The complicating factor, when trying to make any useful statements about "What Operating System Whatever does on SSDs" is that SSDs are totally free to do whatever crazy stuff internally, so long as they can present a coherent block device abstraction over the IDE/SATA/etc. bus.
TRIM happens to make the design problem easier; because it allows you to throw data on the floor when the OS says that they are no longer needed, rather than having to treat everything that isn't explicitly deleted as still good. However, before TRIM was available, manufacturers tended to do their best to design around its absence. This generally had a cost(either monetary, with lots of reserve flash driving up the cost, performance, with the drive either starting fast and taking a nosedive or just not being all that fast, or predictability, a given write operation could take milliseconds or it could take literal seconds, depending on the drive state).
I would agree that the desire for "cheap" is arguably behind this problem; but I would disagree about "shiny". The problem isn't that the protocol is general purpose(particularly in those cases where the receiver was sold in a set that contained a mouse and/or keyboard in addition to the little PPT remote...) but that absolutely no useful effort was made to apply what we already know about authentication and encryption. For just slightly more, you could just have a bluetooth device that(while certainly not free of security issues throughout its history) at least takes security into account, and isn't some poor bastard who knows very little about crypto reinventing the wheel at the last second.
In many of these cases the little proprietary receiver dongle accepts arbitrary keystrokes, not just the ones that the remote has buttons for, because it is exactly the same item as the one being sold(under that brand, or one or more others) in a package with a wireless keyboard and often a mouse as well. Some kits come with everything in one box, receiver, keyboard, mouse, little powerpoint clicker widget.
In other cases, I imagine, the engineer in charge of knocking together the receiver unit (correctly) realized that implementing a general-purpose system for taking arbitrary keycodes encapsulated in whatever the proprietary RF protocol is and dumping them to the host system just like any USB HID device wouldn't be much harder than implementing just the 6 keycodes found on revision 1 of Product X and would save him from having to do it again when revision 1.1 adds another couple of buttons, and revision 2.0 has to have a special button for the ribbon interface, or whatever it happens to be.
I wonder where the good Mr. Groves stands on Intel's large R&D base in Israel... Should we be taxing that, or is this only a tax for his dirty competitors who are fabbing with TSMC?
I'm guessing that this will end "climategate" about as well as further scientific research has managed to shut up the mercury-militia/autism-antivaxers.
Somebody really should have told that hack Picasso that if it isn't PG it isn't art before he went and wasted his time on Guernica...
(this is not to say that Doom is art, indeed, it probably isn't, being roughly on the level of some of the more derivative Rambo clones; but the "ooh, look, it's too bloody and gross to be art!!!" gambit is unbelievably pathetic. Pretty much every "real artist" who isn't painting pastoral landscapes or naked women is busy painting assorted martyrdoms, battles, famous assasinations, and the like.)
Incidentally, why hasn't somebody(either a Valve person for publicity, or a 3rd party just because they can) painted "Ravenholm" in cubist-horror style yet? (or, for extra credit, done a game in that graphical style)...
Even there, unless you want "teacher/institution quality" metrics to just be a referendum on unemployment rates and consumer confidence indices at 5 years after graduation(consider, for instance, the.com bubble. It made a lot of CS grads very happy indeed, and its bursting made a lot of CS grads very unhappy; but, if anything, the happier the CS grads in the "real world" were, the shittier the CS programs were becoming, because they were filling with people who figured that a CS degree was a ticket to easy street.), you would really need to build a metric that compares relative to peers at other institutions, ideally peers who are intellectually as similar as possible.
Arguably, his stupidity was in saying that "video games can never be art". Saying that "no video game has yet reached the level of art" is controversial; but a respectable enough empirical opinion, particularly back in the bad old days.
To say that they can "never" be art is either to make a stupid and almost certainly wrong prediction about the technological future, or to attempt to impose a definition of "art" so special-purpose that the statement "video games can never be art" is basically just a tautology masquerading as an insight.
For only 50-100% more than you have been paying for your misshapen generics, Microsoft's battery partners are proud to announce their new line of Premium Dimensionally Certified(tm) batteries: "Because that widget was expensive, and you wouldn't want something to happen."
