for faster transient response, PCs should migrate towards single-rail power distribution (something like 24V with 20-35V tolerance for easy UPS) to avoid triple, quadruple, etc.-tuple conversions... converting directly from a single higher voltage source decreases the load on intermediate regulators, reducing conduction and switching losses across the board.
PSUs with efficiencies over 90% are possible but every 1% over 85% is more expensive than the last... synchronous rectification alone doubles the number of required high-speed, high-current MOSFETs and other parts.
3) To get media endorsement for the PSP, Sony has to demonstrate that the media and platform are a perfect lock-in solution. FALSE
Ok, maybe not 'perfect' lock-in (should have said 'secure' instead of 'lock-in') but it is much easier to sign deals with movie/record studios and game publishers if you have a reasonably hack-proof device and some means to upgrade the security model after it has been compromised.
Game publishers go with consoles because of: 1- reduced development and support costs (standard hardware and software environment) 2- reduced copyright violations (mod-chip or other non-obvious work-around required) 3- better accessibility to the general public (PC gamers are a large minority)
In today's DRM-obsessed world, I am betting that locking down consoles as best as possible is a priority... I would not be surprised if the major publishers' contracts included a compensation clause for software hacks that may work around or disable content protection.
Slashdot Mirror
The problem with DIY/custom laptops is that they usually end up substantially more expensive overall than similarly configured serially-produced Windows laptops.
I also put together my own desktops because I like having absolute control over the box's content and the bottom-line cost is comparable to similar pre-assembled boxes. For laptops though, a similar custom would have cost me over $300 more... so, I saved $300 by paying the "Windows tax" and I got a sort-of-free copy of XP Home in the process.
Why would you do that?
Current DRAM chips were there years ago: current DRAM chips are around 1Gbit per square centimeter. On a ~5" wafer, this means ~40 potentially working chips per wafer and 40Gbits/wafer, four times as much.
And as far as downtime reduction goes, NRAM would be no good unless the server has time to suspend-to-RAM... so you would still need an UPS or ultra-capacitors to cover this.
I am particularly impressed by the way some 'upgrades' cost more than the equivalent after-market retail parts.
Back when I was shopping for a laptop on dell.ca, "upgrades" from 2x256MB to 1x512MB cost $150, pretty much the same price as an after-market 512MB SODIMM... and $400 for a 2x512MB upgrade when the after-market price for a 1GB SODIMM was $300. Same goes with HDDs... the 60GB to 80GB upgrade cost $200 but an after-market 80GB drive costs $150.
The USA site's prices make more sense but the Canadian site's prices (at least when I checked early this year) need a grossly needed reality check. I am surprised there have not been class-action suit to get refunds for what I would brand fraudulent pricing. (Paying more for an upgrade than the fair market value of the same/equivalent/superior after-market solution.)
There is also a section of the Linux Kernel Mailing List about C++ in the kernel there: http://www.tux.org/lkml/#s15-3
There is such a thing as 'dialects': variations of a language to suit specific needs, like machine-binary, embedded and real-time Javas. Vanilla Java is all bytecode within a contained runtime environment, it cannot have any direct binary interface with the outside world. The other Java dialects can do things that may be technically impossible or simply forbidden by plain Java.
As for Linux having C++ in it, with the number of times I have seen Linus quoted saying he would not put any C++ in the kernel, it seems unlikely that the kernel would contain actual C++. Since mixing plain C with C++ often causes maintenance nightmares, any C++ code in Linux should be in some isolated sub-module, well outside everything else's way.
How many 'close to the hardware' aspects of C could not be implemented in machine-binary Java?
Give machine-binary Java inline-ASM and that takes care of the 1% of situations where absolute control is necessary or justifiable. Or just link with external libraries like we already do in C/Pascal/etc.
Actually, since Java has some machine binary compilers, nothing stops people from rewriting most of the JRE in Java.
Before people could use text editors, someone had to input them as direct machine language... same thing goes even for ASM compilers. The first compiler and runtime environment generation has to be created from existing technologies but nothing (other than motivation or justification) is preventing it from coming full-circle.
How many pre-approved credit cards have you received over the last few years?
In many cases, all you need to activate a pre-approved card is to call a number and enter the activation code. Anyone who might go through your mailbox could potentially go away with it and use it. Most of the time, these come from super-stores promoting their branded cards but phishers could also get card issued by using collected information.
When activating a credit card that you may not have ever requested is this trivial, I see making card issuers fully liable as a perfect solution to unsollicited credit offers. At least three people I know got bills for credit cards they never had and had to pay the $500 maximum liability.
4) phishers use your SSN and other info to open bank account and apply for credit cards at other institutions where you have not registered your SecureID yet.
5) Repeat from #1.
Players trivial to unlock are relatively easy to get but players explicitly marketed as regiion-free and shipped as such is a different matter.
Last time I read about region-coding, the DVD Forum said that Macrovision-free or region-less DVD players were not allowed to use the "DVD-Video" logos on their devices.
