My suspicion would be that the level of detail that is commercially viable will depend largely on the availability of tools to generate it more efficiently(and recycle without being too blatant what you've already generated)...
Something like a high-resolution 3D laser scan and motion capture isn't cheap; but if you have the capability to take a library of captured actors and then programmatically mix-and-match and slightly randomize certain parameters to generate unlimited NPCs, the start-up cost is high; but the incremental cost of adding additional mooks becomes relatively cheap.
Same thing would go for programmatically generated trees, grabbing textures with (relatively) cheap high-resolution digital cameras, and so on.
According to TFA, the publishers went 'agency' in order to try to stop Amazon from 'cornering the market' by selling books cheaply. Now they are under DoJ fire for what was essentially an attempt to set an artificial price floor across the industry.
Squeeze. Crunch.
Y'know what might have been a better plan? Not Insisting on the DRM that makes it possible, and easy, for an incumbent seller to lock in large numbers of buyers and obtain the market power needed to then put the publishers on the rack... It's not as though the story of iTunes went exactly that way with team RIAA or anything...
If DRM actually magically worked, there might be some business case for accepting a smaller slice of an impregnable walled garden; but the present state of it is trivially weak for all the common book formats. Good work on stopping no pirates and giving large retailers the power to cut your throat, guys...
This is also what happens when you talk about whatever fuckup occurred on the dark end of your super-fancy-don't-look-too-closely-contemporary-JIT-outsourced-supply-chain...
Consumer electronics widgets are constantly having their ship dates quietly revised, usually with a terse announcement from some PR flack that 'Release of Widget Foo has been moved from late Q1 to mid Q2'. Or they just go up for pre-order and take longer than initially promised to ship. Annoying; but hardly unusual.
In this specific case, I'm a bit surprised that they didn't have somebody plug one into their laptop and then wonder why the NIC wasn't working slightly earlier in the process; but so it goes.
If you are using a persistent/tmp, 'root' is anybody who mounts the HDD...
Now, if you want scrollback data to be persistent across reboots, you have to suck it up and dump it to disk somewhere(user home might be a slightly better idea, since support for encrypting your home directory is slightly easier than for encrypting the entire disk); but if that isn't a requirement you probably don't want it touching the disk at all(unless you are in a very memory constrained and swapless environment where getting OOM killed every time something unexpectedly verbose happens would be really annoying).
Given the little fuss over Apple's quiet-but-massive location tracking of cellular iDevices some time back, it would not be a total surprise to learn that they are planning on using their comparatively pervasive handsets in order to produce their own traffic data... If you have a street map, and you know how fast and in what direction the iDeviced commuters are moving, you may be able to draw useful inferences from there.
Bah. To have the honor of being a part of a glorious iProduct is all the recognition any puny NIH product could possibly desire, or conceivably deserve...
Still reasonably common. I don't have exact figures or anything; but you'll still quite often see those distinctively chunky, rectangular passives placed in neat rows just slightly behind the ethernet jacks on devices where space constraints aren't a huge issue. They essentially double the board footprint of the jack, so I assume that they've been stamped out in laptops, classier switches where density counts, and the like; but cheapie switches, home router boxes, NICs of indifferent quality, and so forth are still using them. Sometimes even pin-through-hole DIPs, no less, not even surface mount...
They do tend to have all the magnetics for at least one port, often two or four, crammed into one package, I don't think I've ever encountered discrete ones on anything remotely recent; but magnetics separate from jacks are still around.
There are certainly excellent reasons to never touch Zuckerburg's kool-aid; but that isn't really the core of the problem here:
Facebook is one stop shopping for the petty snoop; but the problem(in this context, there are other contexts, with their own problems) is the number of petty snoops who, de facto, have enough power over you to force you to use your own credentials to defeat whatever trivial privacy barriers get in their way. Facebook makes it dangerously simple; but the fact that HR flacks or educational admin types have, and shamelessly exercise, the ability to demand access is a more fundamentally problematic thing...
Not to worry, citizen, our house counsel is on call during HR's operating hours in order to provide a nebulous-but-entirely-legal justification for any hiring and firing decisions we may wish to make.
The moral of the story(as always) would appear to be that purely rules-based protections(even when they aren't fundamentally flawed by design, as facebook's certainly are) are essentially useless in the face of a real power imbalance.