The thing to remember about "bullshit" courses of study is that, while being mediocre at them is a very low effort enterprise(so, if you are at a school where passing mediocre students is standard practice, they are among the easier degrees), their difficultly actually ramps up enormously at the high end precisely because they are "bullshit".
Without any sort of real-world objectives, where the universe sets the difficultly level(i.e. Write a program that does X without crashing all the damn time, launch a rocket, grow a potato that isn't full of horrid worms, successfully remove a spleen without the patient bleeding out), the exercise consists largely of competing against others in the field, constantly striving to be cleverer and more novel than they are. This leads to heavy jargon-churn, constant need for fad adaptation, and other such fun. Since bullshit fields don't really have "workaday" problems that actually need to be solved but don't require novel brilliance, those who enter them either just leave and do something else entirely, or keep treading water against their peers until they retire.
Ah; but that would ruin the university and/or lecturer's numbers. And, thanks to various naive, "Hey, let's run this school like a business, punish failures, rewards successes" schemes, you can look bad because you failed too many people, regardless of whether they deserved it or not. Nobody seems to have figured out a quality metric that manages to capture "your quality, as expressed by the delta between the performance of these students under your tutelage vs. their hypothetical performance under other conditions" rather than a basic "what grades did your students get?"(the latter, perversely, makes people who provide honest feedback about bad performance look like bad teachers, while rewarding those who provide dishonest feedback about bad performance. Clearly an excellent metric.)
The less polite; but not substantially less accurate, version of TFA's headline might be "Qualcomm Maker attempts to pass off responsibility for kernel-mode portion of 3D Snapdragon driver"...
While it would certainly be nice if scientists, as a class, were better at public communication, I think that this consideration misses an important point:
If somebody happens to be the best available information source on a given issue, failure to communicate with them is a major failing on your part.
All men may be created equal; but only some of them are worth consulting for advice.
I still want to know what Intel has up their sleeve with light peak.
10Gb/s over fiber isn't new or all that interesting; but, in the networking world, it isn't all that cheap. What have they done to make the equivalent of shoving a 10GbE fiber interface into random bits of cheap consumer electronics remotely viable?
Second, while neither optical cables nor optical connectors are quite in "will die if you give them a funny look" territory, they definitely won't stand up to the kind of abuse that even ghastly quality copper will, never mind being rolled over by chairs and filled with pocket fuzz. What do they have in mind?
It varies. The rated wattage(which is sometimes even not a lie) of "enthusiast" power supplies has been climbing steadily, as the ability to purchase and cool seriously toasty chips, often several per system, has become substantially cheaper and more widespread. Though, it should always be noted that "rated wattage" means "power it can deliver without Something Bad happening" not "Power it actually uses". The rated wattage of high-density servers has increased substantially for similar reasons, and those wattage ratings usually aren't lies.
On the other hand, performance-per-watt has been improving markedly over the years, and the efficiency of non-crap PSUs has made decent strides. Also, the ability of modern chips(CPUs and GPUs particularly, though often smaller stuff as well) to intelligently throttle themselves when unused or lightly used has improved pretty markedly.
Thus, it is basically impossible to say whether "old computers use more power than new ones". It's just too general a statement to be meaningful. On the one hand, nobody on the high end blinks at a 150watt TDP processor, where(back in the day), hitting 60 watts on your crazy overclock was considered seriously hardcore. On the other hand, an ever increasing proportion of computers in actual use are laptops which cram the entire computer, and a monitor, into a 60watt or less AC adapter, and conserve power on idle like their (battery) lives depend on it.
In general, the ability of people, who don't have $250,000 to spare and a datacenter to house the beast, to buy very energy-hungry computers has increased significantly over time(1.5 kilowatt PSUs definitely didn't use to be retail items). On the other hand, though, performance per watt has gone through the roof(and CRTs have largely gone out the window). Whether ditching an old piece of gear for a new one is usually a case-by-case thing. If your mid-90's junker can be replaced by a weedy little plastic ARM box that runs off a wall wart that weighs under a hundred grams, the replacement almost certainly saves energy. Replacing a late-model PIII and a 17-inch CRT with a new Quad-core and a 28-inch LCD probably won't(though you will get surprisingly close to breaking even, and the performance will be a lot better).
