I do believe Samsung has an ARM Architecture license, too. Anyone selling a 1GHz or above Cortex A8 has access to a modified core, because ARM only certifies this to 800MHz (in a small enough process) based on their design. The "no mess" license doesn't permit overclocking.
One reason Apple does customization is the obvious one -- they save money. In particular, they have RAM on the same "A4" module (it's a "package on package" module).
This means they're getting all their RAM included, so no worries about supply, much less board space, no DDR layout issues, etc. Also saves a little board space. And since Samsung makes the module (the same Cortex A8 design, as well as a higher spec PowerVR, is found in the Galaxy phones and pads), there's little change of any DRAM supply issues, Samsung being a giant in DRAM.
Of course, Apple has to be a little nervous about Samsung, now that they're also in full competition. As the chip provider, they'll know precisely what's going in to Apple's next product and, as with the Galaxy series, ensure they outperform Apple. So I kind of believe that Apple may be looking at alternate vendors. They have had their own IC design team for awhile, and last year they bought Intrinsity, Inc., the company that worked with Samsung on tweaking the A8 design up to 1GHz.
There are quite a few companies with the ARM Architecture license. It was needed by anyone delivering an A8 at 1GHz, since that's beyond the point of ARM's certification. Apple has an Architecture license, though they probably didn't need it yet, since all their stuff so far is just stripped down versions of Samsung SOCs. The shopping for a foundry in Taiwan might be true, though... with Samsung emerging as a big player in both smartphone and tablet, Apple might be getting a little nervous about their supply. Or even the simple fact that every iPhone, iPod, or iPad sold is also boosting Samung's economies of scale. Not that the world's second largest chip maker need worry all that much, anyway.
Most of the media decoding on these devices isn't done on the CPU, or even the DSP or GPU. To really keep power down, there are dedicated DCT acceleration engines in hardware. The Tegra 2, for example, can play 1080p using only about 400mW. The playback engine isn't totally dedicated to one specific video type... nVidia claims support for H.264, MPEG1/2, VP6, VP8, VC-1, and other DCT-based standards.
When the SGX543 was announced, PowerVR were actually showing it off in a four-core version. Otherwise, it's basically an SGX540.. maybe a little faster in most implementations, since the original design targeted 65nm chips, and no one's using that for this year's smartphones and tablets. This is a synthesizable core, as are all their "chips", so this can be implemented between 1 and 16 cores. The cores are computing cores, capable of GPGPU computing (OpenCL, etc). Desktop GPUs, of course, can have hundreds of such processors, but this is fairly new in chips designed for mobile computing. Of course, since this a PowerVR design, they're going to be in everyone's SOCs before you know it.
Apple's SOC names are only adding to the confusion... does an A8 SOC contain a Cortex A8 or Cortex A9 CPU? And how many? Most of the new tablets and even smart phones are going to A9 and dual core, at 1GHz or more. That's a good 2.5x faster than an iPad, at least at the metal (Android is going to be slower on some things per clock cycle, at least until is starts using the NEON vector instructions
Freedom remains a choice. And in its own weird way, Apple may have helped promote this. By being so repressive and so successful in the smart phone and tablet markets, they made openness a big selling point among the competition.
So sure, Apple is the dark side of freedom in computing, but I'm not sure the overall effects winds up a negative sum. Certainly, it was the prospect of close/proprietary search that mobilized Google to buy Android and open source it all (well, everything at the OS level). And probably at least in part the need to be an alternative to the iPhone that kept them from locking it into the Android Market.
And while Apple certainly didn't invent as much of the things in these new systems as the media likes to believe, they certainly did get it to market early, and make it popular. I'll never use an Apple product, but I can appreciate that some of the things I do use are better because they had to take on Apple's things.
Actually, in series you get a pretty strong voltage, but exactly the same current you get with a single cell. Put them in parallel to increase the current.
Actually, R is variable. As shipped, R is very, very high....to high to measure with a conventional ohmmeter. As the paint wears off, R is decreasing. It seems to vary according to how the whole assembly is held, too, just worn pain may not be enough to cause problems.
The ultimate current-limiting circuitry -- an open circuit -- is nearly free. That seems to be the intended purpose of the paint, and it works, as long as the paint stays in place.
I think the problem is that they're re-using metal parts from the Kindle2 or some other cover, where this didn't matter. So changing that in any significant way isn't free. In particular, going to plastic. The plastic itself is very cheap, but making a mold for it will runs in the tens of thousands of dollars. They certainly have the volume to justify that, but apparently, a little paint seemed like a simpler idea.
It's not a continuous metal bar... that's why finding this is kind of tricky.
There are two clips, and they slide against each other, with a spring tensioning the top clip to its resting position. I think you have see wear on the hidden metal parts (within the case), not just on the contact tabs, before any significant current will flow. And even this, it's only when things are held just-so. These cases are blamed for crashing the processor, not taking the whole Kindle down, exploding batteries, or other evil you'd expect if this were a constant short circuit.