Maybe you are right and they do sell more expensive licenses that remove some of these requirements... but the licenses I read about require both. It would be possible for chinese/taiwanese/etc. manufacturers to male logo-less, macrovision-free and region-less players but these may not be legal for import (due to studio lobbying / DMCA and equivalents) in some key markets, like the USA.
I stopped buying into entertainment material when medias and studios started making asses of themselves with copyright infringement, region-coding, DRM and the rest. This is generally unfair for pretty much everyone beyond the studios. I'd rather be a 'pirate' than buy into lock-in/out policies.
At least, we are seeing increasingly more people get frustrated with incompatible DRM schemes and artificially non-interoperable systems/devices. This should help media-happy politicians realize that accepting media money for campaign financing can potentially ruin their careers in the long run.
Inductive charging itself is nothing new, is very simple and fairly efficient. All it requires is close range and decent alignment. IIRC, inductive chargers are already in use for some artificial hearts but they still need to change the battery every year or so due to premature capacity degradation from continuous high-drain operation. Without inductive chargers, people with artificial hearts would have to be opened up to replace batteries at least twice a week.
As far as diode 'losses' go, this is not about resistance. In any semiconductor junction (like diodes), there is a minimum energy (voltage) that electrons must have to jump across the junction. For ordinary silicon diodes are around 0.7V, schottky diodes can go as low as 0.2V (sometimes a little less under very low currents) and the lowest I have seen were germanium that went down to 0.12V.
Since batteries degrade even when they are in cold storage, even the best batteries need to be replaced or serviced every ~10 years. Because stimulators are sufficiently low-power to be limited by the battery's shelf life, it makes no sense to make these rechargeable. (Rechargeable batteries are far more likely to fail/leak and they have lower power densities.)
One could work around the battery issue by using capacitors but ultra-capacitors have far lower energy densities than batteries: a D-cell can provide up to 160kJ while a similarly sized ultra-capacitor (2.5V, 2600F) could provide about 15kJ at best. One problem with these is that they have relatively high leakage (self-discharge) current, somewhat like rechargeable batteries.
This does not work particularly well.
First, RF energy is fairly weak. Second, it is readily absorbed by tissues and water. Third, even the lowest forward voltage rectifiers still need more than 0.1V to conduct while antenna voltages are typically under 0.01V.
It would be possible to use a high-Q resonnant circuit to boost the voltage to a rectifiable level but such circuits are effective only over a very small frequency band which pretty much requires a tuned source, effectively ruling out most random external sources.
Since cell. phones transmit at most 1W while talking and are likely to be more than 1m away from such a coil and not directly in front of them, the power density would be under 1mW per square inch. Add skin absorbption (assuming the coil is directly under the skin) and you are left with less than 500uW, assuming there is always at least one phone operating at its maximum output power within 1m in front of your coiled side.
Possible? Yes, if the typical implant requires less than 100uW and there is always someone talking on his/her cell. phone within ~2m of you 24 hours per day, 365 days per year. Practical? No.
BTW, your North-America 1900MHz RF-charged pace-maker would not work in 1800MHz Europe and areas that either have no cell. service or have already upgraded to G3.
Region-dependent life-saving devices? I'll pass.
There is a simple solution to this:
"Should the customer decide to cancel the service before the this contract's expiration, the company shall collect a $200 termination fee and recover the phone. Should the customer have lost the phone and be unable to return it within 30 days from termination, the company shall invoice an additional amount up to the cash discount given when the service was contracted."
When you use customized OSS running on some company's servers, how does having the source code as a user helping you fix bugs for yourself?
Even if you go through the code and fix bugs for yourself, this does not magically update the company's servers and there is no guarantee that the company would accept your patch. And unless you actually have a personnal interest in the modifications themselves, the customized code provides no advantages to you. If those mods are to interface proprietary hardware and network protocols, being forced to release the code could be a major liability both ways.
I personally despise the GPL's anti-integration nature. My ideal license would be betweem GPL and BSD: one where only contributions (changes, fixes, enhancements, etc.) to so-licensed code would have to be published. This way, open code stays open without 'infecting' everything that uses it.
Thankfully, most of the most useful libraries I have ever needed were either public-domain, copyleft or BSD-ish.
In eMule, there is an option to save logs to disk but it is not enabled by default.
If the RIAA followed due process, they would have filed a complaint with the RCMP or local police, the police would have investigated the lead and verified the *AA's claims, then they would have requested a warrant and raided the house.
The RIAA takes justice in its own hands to extort money from people through out-of-court settlements. Last time I checked, justice self-service and extortion were both illegal in the USA, Canada and most other civilized countries. All we need now is a good definition for the MAFIAA acronym to would accurately represent the MPAA/RIAA/etc.'s self-serving extortionist nature.
You're right... my mind slipped up on that one.
I shouldn't think too many sentenses ahead, that's what usually happens when I do.
If I was planning something big, I would not be using off-the-shelf services... putting together a simple and secure point-to-point protocol is not horribly difficult. I guess the FBI's next step will be to ask for permission to arrest, interrogate and search people who used unknown protocols. (Sort of like the guy who got arrested because someone thought 'lynx' was a hacking tool... guess I should not use elinks too often.)