This is why the energy economics of using obsolete x86s as networking gear are usually pretty weak, while those of getting rid of generic business desktop c. 1999 just to replace it with generic business desktop c. 2009 probably won't be nearly as exciting(unless the ACPI on the old box was painfully broken, as it not infrequently was)...
This would probably work. It has the look of an industrial/specialty system, though, so it could easily double the cost of the project.
Also worthy of consideration, a fairly wide variety of mini-ITX and smaller motherboards, particularly those designed for industrial/kiosk/signage applications actually include LVDS headers right on the board. This is of no use to you if you absolutely need a standard desktop board, or a spiffy graphics card; but if your option is either "Spend 80-150 for a motherboard. Spend 200-400 for a specialty DVI to LVDS board because you are buying quantity 1, have it dangle awkwardly." or "Spend 200-400 for a motherboard with LDVS natively." the second starts to look pretty good...
If this sounds attractive, Logicsupply.com has a good range. They may or may not have the best price on a given board; but they have an excellent collection all in one place.
There is no "mode control" as such. Basically, the way it works is that if the backlight is brighter than the ambient light falling on the screen, it is in the transmissive "mode"(which gives you lower resolution; but full color. If the ambient light is strong enough to wash out the backlight, or the backlight is off, you are in the "reflective" mode, which is much sharper; but B/W. At least in the version used in the OLPC, there are some intermediate states, where you get enough backlight for a hint of color; but the ambient light is strong enough that it looks sort of like a B/W image with pastel shading, rather than a full color image.
On any existing laptop, you "control the mode" with whatever mechanism you would ordinarily control the display backlight with. Turn the backlight down low enough, and you are "in reflective mode". Crank it up and you are "In transmissive mode"(unless you are in direct sunlight, in which you probably don't have a choice.
Any device designed with this display in mind would, in all likelyhood, give you slightly more "audio-like" backlight control(i.e. There would be the usual intensity control; but there would also be a discrete on/off toggle, so you could set your preferred indoor level, then hit just one button and save a whole lot of power while you are outside, then give the same button a tap to go back to your preferred level when you go back indoors). In any case, though, there is no magic "mode bit" that needs to be set, or special command, or pin that has to be toggled or anything. Purely a "however you would normally control the backlight" thing. The only real design difference is that, unlike a normal LCD, this thing can actually be useful with the backlight totally off, so you will probably want to incorporate a somewhat wider than usual range of brightnesses and/or an audio style "mute" button...
The Pixel Qi screens are the evolutionary descendants of the OLCP ones, and might well have not traded off cost and quality quite as agressively; but I'm assuming that at least half of this price is the "It's a quantity-one sale of a previously unavailable item to a cost-insensitive enthusiast" premium.
The ruthless margin-slashers who do purchasing for the big OEMs are just going to give you a thousand-yard-stare and a hollow laugh if your quoted price is much above a standard LCD of the same size, so Pixel Qi are either utter morons, or offer much more reasonable terms in quantities of 10,000+
Plus, while the maker shed is a noble operation, and sometimes a useful place to get stuff that would be hard to find in small quantities elsewhere, they aren't what you would call an "everyday low prices" kind of operation...
It's such a pity that they never settled on a standardized physical connector for the LVDS+backlight power connection that virtually all laptop screens use(are there any internal displayport devices in the wild yet?) Electrically, they are usually much the same, at least within a given size class(obviously, the current required for a 17inch DTR LCD backlight is going to be a little bit higher than that needed for a 8 inch netbook LCD, so a diffferent connector might be needed); but there was no real standardization. For basic economic reasons, and the fact that there are fewer OEMs than there are brand names, there are a lot of identical connectors lurking out there if you take a screwdriver to the problem; but there is nothing resembling a proper, consumer accessible, "standard", on the order of DVI or molex...