Facebook is a bit novel in that it produces such a very juicy target for lifestyle police, and one that is fairly persistent; but it isn't as though there is any conceivable privacy policy/enforcement mechanism that could protect you from somebody who has the real world power to make you defeat it for them.
A video in would actually be much more broadly useful(surely the IT minions of the world who occasionally have to deal with a headless box would pay a premium for a laptop that could use its own monitor, keyboard, and mouse as a KVM with just a single cable connection?); but I've never seen one, not even a vaporware or overpriced specialty one...
More generally, though, driving the internal monitor is hardly an impossible problem(either through feeding a video output, or agreeing on some standard way of grabbing the internal GPU's framebuffer), it's just one that doesn't presently have a general solution, so there is no way for a peripheral maker, or makers, to sell something to the owners-of-various-laptops in quantities enough to get an attractive price.
As expensive niche products, mostly aimed at specialist professionals who Just Can't Not Have their oddball expansion cards with them, you can purchase expresscard-connected PCIe enclosures that will drive external monitors if a graphics card is added. You can also get very specific, rather expensive, laptops(the Sony Vaio Z with its curious 'Power Media Dock', for instance) that do offer an external GPU module that more-or-less seamlessly grabs the internal display and takes over graphical duties.
What they didn't mention, for competitive reasons, are the parts of their datacenter now running MySQL on 23rd century technology, helpfully provided through some sort of plot hopefully less lame than that one with time-travel and whale extinction...
Their GPU compute division might also have a hand in it: Linux users as customers for desktop and gaming cards aren't wildly compelling; though certainly not nonexistent; but the people buying racks and racks of Tesla enclosures are an entirely different matter. Nvidia has no obvious interest in more OSS in their bits of that particular arrangement; but they certainly want it to work smoothly.
Better linux support for the notoriously eccentric ARM SoCs of the world certainly isn't a bad thing; but it does seem likely that Nvidia's interests align with Linux's interests in roughly the same way that IBM's do:
They are entirely happy to see a cheap and capable OS available to sit between their expensive hardware and their proprietary software; but they aren't exactly thinking of changing the status of either of those two...
There have been a few stabs at it, I think both ATI and Nvidia have released more-or-less-orphaned-on-launch partnerships with some laptop outfit or other, using proprietary cabling.
My understanding is that there are a few major hurdles:
Historically, there really haven't been any good standardized high-bandwidth interfaces to the outside world on laptops. The proprietary docking station port, if provided, might connect directly to the PCI bus; but your next best bets were relatively lousy things like PCMCIA or USB. Even with PCIe, you get 1x from an expresscard slot; but the standards for external cabling for anything beefier than that have been languishing in the PCIe SIG forever...
Unless you are content to use an external monitor only, an 'expansion GPU' both has to have access to all the usual bandwidth that a GPU requires and have enough bandwidth(and suitable software/firmware cooperation) to dump its framebuffer back to whatever internal GPU is driving the laptop screen. You can get(albeit at unattractive prices) enclosures that connect to the 1xPCIe lane in an expresscard slot and break that out into a mechanically 16xPCIe card enclosure with supplemental power. Assuming the BIOS isn't a clusterfuck, you can pop in an expansion card just as you would on a desktop. That only gets you the video outs, though, it doesn't solve the trickier and more system-specific problem of driving the laptop screen.
Docking stations: At present, laptop manufacturers get to designate one line as 'enterprise' by including the necessary connector, and then charge a stiff fee for the proprietary docking station as your only option to drive a few extra heads. I imagine that this blunts the enthusiasm of the major enterprise laptop players for a well-standardized and high bandwidth external connector.
The great thing about whatever card is currently going for $50 nominal, $20-$30 if the rebate gods smile upon you, is the video outputs:
Back in the bad old days, buying a bottom-barrel graphics card meant getting a single VGA out(and possibly one where the manufacturer had cheaped out so hard that analog quality issues were visible...), and lousy performance, and the PCI ones that you needed to run more than one, once your AGP slot filled, were always mysteriously overpriced(alas, this still seems to be the case with 16x vs. 1x PCIe versions...)