Supersizing an UPS has three advantages:
1- larger batteries
2- beefier components
3- more efficiency at low to intermediate loads
Of those three, #1 counts double: larger batteries not only have more capacity but they also lose less of their effective capacity for a given load current. So, a 50% larger battery can yield a 75% increase in usable backup power for a given load. The other two factors are relatively minor compared to this.
If you want to power a UPS with solar panels and some dynamo/alternator attached to a bike, you would need a battery or ultracapacitors to help regulate and store the energy. Otherwise, your UPS (and yourself if you implemented the bicycle aspect) would have a really hard time generating steady output and keep the attached equipment running properly.
As far as plugging a laptop in a 1.5kVA UPS for extended battery run, I would be somewhat surprised if 1.5kVA UPSes ever overheated at something like 10% load under otherwise normal operating conditions.
If I modded my BK1500 for extended run with external batteries, I would do a typical-load run-down test to verify that it can take it without overheating then add and test the 'missing' 60mm fan for peace of mind.
I dare you to find one high-power linear regulator in your PC... all but the lightest loads are powered from linear regulators now. With modern integrated SMPS chips, putting switchers everywhere is relatively inexpensive.
Look at CPU voltage regulators... the Prescott VRM must be able to provide over 100A to the socket at as little as 1.4V. This would mean your low-drop-out regulator would be dissipating up to 190W. In the real world, motherboard manufacturers use multi-phase switchers fed from 12V.
Look at video cards... all current boards have onboard switching regulators... they all feature a Maxim, TI or other power modulator switches, some surface-mount MOSFETs all mounted around inductors and filter capacitors.
Look at next-generation Pentium-M, these will have on-board PWM regulator.
Look at recent TI newsletters, there have been lots of promos for chips designed for point-of-load power regulation.
Switchers are also into most voltage-sensitive devices to buck/boost battery voltage so they can be used more effectively and completely.
Switchers are everywhere and much more practical than you appear to think.
Was that a rear or front pocket?
If I put a nano in a rear pocket, scratching the display would be the least of my worries and very much in line with cracked (rather than simply scratched) displays.
eDonkey is only a client application, there are plenty of independent eD2k servers out there that will continue to operate without it.
eMule also has it own kademlia network for distributed content indexing but it requires a server to fetch some clients to bootstrap off of - very much like Gnutella. If you do not mind hunting down a bootstrap IP:Port yourself and entering it manually, then you can use Kad without eD2k servers.
Hook up your UPS to a two 70AH car batteries and let it run unplugged at full-load for an hour, it will be hotter. I read many stories about UPSes overheating or even catching fire when run for too long beyond what their stock batteries would normally allow at full-load. UPSes meant for extended runs have fans... look at APC Back-UPS XS and RS series... the XS has no external battery connector and no fan while the equivalent RS has both the external battery connector and a fan.
Laptop bricks only provide 12-20VDC, the laptop itself has a number of extra converters in it to provide all the other voltages (Vcore, Vram, Vio, Vterm, 3.3V, 5V, etc.) from the battery and charge-controller circuitry. The laptops' external brick is there to decouple non-essential AC power circuitry from the mobile components, the laptop still requires local bulk power regulation. There is also the problem that an external PSU would have slower transient response times. With today's systems where the load can change by 10A in microseconds or less, an external PSU would probably need a secondary regulator (at least a large capacitor bank) inside the case. For a laptop, this is not an issue because everything goes through the battery controller. For ITX, this is not as much of an issue because they are mostly low-power systems.
Since technology is moving towards local voltage regulators for faster transient response, PCs should migrate towards single-rail power distribution (something like 24V with 20-35V tolerance for easy UPS) to avoid triple, quadruple, etc.-tuple conversions... converting directly from a single higher voltage source decreases the load on intermediate regulators, reducing conduction and switching losses across the board.
PSUs with efficiencies over 90% are possible but every 1% over 85% is more expensive than the last... synchronous rectification alone doubles the number of required high-speed, high-current MOSFETs and other parts.
3) To get media endorsement for the PSP, Sony has to demonstrate that the media and platform are a perfect lock-in solution.
FALSE
Ok, maybe not 'perfect' lock-in (should have said 'secure' instead of 'lock-in') but it is much easier to sign deals with movie/record studios and game publishers if you have a reasonably hack-proof device and some means to upgrade the security model after it has been compromised.
Game publishers go with consoles because of:
1- reduced development and support costs (standard hardware and software environment)
2- reduced copyright violations (mod-chip or other non-obvious work-around required)
3- better accessibility to the general public (PC gamers are a large minority)
In today's DRM-obsessed world, I am betting that locking down consoles as best as possible is a priority... I would not be surprised if the major publishers' contracts included a compensation clause for software hacks that may work around or disable content protection.
Wasn't two-digit years the main reason behind the Y2K gold-rush?
But yeah, the near-future context would have resolved this sort of ambiguity.