It would certainly have been even more superior-looking when compared to the sort of handsets that tracfone deals in; but they still would have had to solve the basic cost-of-data problem.
If the various Kin-social-integration-cloud-data stuff consumes smartphone levels of data, carriers are going to want to charge smartphone-level prices for it, prepaid or contract. Customers, sensibly enough, are going to want a smartphone if they are paying smartphone rates for connectivity.
Since the prepaid services tend to cater to a mixture of thrifty-people-who-don't-care-much in all income brackets, along with poor people and people with questionable credit ratings, the data costs would continue to be an issue. The handset competition would be less fierce, because the prepaid guys mostly get stuck with some real crap; but prepaid customers are probably, in aggregate, even less excited than verizon customers about paying stiff charges for a non-smartphone.
Or, in a view where not quite everything is moved by nefarious wheels-within-wheels or filled with secret Combine electronics, it could simply be that it made good business sense for Fark to move into the niche that the Drudge Report had no intention of even trying to fill. Competition is painful and can be dangerous. You frequently see such attempts to segment a market.
Which actually makes having killed it seem like a pretty decent idea.
Well golly, should I choose this here net-book that does absolutely nothing unless my Palm brand smartphone is within a couple of meters, or should I choose this one that is pretty much like a normal computer, only smaller and cheaper and slower, that I can tether my phone to if I really want to....
Had Foleo been some sort of WebOS device that could either function alone like a normal netbook, syncing data and favorites and stuff back to your phone the next time they were in contact and, if the two were directly communicating with each other, the user being able to "pass" running application cards back and forth between the two, that would have kicked ass. As it was, though, Foleo was sort of the last gasp of old-school palm, offering an expensive and inflexible way to give your palm phone a larger screen and keyboard. They might as well have tied it to the railroad tracks as released it in front of the netbook explosion.
Perhaps we can conclude that the CEO of Microsoft then and the lead on Bob had a better "working relationship" than do the CEO of Microsoft now and the lead on Kin?
Slashdot Mirror
The complicating factor, when trying to make any useful statements about "What Operating System Whatever does on SSDs" is that SSDs are totally free to do whatever crazy stuff internally, so long as they can present a coherent block device abstraction over the IDE/SATA/etc. bus.
TRIM happens to make the design problem easier; because it allows you to throw data on the floor when the OS says that they are no longer needed, rather than having to treat everything that isn't explicitly deleted as still good. However, before TRIM was available, manufacturers tended to do their best to design around its absence. This generally had a cost(either monetary, with lots of reserve flash driving up the cost, performance, with the drive either starting fast and taking a nosedive or just not being all that fast, or predictability, a given write operation could take milliseconds or it could take literal seconds, depending on the drive state).
I would agree that the desire for "cheap" is arguably behind this problem; but I would disagree about "shiny". The problem isn't that the protocol is general purpose(particularly in those cases where the receiver was sold in a set that contained a mouse and/or keyboard in addition to the little PPT remote...) but that absolutely no useful effort was made to apply what we already know about authentication and encryption. For just slightly more, you could just have a bluetooth device that(while certainly not free of security issues throughout its history) at least takes security into account, and isn't some poor bastard who knows very little about crypto reinventing the wheel at the last second.
In many of these cases the little proprietary receiver dongle accepts arbitrary keystrokes, not just the ones that the remote has buttons for, because it is exactly the same item as the one being sold(under that brand, or one or more others) in a package with a wireless keyboard and often a mouse as well. Some kits come with everything in one box, receiver, keyboard, mouse, little powerpoint clicker widget.
In other cases, I imagine, the engineer in charge of knocking together the receiver unit (correctly) realized that implementing a general-purpose system for taking arbitrary keycodes encapsulated in whatever the proprietary RF protocol is and dumping them to the host system just like any USB HID device wouldn't be much harder than implementing just the 6 keycodes found on revision 1 of Product X and would save him from having to do it again when revision 1.1 adds another couple of buttons, and revision 2.0 has to have a special button for the ribbon interface, or whatever it happens to be.
I wonder where the good Mr. Groves stands on Intel's large R&D base in Israel... Should we be taxing that, or is this only a tax for his dirty competitors who are fabbing with TSMC?