Actually, its worse with a Lithium battery. Yes, a partially cycles Lithium cell seems to last much longer, just as the NiMh cells used in hybrid cars. iPhone owners experience this with their non-replaceable iPhone cells... if you run the iPhone down all the way each day, it won't last out your AT&T contract. If you charge it regularly, it does.
The big problem is that completely discharging a Lithium cell can damage it. One of the contacts typically acts as a parasitic reverse-polarity battery and consumes itself, below a certain level. Most Lithium-cell gear will shut off power entirely before this happens. No idea if the case contact on a Kindle run though power management, or just off the battery though a current limiter (resistor or some other device).
The main problem was simple: it was a "feature" phone, but required a full smart phone data plan... an extra $30 per month or whatever. That made it, in essence, in competition directly with iPhones and Androids.
Even High School Kids aren't that stupid... at least, the small percentage who's parents would shell out the cash for a smart phone (I have one kid in HS, one in College, neither gets a smart phone).
I think the plug-pulling came from the results: by most accounts, they sold some few thousands of phones in those six weeks. Awfully close to zero, and those few? Probably sold by accident.
It's Google's decision to make it the carriers' responsibility to deliver an up-to-date OS release, or simply ignore the old devices. In that, yeah, it's partly Google's blame.
The UI stuff... different user home shells, I claim that's no big deal. You don't have to use HTC Sense or MotoBlur; there are close to a dozen alternate shells in the Android Market. Google might help here by allowing users with other home shells to download the bog standard Android shell. But it's not as if these things take more than a few minutes to learn, anyway.
The big problem is building Android too much like Linux used to be built... they need a modular architecture. Right now, your ROM is a big amalgam of built-in apps, OS, drivers, etc. What the really need is a hardware abstraction layer. The OEMs write very simple drivers for cameras, I/O devices, etc -- anything you need, just a basic driver. The OS then recognizes these, wraps them in higher level Linux or Android drivers as needed.
The result: Neither phone OEMs nor carriers ever HAVE to touch Android proper. They can update HAL modules independently of the whole OS, etc. And Google could push out a whole OS update in the Android Market that works on every device, no need to worry about embedded drivers.
That was one big win on the IBM PC -- they wrote the BIOS, which is really just "HAL for Dummies". So no need to worry a great deal about which OS the machine is running, etc. Sure, in those days, you didn't really have full hardware abstraction much, but that's totally practical in a machine as capable as today's Android devices today.
Flaws in measurement aside, it's not a continuous wire. The clips are actually part of two metal rails, presumably painted, which slide against each other, with a spring in-between. You would probably get some real current flowing if the [presumed] paint wore off internally; wear on the clips is not enough to cause a problem.
Now, I just looked at one of these owned by a co-worker -- he wasn't going to let me take it apart. It's possible the metal gives way to plastic under the leather cover, but I have no evidence of that -- sure seems like metal all the way. Stripped of paint and held just so, I'm sure it's less than 2 ohms. But you're probably going to get trouble with significantly higher resistance than that. Ok, not 2 megaohms:-)
The big problem with MediaFlo, DVB-H, etc. is that they were essentially obsolete before they were really finished. In both cases, the idea was to use a single TV channel, 6-8MHz, to transmit up to 32 sub-channels for mobile devices (a single MPEG-2 TS stream, which can carry up to 32 sub-channels). This pretty much means you're viewing a tiny image on an old-fashioned cellphone. The maxium resolution for DVB-H is 320x240|288 or so; they were really planning for tiny cellphone screens. Not sure about resolution, but the rate per digital sub-channel on MediaFlo, for example, is 200-250kb/s... even YouTube aims higher than that for their lower resolution SD video.
In short, this is going to awful on a 2010-vintage smartphone... our smartphones are already typically 800x480 or better, full standard-definition, DVD-quality or thereabouts (a DVD is actually 720x480 NTSC, 720x576 PAL). And we already have IPTV of various sorts on these devices. A separate super-low-resolution, subscription-based broadcast TV option is not something anyone wants.
LTE supports 81.6Mb/s and 200 active data clients on a 5MHz channel, using 4x4 MIMO, 43.2Mb/s with 2x2 MIMO. Not quite as bad as you let on here. Obviously, if you had 200 data users on a cell, they're not all getting the high speed they're after... but no different than the 3G situation -- about 21Mb/s via HSPA or 56Mb/s with HSPA+ (2x2 MIMO and 64QAM).
AT&T has actually been buying up 700MHz spectrum for years now. Along with the national Block B 12MHz they bought at FCC Auction 73, they own up to 12MHz of Block C from some of these other purchases... I hadn't realized they bought Aloha Partners sometime back in 2007. So that's up to 24MHz in some areas, even before you factor in this new 6MHz block. LTE doesn't support more than 20MHz per channel, but in aggregate, AT&T may have up to 480Mb/s of LTE per cell. Not too shabby.