Now, even a cheap POS, low-profile, fanless, last-gen whatever card has at least a DVI and a VGA connector, possibly even an HDMI if you splurge an extra $10 for the lousiest 'media PC' part currently on sale. Plus, computers with enough PCIe slots to run more than one are downright common, so you can get some serious pixel area for dirt cheap...
Arguably, any business whose strategy to prevent staff gaming involves the GMA950, rather than software administration or hiring responsible people deserves the huge number of games of Farmville and Snood currently being played on their network...
The former is an intel Atom, running in the GTL mode that allows it to pair with core logic designed for P4s, using an Nvidia chipset instead of the bottom-barrel GMA-950-hobbled Intel chipsets that were the cheap(but not especially power-efficient or high-performance) thing to pair with Atom parts.
With the later Atom revisions, Intel moved to a new chipset interface that they assert Nvidia has to right to interface with(they moved to a different, also Nvidia-disallowed, interface for desktop and server/workstation parts). Here, with 'ION2', it's an intel CPU and chipset, combined with Nvidia's lowest-end discrete GPU.
Not much Nvidia can do about it; but ION actually became a suckier value proposition in its second revision. The first was worse only in that it was slightly more expensive. The second cost more and required more board space and power(because it wasn't replacing an inferior part with a better one; but adding an additional GPU, and possibly some RAM, since access to system RAM was pitifully slow for the outboard GPU).
The main reason that integrated GPU performance matters(aside from the fact that it is all the GPU you get in any too-cheap or too-skinny device that doesn't have a discrete option) is that it defines the (overwhelmingly common) baseline for what 'PC graphics' means. If that situation is uniformly awful, GPU intensive stuff will continue to be fairly niche, which leads to a chicken-and-egg issue: if integrated graphics suck, the market for GPU intensive stuff will be constrained, which will reduce the incentive to improve GPU performance, and so it goes...
Slashdot Mirror
What else will enable rich multimedia experiences on IE6 Legacy Essentials for Positronic Brainstem Implants Enterprise Edition?
My suspicion would be that the level of detail that is commercially viable will depend largely on the availability of tools to generate it more efficiently(and recycle without being too blatant what you've already generated)...
Something like a high-resolution 3D laser scan and motion capture isn't cheap; but if you have the capability to take a library of captured actors and then programmatically mix-and-match and slightly randomize certain parameters to generate unlimited NPCs, the start-up cost is high; but the incremental cost of adding additional mooks becomes relatively cheap.
Same thing would go for programmatically generated trees, grabbing textures with (relatively) cheap high-resolution digital cameras, and so on.
A new entrant into the solid-state heat pumps field would be rather nice... I wonder if they can get this thing to scale up a bit?
According to TFA, the publishers went 'agency' in order to try to stop Amazon from 'cornering the market' by selling books cheaply. Now they are under DoJ fire for what was essentially an attempt to set an artificial price floor across the industry.
Squeeze. Crunch.
Y'know what might have been a better plan? Not Insisting on the DRM that makes it possible, and easy, for an incumbent seller to lock in large numbers of buyers and obtain the market power needed to then put the publishers on the rack... It's not as though the story of iTunes went exactly that way with team RIAA or anything...
If DRM actually magically worked, there might be some business case for accepting a smaller slice of an impregnable walled garden; but the present state of it is trivially weak for all the common book formats. Good work on stopping no pirates and giving large retailers the power to cut your throat, guys...
You link to the PR statement issued a week after independent researchers caused a bit of a stir by releasing a tool to read back the tracking data from your iDevice.
"Transparency" is what you admit to before you are caught, not what you say afterwards...
Based on the amount of complaining about supply issues, apparently not enough.
This is also what happens when you talk about whatever fuckup occurred on the dark end of your super-fancy-don't-look-too-closely-contemporary-JIT-outsourced-supply-chain...
Consumer electronics widgets are constantly having their ship dates quietly revised, usually with a terse announcement from some PR flack that 'Release of Widget Foo has been moved from late Q1 to mid Q2'. Or they just go up for pre-order and take longer than initially promised to ship. Annoying; but hardly unusual.
In this specific case, I'm a bit surprised that they didn't have somebody plug one into their laptop and then wonder why the NIC wasn't working slightly earlier in the process; but so it goes.
If you are using a persistent /tmp, 'root' is anybody who mounts the HDD...