I'm guessing that this will end "climategate" about as well as further scientific research has managed to shut up the mercury-militia/autism-antivaxers.
This is to say, not at all.
Somebody really should have told that hack Picasso that if it isn't PG it isn't art before he went and wasted his time on Guernica...
(this is not to say that Doom is art, indeed, it probably isn't, being roughly on the level of some of the more derivative Rambo clones; but the "ooh, look, it's too bloody and gross to be art!!!" gambit is unbelievably pathetic. Pretty much every "real artist" who isn't painting pastoral landscapes or naked women is busy painting assorted martyrdoms, battles, famous assasinations, and the like.)
Incidentally, why hasn't somebody(either a Valve person for publicity, or a 3rd party just because they can) painted "Ravenholm" in cubist-horror style yet? (or, for extra credit, done a game in that graphical style)...
Even there, unless you want "teacher/institution quality" metrics to just be a referendum on unemployment rates and consumer confidence indices at 5 years after graduation(consider, for instance, the .com bubble. It made a lot of CS grads very happy indeed, and its bursting made a lot of CS grads very unhappy; but, if anything, the happier the CS grads in the "real world" were, the shittier the CS programs were becoming, because they were filling with people who figured that a CS degree was a ticket to easy street.), you would really need to build a metric that compares relative to peers at other institutions, ideally peers who are intellectually as similar as possible.
Arguably, his stupidity was in saying that "video games can never be art". Saying that "no video game has yet reached the level of art" is controversial; but a respectable enough empirical opinion, particularly back in the bad old days.
To say that they can "never" be art is either to make a stupid and almost certainly wrong prediction about the technological future, or to attempt to impose a definition of "art" so special-purpose that the statement "video games can never be art" is basically just a tautology masquerading as an insight.
A problem? I think you mean a golden opportunity!
For only 50-100% more than you have been paying for your misshapen generics, Microsoft's battery partners are proud to announce their new line of Premium Dimensionally Certified(tm) batteries: "Because that widget was expensive, and you wouldn't want something to happen."
The thing to remember about "bullshit" courses of study is that, while being mediocre at them is a very low effort enterprise(so, if you are at a school where passing mediocre students is standard practice, they are among the easier degrees), their difficultly actually ramps up enormously at the high end precisely because they are "bullshit".
Without any sort of real-world objectives, where the universe sets the difficultly level(i.e. Write a program that does X without crashing all the damn time, launch a rocket, grow a potato that isn't full of horrid worms, successfully remove a spleen without the patient bleeding out), the exercise consists largely of competing against others in the field, constantly striving to be cleverer and more novel than they are. This leads to heavy jargon-churn, constant need for fad adaptation, and other such fun. Since bullshit fields don't really have "workaday" problems that actually need to be solved but don't require novel brilliance, those who enter them either just leave and do something else entirely, or keep treading water against their peers until they retire.
Ah; but that would ruin the university and/or lecturer's numbers. And, thanks to various naive, "Hey, let's run this school like a business, punish failures, rewards successes" schemes, you can look bad because you failed too many people, regardless of whether they deserved it or not. Nobody seems to have figured out a quality metric that manages to capture "your quality, as expressed by the delta between the performance of these students under your tutelage vs. their hypothetical performance under other conditions" rather than a basic "what grades did your students get?"(the latter, perversely, makes people who provide honest feedback about bad performance look like bad teachers, while rewarding those who provide dishonest feedback about bad performance. Clearly an excellent metric.)
The less polite; but not substantially less accurate, version of TFA's headline might be "Qualcomm Maker attempts to pass off responsibility for kernel-mode portion of 3D Snapdragon driver"...
Later to be replayed in the "overheating time capsule" debacle.
While it would certainly be nice if scientists, as a class, were better at public communication, I think that this consideration misses an important point:
If somebody happens to be the best available information source on a given issue, failure to communicate with them is a major failing on your part.
All men may be created equal; but only some of them are worth consulting for advice.
I still want to know what Intel has up their sleeve with light peak.