Verizon has been doing much the same thing... they spent $4.7 billion for the national 700MHz Block C (and a few licenses in Block A as well) they won in Auction 73, 22MHz wide. And another $4.66 billion buying up 700MHz spectrum owned by regional companies... no idea just how much, or where. But both companies are well situated for 4G, and clearly, the scarcity of this commodity is driving the price up. This has Verizon with up to 90MHz of aggregate spectrum in places (3G + 4G), over 45MHz through most of the country.
It'll be interesting to see if Echostar hangs on to their 6MHz Block E ($722 million), given they could better than double their money on it now. And the FCC still plans to run the Block D auction again, probably next year (Block D has to be shared with public service use).
Both UHF and VHF get attenuated, not entirely blocked, by objects: trees, houses, skyscrapers, etc... and both diffract around said objects. That's why your cellphone works indoors.
Higher frequency UHF is strongly attenuated by foliage, lower frequency UHF and VHF, no so much. As a digital radio designer, I did a 2.4GHz ISM band radio, narrowband, fairly long range, but it stopped dead at the edge of a forest. My next radio, at 435MHz (used to control robots) went completely through the forest, over a slight hill, and out to a roadway 1/2 mile away.
That's big reason so much money changed hands in the FCC's 700MHz auction. Verizon nabbed 22MHz there, AT&T 12MHz. So Flo TV's 6MHz, 716-722MHz, may not seem like much, but it boosts AT&Ts 4G band by 50%. They're closing down Flo TV in March, AT&T is expected to launch their LTE-based "4G" (not real 4G yet, but a first step) sometime over the summer. Unless, like everyone else has been, they're six months late.
There's no value in a 50 mile range for cellular.... no matter how powerful the tower, the handset is still going to be limited to about a watt, maybe less (US cellular devices can run up to 3W, but handsets usually max out a 1W or less... the cell tower actually tweaks the handset's output to something it can hear).
The 700MHz band is bandwidth limited compared to Sprint, for example.. the Sprint/Clear/Comcast WiMax network has about 90MHz in the 2500MHz band, so they have a higher peak capacity, for sure. But they're going to need more power per cell in cities, if not more cells, to get into buildings. And they're going to have issues with rural coverage -- as Sprint and T-Mobile already do for 2G and 3G. They're limited to 1900MHz, while AT&T and Verizon both have slots at both 850MHz and 1900MHz. My house is centered in 26 acres of forest. I can get Verizon in my cellar, AT&T though most of the house, but T-Mobile is pretty much outdoors only, while Sprint, last I checked, is available at the end of my driveway.
Of course, you can always deal with the limited bandwidth issue by using more, lower powered cells in highly populated areas. The only real fix for the higher frequency stuff outside of high population areas is more cells, something historically just never done by these guys,
There are multiple problems here, none likely to be fixed. The first problem is the UI... until Windows is as slick as iOS, WebOS, or Android (take your pick) on a finger based touch screen, the OS fails for tablets. Consumers at this level have already rejected stylus-based input, at least as the normal UI means.
Then you have the general unusability of much of Windows software. This is inherently a bottom-of-the-barrel device, performance-wise. To support it properly, a developer may have to give up on full support for higher-end PCs, or release multiple versions... negating the one supposed big advantage to Windows on a tablet: all that software. And where table performance is adequate, you have the same UI issue. The Windows API isn't robust enough to simply have the OS say "tablet", and every program suddenly work with a touch interface. So apps have to change too, again negating the Windows advantage.
And finally, the co-evolution of Windows and Intel's x86 agenda, until recently. Intel has been very successful at replacing add-on hardware with x86 software. They fell a bit on AVC decoding, but the GPUs have picked up that slack. And it's great... until you have to consider power. Which is the single most critical factor in building a successful tablet -- all day computing. If I have to worry that my battery's going to die over an 8-12 hour workday, no matter what I do on the tablet, that's a big fail. And certainly, many cheap ARM tablets fail here, too.. but usually by holding back on expensive components, like batteries, or more modern ARM SOCs with better power savings and alternate function units. For example, video: the better ARM SOCs don't use the CPU or GPU as the main video decoding engine, they have a separate unit for this, which dramatically reduces power consumption when displaying video. Windows systems have embraced the "space computing cycle" fact of the desktop to "let the CPU just do it", saving hardware money, but costing in power. This is not easily divorced from Windows, not even from Linux necessarily, as it's been the big evolutionary direction of the PC platform for 20+ years... ever since Intel noticed a faster wordprocessor wasn't going to drive PC upgrades anymore.
They don't cost anything significant in the base device. USB ports run about $0.50 each, for example. They do cost as an add-on... and you still can't get digital video out from an iPad. Why a device released in 2009 only offers CVBS or VGA out makes absolutely no sense.
But I do believe you have the answer here: Apple won't change. They're not going to offer additional ports, simply because they want customers to pay them for extra things. People who really need to hook their video or still cameras to an iPad will either live with the clumsy adapters, or more likely, buy a different device (that's one primary use of a tablet computer for me -- I would not even consider one that didn't have at least a card reader and one USB port as a built-in).