Now, if you want scrollback data to be persistent across reboots, you have to suck it up and dump it to disk somewhere(user home might be a slightly better idea, since support for encrypting your home directory is slightly easier than for encrypting the entire disk); but if that isn't a requirement you probably don't want it touching the disk at all(unless you are in a very memory constrained and swapless environment where getting OOM killed every time something unexpectedly verbose happens would be really annoying).
Given the little fuss over Apple's quiet-but-massive location tracking of cellular iDevices some time back, it would not be a total surprise to learn that they are planning on using their comparatively pervasive handsets in order to produce their own traffic data... If you have a street map, and you know how fast and in what direction the iDeviced commuters are moving, you may be able to draw useful inferences from there.
Bah. To have the honor of being a part of a glorious iProduct is all the recognition any puny NIH product could possibly desire, or conceivably deserve...
Still reasonably common. I don't have exact figures or anything; but you'll still quite often see those distinctively chunky, rectangular passives placed in neat rows just slightly behind the ethernet jacks on devices where space constraints aren't a huge issue. They essentially double the board footprint of the jack, so I assume that they've been stamped out in laptops, classier switches where density counts, and the like; but cheapie switches, home router boxes, NICs of indifferent quality, and so forth are still using them. Sometimes even pin-through-hole DIPs, no less, not even surface mount...
They do tend to have all the magnetics for at least one port, often two or four, crammed into one package, I don't think I've ever encountered discrete ones on anything remotely recent; but magnetics separate from jacks are still around.
There are certainly excellent reasons to never touch Zuckerburg's kool-aid; but that isn't really the core of the problem here:
Facebook is one stop shopping for the petty snoop; but the problem(in this context, there are other contexts, with their own problems) is the number of petty snoops who, de facto, have enough power over you to force you to use your own credentials to defeat whatever trivial privacy barriers get in their way. Facebook makes it dangerously simple; but the fact that HR flacks or educational admin types have, and shamelessly exercise, the ability to demand access is a more fundamentally problematic thing...
Not to worry, citizen, our house counsel is on call during HR's operating hours in order to provide a nebulous-but-entirely-legal justification for any hiring and firing decisions we may wish to make.
The moral of the story(as always) would appear to be that purely rules-based protections(even when they aren't fundamentally flawed by design, as facebook's certainly are) are essentially useless in the face of a real power imbalance.
Facebook is a bit novel in that it produces such a very juicy target for lifestyle police, and one that is fairly persistent; but it isn't as though there is any conceivable privacy policy/enforcement mechanism that could protect you from somebody who has the real world power to make you defeat it for them.
A video in would actually be much more broadly useful(surely the IT minions of the world who occasionally have to deal with a headless box would pay a premium for a laptop that could use its own monitor, keyboard, and mouse as a KVM with just a single cable connection?); but I've never seen one, not even a vaporware or overpriced specialty one...
More generally, though, driving the internal monitor is hardly an impossible problem(either through feeding a video output, or agreeing on some standard way of grabbing the internal GPU's framebuffer), it's just one that doesn't presently have a general solution, so there is no way for a peripheral maker, or makers, to sell something to the owners-of-various-laptops in quantities enough to get an attractive price.
As expensive niche products, mostly aimed at specialist professionals who Just Can't Not Have their oddball expansion cards with them, you can purchase expresscard-connected PCIe enclosures that will drive external monitors if a graphics card is added. You can also get very specific, rather expensive, laptops(the Sony Vaio Z with its curious 'Power Media Dock', for instance) that do offer an external GPU module that more-or-less seamlessly grabs the internal display and takes over graphical duties.
You just can't, at present at least, get both.
What they didn't mention, for competitive reasons, are the parts of their datacenter now running MySQL on 23rd century technology, helpfully provided through some sort of plot hopefully less lame than that one with time-travel and whale extinction...
Their GPU compute division might also have a hand in it: Linux users as customers for desktop and gaming cards aren't wildly compelling; though certainly not nonexistent; but the people buying racks and racks of Tesla enclosures are an entirely different matter. Nvidia has no obvious interest in more OSS in their bits of that particular arrangement; but they certainly want it to work smoothly.
Better linux support for the notoriously eccentric ARM SoCs of the world certainly isn't a bad thing; but it does seem likely that Nvidia's interests align with Linux's interests in roughly the same way that IBM's do:
They are entirely happy to see a cheap and capable OS available to sit between their expensive hardware and their proprietary software; but they aren't exactly thinking of changing the status of either of those two...