10Gb/s over fiber isn't new or all that interesting; but, in the networking world, it isn't all that cheap. What have they done to make the equivalent of shoving a 10GbE fiber interface into random bits of cheap consumer electronics remotely viable?
Second, while neither optical cables nor optical connectors are quite in "will die if you give them a funny look" territory, they definitely won't stand up to the kind of abuse that even ghastly quality copper will, never mind being rolled over by chairs and filled with pocket fuzz. What do they have in mind?
Yeah, you'll need a converter, then. The one I linked may or may not be your best shot; but it should be representative of the sort of thing you want.
It varies. The rated wattage(which is sometimes even not a lie) of "enthusiast" power supplies has been climbing steadily, as the ability to purchase and cool seriously toasty chips, often several per system, has become substantially cheaper and more widespread. Though, it should always be noted that "rated wattage" means "power it can deliver without Something Bad happening" not "Power it actually uses". The rated wattage of high-density servers has increased substantially for similar reasons, and those wattage ratings usually aren't lies.
On the other hand, performance-per-watt has been improving markedly over the years, and the efficiency of non-crap PSUs has made decent strides. Also, the ability of modern chips(CPUs and GPUs particularly, though often smaller stuff as well) to intelligently throttle themselves when unused or lightly used has improved pretty markedly.
Thus, it is basically impossible to say whether "old computers use more power than new ones". It's just too general a statement to be meaningful. On the one hand, nobody on the high end blinks at a 150watt TDP processor, where(back in the day), hitting 60 watts on your crazy overclock was considered seriously hardcore. On the other hand, an ever increasing proportion of computers in actual use are laptops which cram the entire computer, and a monitor, into a 60watt or less AC adapter, and conserve power on idle like their (battery) lives depend on it.
In general, the ability of people, who don't have $250,000 to spare and a datacenter to house the beast, to buy very energy-hungry computers has increased significantly over time(1.5 kilowatt PSUs definitely didn't use to be retail items). On the other hand, though, performance per watt has gone through the roof(and CRTs have largely gone out the window). Whether ditching an old piece of gear for a new one is usually a case-by-case thing. If your mid-90's junker can be replaced by a weedy little plastic ARM box that runs off a wall wart that weighs under a hundred grams, the replacement almost certainly saves energy. Replacing a late-model PIII and a 17-inch CRT with a new Quad-core and a 28-inch LCD probably won't(though you will get surprisingly close to breaking even, and the performance will be a lot better).
This is why the energy economics of using obsolete x86s as networking gear are usually pretty weak, while those of getting rid of generic business desktop c. 1999 just to replace it with generic business desktop c. 2009 probably won't be nearly as exciting(unless the ACPI on the old box was painfully broken, as it not infrequently was)...
This would probably work. It has the look of an industrial/specialty system, though, so it could easily double the cost of the project.
Also worthy of consideration, a fairly wide variety of mini-ITX and smaller motherboards, particularly those designed for industrial/kiosk/signage applications actually include LVDS headers right on the board. This is of no use to you if you absolutely need a standard desktop board, or a spiffy graphics card; but if your option is either "Spend 80-150 for a motherboard. Spend 200-400 for a specialty DVI to LVDS board because you are buying quantity 1, have it dangle awkwardly." or "Spend 200-400 for a motherboard with LDVS natively." the second starts to look pretty good...
If this sounds attractive, Logicsupply.com has a good range. They may or may not have the best price on a given board; but they have an excellent collection all in one place.
There is no "mode control" as such. Basically, the way it works is that if the backlight is brighter than the ambient light falling on the screen, it is in the transmissive "mode"(which gives you lower resolution; but full color. If the ambient light is strong enough to wash out the backlight, or the backlight is off, you are in the "reflective" mode, which is much sharper; but B/W. At least in the version used in the OLPC, there are some intermediate states, where you get enough backlight for a hint of color; but the ambient light is strong enough that it looks sort of like a B/W image with pastel shading, rather than a full color image.
On any existing laptop, you "control the mode" with whatever mechanism you would ordinarily control the display backlight with. Turn the backlight down low enough, and you are "in reflective mode". Crank it up and you are "In transmissive mode"(unless you are in direct sunlight, in which you probably don't have a choice.