The earlier PC tablets were like the earlier compact laptops -- they cut out features and raise the price, sometimes dramatically. As a result, tablets and laptops with 10-12" screens were niche items.
The creation of the netbook changed this for small laptops. Sure, they were yet-even-more underpowered, though in an era in which most users don't need the power of the PC they've got, this is acceptable in a $200-$300 computer.
Apple took a different tack. They always sell overpriced laptops, and the reason they don't have an answer to the netbook, per se, is that selling a lower cost Macbook would destroy both their record-level margins on Mac hardware, and worse-yet, their ability to sell higher spec Macs at 2x-3x the same price as a PC.
Their solution, of course, is selling the inflated iPod that became the iPad. That's the same guts, larger screen and battery, as the iPod/iPhone... and it continues their high profit margins. Thing is, a $500 iPad is still only $500... toy money for the well-off. This is very different than the $2,000-$3,000 PC tablets that didn't compare, performance-wise, to a $500 generic PC notebook. The other ARM tablets suggest a similar level of acceptability.
And the PC tabs might too, in a vacuum. Building a tablet from Netbook DNA gets you a $200-$300 tablet, if not for 2011, then certainly for 2012. About all you're doing is taking stuff off the existing Netbook BOM... maybe add $20 for a touchscreen, but some Netbooks already had these, too. Another $20-$25 if you need a cellular LAN.
But there's no vacuum... what does the PC tablet bring that the ARM tablet doesn't? PC apps? Not so much... particularly for Windows. Sure, they'll technically run, but they don't work via a touch UI, they want far more resources than the tablet has available, etc. You're still selling worst-in-class performance. ARM tablets will all be running made-for-touch software evolved out of the smartphone world, and they'll all offer best-in-class performance. Sure, an Atom will compare in performance if not necessarily power savings to a dual core ARM Cortex A9 (the standard for all 2011 tablets), but does the x86 offer any advantage at that point? Is Intel building a whole system that can deliver the things tablet buyers will demand?
I very much doubt this, and here's why. Intel has spent the last 20+ years coming up with more work for the x86 to do. This has been very successful in keeping the PC processor centric, shedding nearly any bit of specialized software that came along (modem chips, audio synthesizers, etc) for another x86 device driver. But now there's the tablet -- it's expected to play HD video for 10+ hours straight, for example. To do most of what it does all day. Most of these are achieved by a basic ARM, small GPU, and a bunch of other dedicated units on an SOC that accelerate these specialty jobs. I have real doubts about Intel playing well here... they have the x86 doing too much work to compete against dedicated units. No matter how power light it gets, the dedicated stuff is going to run less power.
The bottom line will be the real value of the x86 tablet -- does it do anything consumers want, well enough to suggest that it's the correct answer? I don't believe we'll find out in 2011... Intel spent years getting the laptop x86 correct. They got hit occasionally, such as when Transmeta first shipped, but did eventually step up to clobber them. They got caught offguard by AMD's push into 64-bit x86 computing, but fixed that, too. Only this time, they're not playing in x86 land. And in fact, they're going up against the world's most popular 32-bit CPU, on a platform that probably won't benefit from Windows compatibility. This isn't going to be such an easy one.
Yeah... I figure, if I'm going going to be a human billboard, I either better like the product for the product's sake, or get paid for the ad placement. So I'll wear an R.E.M. or Springsteen T-Shirt, music I strongly recommend, or one from Martin Guitars or maybe Panasonic camcorders, since I use both of those products. But Abercrombie? A company largely distinguished by delivering the stupidest looking stores known to mankind, and hiring a couple of shirtless boys to hang out in font? That company... they'd have to pay BIG to place an ad on me.
It's a miracle of modern advertising kung-fu that some people will pay several times more for the same cheap Chinese-made t-shirts, if advertising is also included.
You're also ricking the future based on the status quo. Sure, it's free today. But what happens when the developer goes out of business.. can I still get that paid-for app? What happens if Google decides at some point that all those Chrome users would do wonders to their bottom line if they had to pay ($5? $10? $20?) for all these great web services they've been getting for free?
The personal computer revolution was a revolution because it took power away from corporate central administration and gave each individual control of their own computing power, data, etc. Smart phones have largely followed that model, but still ceeded some of that control, particularly in the case of the iPhone -- you've given away the "I can run any program" capability, even if that never personally matters. I like the web for transparent sync among networked devices, but that's a hard thing to make expensive, or proprietary.. data exchange is pretty easy.
But putting most everything online, you give up all real control. That may not matter now, but it could eventually.
Freedoms are rarely lost all at once. And usually, we get short-sighted and yield them, a tiny bit at a time, in return for some seemingly big deal: apps everywhere, a false sense of security, whatever. As a people, particularly in the USA, we seem less educated about these things than most of the world, despite our country's pivotal role in global democracy over the years.