Well, scratch would appear to have support for a clock, and a NOR gate, so an x86 compatible scratch VM is only your sanity away...
Interesting... How many simultaneously? I'm not a huge interface snob, I just need lots and lots of them...
There have been a few stabs at it, I think both ATI and Nvidia have released more-or-less-orphaned-on-launch partnerships with some laptop outfit or other, using proprietary cabling.
My understanding is that there are a few major hurdles:
Historically, there really haven't been any good standardized high-bandwidth interfaces to the outside world on laptops. The proprietary docking station port, if provided, might connect directly to the PCI bus; but your next best bets were relatively lousy things like PCMCIA or USB. Even with PCIe, you get 1x from an expresscard slot; but the standards for external cabling for anything beefier than that have been languishing in the PCIe SIG forever...
Unless you are content to use an external monitor only, an 'expansion GPU' both has to have access to all the usual bandwidth that a GPU requires and have enough bandwidth(and suitable software/firmware cooperation) to dump its framebuffer back to whatever internal GPU is driving the laptop screen. You can get(albeit at unattractive prices) enclosures that connect to the 1xPCIe lane in an expresscard slot and break that out into a mechanically 16xPCIe card enclosure with supplemental power. Assuming the BIOS isn't a clusterfuck, you can pop in an expansion card just as you would on a desktop. That only gets you the video outs, though, it doesn't solve the trickier and more system-specific problem of driving the laptop screen.
Docking stations: At present, laptop manufacturers get to designate one line as 'enterprise' by including the necessary connector, and then charge a stiff fee for the proprietary docking station as your only option to drive a few extra heads. I imagine that this blunts the enthusiasm of the major enterprise laptop players for a well-standardized and high bandwidth external connector.
The great thing about whatever card is currently going for $50 nominal, $20-$30 if the rebate gods smile upon you, is the video outputs:
Back in the bad old days, buying a bottom-barrel graphics card meant getting a single VGA out(and possibly one where the manufacturer had cheaped out so hard that analog quality issues were visible...), and lousy performance, and the PCI ones that you needed to run more than one, once your AGP slot filled, were always mysteriously overpriced(alas, this still seems to be the case with 16x vs. 1x PCIe versions...)
Now, even a cheap POS, low-profile, fanless, last-gen whatever card has at least a DVI and a VGA connector, possibly even an HDMI if you splurge an extra $10 for the lousiest 'media PC' part currently on sale. Plus, computers with enough PCIe slots to run more than one are downright common, so you can get some serious pixel area for dirt cheap...
Arguably, any business whose strategy to prevent staff gaming involves the GMA950, rather than software administration or hiring responsible people deserves the huge number of games of Farmville and Snood currently being played on their network...
There is 'ION' and 'ION2':
The former is an intel Atom, running in the GTL mode that allows it to pair with core logic designed for P4s, using an Nvidia chipset instead of the bottom-barrel GMA-950-hobbled Intel chipsets that were the cheap(but not especially power-efficient or high-performance) thing to pair with Atom parts.
With the later Atom revisions, Intel moved to a new chipset interface that they assert Nvidia has to right to interface with(they moved to a different, also Nvidia-disallowed, interface for desktop and server/workstation parts). Here, with 'ION2', it's an intel CPU and chipset, combined with Nvidia's lowest-end discrete GPU.
Not much Nvidia can do about it; but ION actually became a suckier value proposition in its second revision. The first was worse only in that it was slightly more expensive. The second cost more and required more board space and power(because it wasn't replacing an inferior part with a better one; but adding an additional GPU, and possibly some RAM, since access to system RAM was pitifully slow for the outboard GPU).
The main reason that integrated GPU performance matters(aside from the fact that it is all the GPU you get in any too-cheap or too-skinny device that doesn't have a discrete option) is that it defines the (overwhelmingly common) baseline for what 'PC graphics' means. If that situation is uniformly awful, GPU intensive stuff will continue to be fairly niche, which leads to a chicken-and-egg issue: if integrated graphics suck, the market for GPU intensive stuff will be constrained, which will reduce the incentive to improve GPU performance, and so it goes...