Any device designed with this display in mind would, in all likelyhood, give you slightly more "audio-like" backlight control(i.e. There would be the usual intensity control; but there would also be a discrete on/off toggle, so you could set your preferred indoor level, then hit just one button and save a whole lot of power while you are outside, then give the same button a tap to go back to your preferred level when you go back indoors). In any case, though, there is no magic "mode bit" that needs to be set, or special command, or pin that has to be toggled or anything. Purely a "however you would normally control the backlight" thing. The only real design difference is that, unlike a normal LCD, this thing can actually be useful with the backlight totally off, so you will probably want to incorporate a somewhat wider than usual range of brightnesses and/or an audio style "mute" button...
The Pixel Qi screens are the evolutionary descendants of the OLCP ones, and might well have not traded off cost and quality quite as agressively; but I'm assuming that at least half of this price is the "It's a quantity-one sale of a previously unavailable item to a cost-insensitive enthusiast" premium.
The ruthless margin-slashers who do purchasing for the big OEMs are just going to give you a thousand-yard-stare and a hollow laugh if your quoted price is much above a standard LCD of the same size, so Pixel Qi are either utter morons, or offer much more reasonable terms in quantities of 10,000+
Plus, while the maker shed is a noble operation, and sometimes a useful place to get stuff that would be hard to find in small quantities elsewhere, they aren't what you would call an "everyday low prices" kind of operation...
It's such a pity that they never settled on a standardized physical connector for the LVDS+backlight power connection that virtually all laptop screens use(are there any internal displayport devices in the wild yet?) Electrically, they are usually much the same, at least within a given size class(obviously, the current required for a 17inch DTR LCD backlight is going to be a little bit higher than that needed for a 8 inch netbook LCD, so a diffferent connector might be needed); but there was no real standardization. For basic economic reasons, and the fact that there are fewer OEMs than there are brand names, there are a lot of identical connectors lurking out there if you take a screwdriver to the problem; but there is nothing resembling a proper, consumer accessible, "standard", on the order of DVI or molex...
It would certainly have been even more superior-looking when compared to the sort of handsets that tracfone deals in; but they still would have had to solve the basic cost-of-data problem.
If the various Kin-social-integration-cloud-data stuff consumes smartphone levels of data, carriers are going to want to charge smartphone-level prices for it, prepaid or contract. Customers, sensibly enough, are going to want a smartphone if they are paying smartphone rates for connectivity.
Since the prepaid services tend to cater to a mixture of thrifty-people-who-don't-care-much in all income brackets, along with poor people and people with questionable credit ratings, the data costs would continue to be an issue. The handset competition would be less fierce, because the prepaid guys mostly get stuck with some real crap; but prepaid customers are probably, in aggregate, even less excited than verizon customers about paying stiff charges for a non-smartphone.
Or, in a view where not quite everything is moved by nefarious wheels-within-wheels or filled with secret Combine electronics, it could simply be that it made good business sense for Fark to move into the niche that the Drudge Report had no intention of even trying to fill. Competition is painful and can be dangerous. You frequently see such attempts to segment a market.
Which actually makes having killed it seem like a pretty decent idea.
Well golly, should I choose this here net-book that does absolutely nothing unless my Palm brand smartphone is within a couple of meters, or should I choose this one that is pretty much like a normal computer, only smaller and cheaper and slower, that I can tether my phone to if I really want to....
Had Foleo been some sort of WebOS device that could either function alone like a normal netbook, syncing data and favorites and stuff back to your phone the next time they were in contact and, if the two were directly communicating with each other, the user being able to "pass" running application cards back and forth between the two, that would have kicked ass. As it was, though, Foleo was sort of the last gasp of old-school palm, offering an expensive and inflexible way to give your palm phone a larger screen and keyboard. They might as well have tied it to the railroad tracks as released it in front of the netbook explosion.
Perhaps we can conclude that the CEO of Microsoft then and the lead on Bob had a better "working relationship" than do the CEO of Microsoft now and the lead on Kin?