Tech has been an enabler of freedom. PCs give us control of our own computing resources, so no agency (the government, China, Apple) gets to say what apps we can or cannot run. The internet gives us communications that, while not perfect, are very hard for any one entity to thwart. The internet, being the agent of the free flow of information, helped to take down the Soviet Union and create modern Eastern Europe. Together, computers and search give us the first effective "intelligence amplifier"... we can consult thousands of sources of information in the time people from 20 years ago could manage 2-4 in a library.
But the prospect of centralizing ANYTHING is the prospect of controlling that thing. As benevolent at it begins, it's ultimately at the whim of the aggergator.
Actually, there are several. Data stored outside of your control means that, basically, it's outside of your control. If I'm paranoid, I can unplug any PC from the net... can't with this.
Where is it stored? The value of cracking commonly used PCs means that you get all sorts of attacks against Windows; less against less used OSs (MacOS) and/or those with higher levels of expertise and security (Linux).
But if this is all on a big Google server somewhere, there's now one single point of failure for all of that data. That's going to be one big-ass target, if it's at all successful.
Then there are other issues... like apps. Where are all the apps stored? I gather they run, at least in part, on the device, but are they permanent there, or just cached. Does the user have any choice about upgrades, or do they just happen, even if they break things? Are these stored at Google, or are they on the developer's computers? What happens when the server is down and I want to run my app? What happens when Company X goes out of business, but I've paid for that application -- does it still show up forever when I "sync to the cloud", or am I SOL.
There are some interesting things with the ChromeOS. One is simple: I'm sure I can run ChromeOS apps on a PC, and probably at some point on a smartphone or tablet, as well as on a ChromeOS device. This isn't for everything, but there may be a small class of apps that are really better when fully integrated with the net ("cloud", if you must). I was kind of skeptical, but I use a number of smartphone apps that are certainly enhanced based on cloud sync. While I still have POP3 mail and client on my PC, there are some real advantages to Google's GMail (which I also use).. they're the first ones really doing this right.
And certainly, if they're successful, there will be other folks hitting your data. For example, banks. Right now, I open a channel if I want to sync my financial software to my bank... it's on my terms, and if I'm really concerned about what it's doing, I can always fire up WireShark or something and see for myself. But put a Quicken-like app on ChromeOS, and the banks aren't going to worry about synching to my ChromeOS netbook; they're going to want to sync to my account. How do I manage or even know about what they can see and what they can't... and the other 2000 companies wanting the same access?
The PTO only really searches though the patent database for prior art. This is why all patents ought to be on public display for a year (as in Europe) before being granted.
At least, the 1999 filing date is correct, this will expire in 2019, if it's not struck down before then.
No.. I think Lucas is just doing that for experimental purposes. His real motive... see, he's running out of reasons to re-edit Star Wars. So there's this project, apparently, to produce a fake 3D version, which will no doubt be as unwatchable as all other fake 3D films.
But once the technology is there, he'll be able to resurrect young Harrison Ford, young Carrie Fisher, etc. and remake Star Wars entirely from scratch, in the computer. He'll keep all the old dialog, but none of the original film at all -- all digital.
Laugh now.. laugh while you can, monkey-boy. But after seeing young Jeff Bridges brought back in the new Tron, this is clearly where Lucas wants to go.
And probably every other studio, too. Right now, actors are fairly important, and name-brand actors in particular. But once you can license an actor's likeness, living or not, then you just need a warm body (for a little while) on which to build that digital actor, and some good but non-name voice-over talent. And I'm sure they're working on both of those problems.
The "good" part in advancing the state of the iPhone.. new models, that make you WANT a new one (you, not me, I have no use for an iPhone). If you somehow NEED that upgrade, because Jobs and Co. dangle a shiny new carrot in front of you, that's no way you being forced.
On the other hand, if they're obsoleting the older models via software upgrades, you may well be forced to upgrade, or stop having access to available software. They just did that, after all... iOS 4 barely runs on the iPhone 3, doesn't on earlier iPhone and iPod models. AND they pushed developers to upgrade software for this and the iPad.. so the set of available apps that run on older models is drying up. That's how they force an upgrade.
I replaced the main board, CPU, and RAM on my primary system two weeks ago.. ran me about $500. The whole PC is far more expensive than a smart phone: monitors, hard drives, printers, scanners, keyboard, mouse, etc.
Slashdot Mirror
I do believe Samsung has an ARM Architecture license, too. Anyone selling a 1GHz or above Cortex A8 has access to a modified core, because ARM only certifies this to 800MHz (in a small enough process) based on their design. The "no mess" license doesn't permit overclocking.
One reason Apple does customization is the obvious one -- they save money. In particular, they have RAM on the same "A4" module (it's a "package on package" module).
This means they're getting all their RAM included, so no worries about supply, much less board space, no DDR layout issues, etc. Also saves a little board space. And since Samsung makes the module (the same Cortex A8 design, as well as a higher spec PowerVR, is found in the Galaxy phones and pads), there's little change of any DRAM supply issues, Samsung being a giant in DRAM.
Of course, Apple has to be a little nervous about Samsung, now that they're also in full competition. As the chip provider, they'll know precisely what's going in to Apple's next product and, as with the Galaxy series, ensure they outperform Apple. So I kind of believe that Apple may be looking at alternate vendors. They have had their own IC design team for awhile, and last year they bought Intrinsity, Inc., the company that worked with Samsung on tweaking the A8 design up to 1GHz.
There are quite a few companies with the ARM Architecture license. It was needed by anyone delivering an A8 at 1GHz, since that's beyond the point of ARM's certification. Apple has an Architecture license, though they probably didn't need it yet, since all their stuff so far is just stripped down versions of Samsung SOCs. The shopping for a foundry in Taiwan might be true, though... with Samsung emerging as a big player in both smartphone and tablet, Apple might be getting a little nervous about their supply. Or even the simple fact that every iPhone, iPod, or iPad sold is also boosting Samung's economies of scale. Not that the world's second largest chip maker need worry all that much, anyway.
Most of the media decoding on these devices isn't done on the CPU, or even the DSP or GPU. To really keep power down, there are dedicated DCT acceleration engines in hardware. The Tegra 2, for example, can play 1080p using only about 400mW. The playback engine isn't totally dedicated to one specific video type... nVidia claims support for H.264, MPEG1/2, VP6, VP8, VC-1, and other DCT-based standards.
When the SGX543 was announced, PowerVR were actually showing it off in a four-core version. Otherwise, it's basically an SGX540.. maybe a little faster in most implementations, since the original design targeted 65nm chips, and no one's using that for this year's smartphones and tablets. This is a synthesizable core, as are all their "chips", so this can be implemented between 1 and 16 cores. The cores are computing cores, capable of GPGPU computing (OpenCL, etc). Desktop GPUs, of course, can have hundreds of such processors, but this is fairly new in chips designed for mobile computing. Of course, since this a PowerVR design, they're going to be in everyone's SOCs before you know it.
Apple's SOC names are only adding to the confusion... does an A8 SOC contain a Cortex A8 or Cortex A9 CPU? And how many? Most of the new tablets and even smart phones are going to A9 and dual core, at 1GHz or more. That's a good 2.5x faster than an iPad, at least at the metal (Android is going to be slower on some things per clock cycle, at least until is starts using the NEON vector instructions
Freedom remains a choice. And in its own weird way, Apple may have helped promote this. By being so repressive and so successful in the smart phone and tablet markets, they made openness a big selling point among the competition.
So sure, Apple is the dark side of freedom in computing, but I'm not sure the overall effects winds up a negative sum. Certainly, it was the prospect of close/proprietary search that mobilized Google to buy Android and open source it all (well, everything at the OS level). And probably at least in part the need to be an alternative to the iPhone that kept them from locking it into the Android Market.
And while Apple certainly didn't invent as much of the things in these new systems as the media likes to believe, they certainly did get it to market early, and make it popular. I'll never use an Apple product, but I can appreciate that some of the things I do use are better because they had to take on Apple's things.
Actually, in series you get a pretty strong voltage, but exactly the same current you get with a single cell. Put them in parallel to increase the current.
Actually, R is variable. As shipped, R is very, very high....to high to measure with a conventional ohmmeter. As the paint wears off, R is decreasing. It seems to vary according to how the whole assembly is held, too, just worn pain may not be enough to cause problems.
The ultimate current-limiting circuitry -- an open circuit -- is nearly free. That seems to be the intended purpose of the paint, and it works, as long as the paint stays in place.
I think the problem is that they're re-using metal parts from the Kindle2 or some other cover, where this didn't matter. So changing that in any significant way isn't free. In particular, going to plastic. The plastic itself is very cheap, but making a mold for it will runs in the tens of thousands of dollars. They certainly have the volume to justify that, but apparently, a little paint seemed like a simpler idea.
It's not a continuous metal bar... that's why finding this is kind of tricky.
There are two clips, and they slide against each other, with a spring tensioning the top clip to its resting position. I think you have see wear on the hidden metal parts (within the case), not just on the contact tabs, before any significant current will flow. And even this, it's only when things are held just-so. These cases are blamed for crashing the processor, not taking the whole Kindle down, exploding batteries, or other evil you'd expect if this were a constant short circuit.
Actually, its worse with a Lithium battery. Yes, a partially cycles Lithium cell seems to last much longer, just as the NiMh cells used in hybrid cars. iPhone owners experience this with their non-replaceable iPhone cells... if you run the iPhone down all the way each day, it won't last out your AT&T contract. If you charge it regularly, it does.
The big problem is that completely discharging a Lithium cell can damage it. One of the contacts typically acts as a parasitic reverse-polarity battery and consumes itself, below a certain level. Most Lithium-cell gear will shut off power entirely before this happens. No idea if the case contact on a Kindle run though power management, or just off the battery though a current limiter (resistor or some other device).
No, it was doomed to failure.
The main problem was simple: it was a "feature" phone, but required a full smart phone data plan... an extra $30 per month or whatever. That made it, in essence, in competition directly with iPhones and Androids.
Even High School Kids aren't that stupid... at least, the small percentage who's parents would shell out the cash for a smart phone (I have one kid in HS, one in College, neither gets a smart phone).
I think the plug-pulling came from the results: by most accounts, they sold some few thousands of phones in those six weeks. Awfully close to zero, and those few? Probably sold by accident.
It's Google's decision to make it the carriers' responsibility to deliver an up-to-date OS release, or simply ignore the old devices. In that, yeah, it's partly Google's blame.
The UI stuff... different user home shells, I claim that's no big deal. You don't have to use HTC Sense or MotoBlur; there are close to a dozen alternate shells in the Android Market. Google might help here by allowing users with other home shells to download the bog standard Android shell. But it's not as if these things take more than a few minutes to learn, anyway.
The big problem is building Android too much like Linux used to be built... they need a modular architecture. Right now, your ROM is a big amalgam of built-in apps, OS, drivers, etc. What the really need is a hardware abstraction layer. The OEMs write very simple drivers for cameras, I/O devices, etc -- anything you need, just a basic driver. The OS then recognizes these, wraps them in higher level Linux or Android drivers as needed.
The result: Neither phone OEMs nor carriers ever HAVE to touch Android proper. They can update HAL modules independently of the whole OS, etc. And Google could push out a whole OS update in the Android Market that works on every device, no need to worry about embedded drivers.
That was one big win on the IBM PC -- they wrote the BIOS, which is really just "HAL for Dummies". So no need to worry a great deal about which OS the machine is running, etc. Sure, in those days, you didn't really have full hardware abstraction much, but that's totally practical in a machine as capable as today's Android devices today.
Flaws in measurement aside, it's not a continuous wire. The clips are actually part of two metal rails, presumably painted, which slide against each other, with a spring in-between. You would probably get some real current flowing if the [presumed] paint wore off internally; wear on the clips is not enough to cause a problem.
Now, I just looked at one of these owned by a co-worker -- he wasn't going to let me take it apart. It's possible the metal gives way to plastic under the leather cover, but I have no evidence of that -- sure seems like metal all the way. Stripped of paint and held just so, I'm sure it's less than 2 ohms. But you're probably going to get trouble with significantly higher resistance than that. Ok, not 2 megaohms :-)
The big problem with MediaFlo, DVB-H, etc. is that they were essentially obsolete before they were really finished. In both cases, the idea was to use a single TV channel, 6-8MHz, to transmit up to 32 sub-channels for mobile devices (a single MPEG-2 TS stream, which can carry up to 32 sub-channels). This pretty much means you're viewing a tiny image on an old-fashioned cellphone. The maxium resolution for DVB-H is 320x240|288 or so; they were really planning for tiny cellphone screens. Not sure about resolution, but the rate per digital sub-channel on MediaFlo, for example, is 200-250kb/s ... even YouTube aims higher than that for their lower resolution SD video.
In short, this is going to awful on a 2010-vintage smartphone... our smartphones are already typically 800x480 or better, full standard-definition, DVD-quality or thereabouts (a DVD is actually 720x480 NTSC, 720x576 PAL). And we already have IPTV of various sorts on these devices. A separate super-low-resolution, subscription-based broadcast TV option is not something anyone wants.
LTE supports 81.6Mb/s and 200 active data clients on a 5MHz channel, using 4x4 MIMO, 43.2Mb/s with 2x2 MIMO. Not quite as bad as you let on here. Obviously, if you had 200 data users on a cell, they're not all getting the high speed they're after... but no different than the 3G situation -- about 21Mb/s via HSPA or 56Mb/s with HSPA+ (2x2 MIMO and 64QAM).
AT&T has actually been buying up 700MHz spectrum for years now. Along with the national Block B 12MHz they bought at FCC Auction 73, they own up to 12MHz of Block C from some of these other purchases... I hadn't realized they bought Aloha Partners sometime back in 2007. So that's up to 24MHz in some areas, even before you factor in this new 6MHz block. LTE doesn't support more than 20MHz per channel, but in aggregate, AT&T may have up to 480Mb/s of LTE per cell. Not too shabby.
Verizon has been doing much the same thing... they spent $4.7 billion for the national 700MHz Block C (and a few licenses in Block A as well) they won in Auction 73, 22MHz wide. And another $4.66 billion buying up 700MHz spectrum owned by regional companies... no idea just how much, or where. But both companies are well situated for 4G, and clearly, the scarcity of this commodity is driving the price up. This has Verizon with up to 90MHz of aggregate spectrum in places (3G + 4G), over 45MHz through most of the country.
It'll be interesting to see if Echostar hangs on to their 6MHz Block E ($722 million), given they could better than double their money on it now. And the FCC still plans to run the Block D auction again, probably next year (Block D has to be shared with public service use).
Both UHF and VHF get attenuated, not entirely blocked, by objects: trees, houses, skyscrapers, etc... and both diffract around said objects. That's why your cellphone works indoors.
Higher frequency UHF is strongly attenuated by foliage, lower frequency UHF and VHF, no so much. As a digital radio designer, I did a 2.4GHz ISM band radio, narrowband, fairly long range, but it stopped dead at the edge of a forest. My next radio, at 435MHz (used to control robots) went completely through the forest, over a slight hill, and out to a roadway 1/2 mile away.
That's big reason so much money changed hands in the FCC's 700MHz auction. Verizon nabbed 22MHz there, AT&T 12MHz. So Flo TV's 6MHz, 716-722MHz, may not seem like much, but it boosts AT&Ts 4G band by 50%. They're closing down Flo TV in March, AT&T is expected to launch their LTE-based "4G" (not real 4G yet, but a first step) sometime over the summer. Unless, like everyone else has been, they're six months late.
There's no value in a 50 mile range for cellular.... no matter how powerful the tower, the handset is still going to be limited to about a watt, maybe less (US cellular devices can run up to 3W, but handsets usually max out a 1W or less... the cell tower actually tweaks the handset's output to something it can hear).
The 700MHz band is bandwidth limited compared to Sprint, for example.. the Sprint/Clear/Comcast WiMax network has about 90MHz in the 2500MHz band, so they have a higher peak capacity, for sure. But they're going to need more power per cell in cities, if not more cells, to get into buildings. And they're going to have issues with rural coverage -- as Sprint and T-Mobile already do for 2G and 3G. They're limited to 1900MHz, while AT&T and Verizon both have slots at both 850MHz and 1900MHz. My house is centered in 26 acres of forest. I can get Verizon in my cellar, AT&T though most of the house, but T-Mobile is pretty much outdoors only, while Sprint, last I checked, is available at the end of my driveway.
Of course, you can always deal with the limited bandwidth issue by using more, lower powered cells in highly populated areas. The only real fix for the higher frequency stuff outside of high population areas is more cells, something historically just never done by these guys,
Unfortunately, Windows + Tablet = fail, too.
There are multiple problems here, none likely to be fixed. The first problem is the UI... until Windows is as slick as iOS, WebOS, or Android (take your pick) on a finger based touch screen, the OS fails for tablets. Consumers at this level have already rejected stylus-based input, at least as the normal UI means.
Then you have the general unusability of much of Windows software. This is inherently a bottom-of-the-barrel device, performance-wise. To support it properly, a developer may have to give up on full support for higher-end PCs, or release multiple versions... negating the one supposed big advantage to Windows on a tablet: all that software. And where table performance is adequate, you have the same UI issue. The Windows API isn't robust enough to simply have the OS say "tablet", and every program suddenly work with a touch interface. So apps have to change too, again negating the Windows advantage.
And finally, the co-evolution of Windows and Intel's x86 agenda, until recently. Intel has been very successful at replacing add-on hardware with x86 software. They fell a bit on AVC decoding, but the GPUs have picked up that slack. And it's great... until you have to consider power. Which is the single most critical factor in building a successful tablet -- all day computing. If I have to worry that my battery's going to die over an 8-12 hour workday, no matter what I do on the tablet, that's a big fail. And certainly, many cheap ARM tablets fail here, too.. but usually by holding back on expensive components, like batteries, or more modern ARM SOCs with better power savings and alternate function units. For example, video: the better ARM SOCs don't use the CPU or GPU as the main video decoding engine, they have a separate unit for this, which dramatically reduces power consumption when displaying video. Windows systems have embraced the "space computing cycle" fact of the desktop to "let the CPU just do it", saving hardware money, but costing in power. This is not easily divorced from Windows, not even from Linux necessarily, as it's been the big evolutionary direction of the PC platform for 20+ years... ever since Intel noticed a faster wordprocessor wasn't going to drive PC upgrades anymore.
They don't cost anything significant in the base device. USB ports run about $0.50 each, for example. They do cost as an add-on... and you still can't get digital video out from an iPad. Why a device released in 2009 only offers CVBS or VGA out makes absolutely no sense.
But I do believe you have the answer here: Apple won't change. They're not going to offer additional ports, simply because they want customers to pay them for extra things. People who really need to hook their video or still cameras to an iPad will either live with the clumsy adapters, or more likely, buy a different device (that's one primary use of a tablet computer for me -- I would not even consider one that didn't have at least a card reader and one USB port as a built-in).
The earlier PC tablets were like the earlier compact laptops -- they cut out features and raise the price, sometimes dramatically. As a result, tablets and laptops with 10-12" screens were niche items.
The creation of the netbook changed this for small laptops. Sure, they were yet-even-more underpowered, though in an era in which most users don't need the power of the PC they've got, this is acceptable in a $200-$300 computer.
Apple took a different tack. They always sell overpriced laptops, and the reason they don't have an answer to the netbook, per se, is that selling a lower cost Macbook would destroy both their record-level margins on Mac hardware, and worse-yet, their ability to sell higher spec Macs at 2x-3x the same price as a PC.
Their solution, of course, is selling the inflated iPod that became the iPad. That's the same guts, larger screen and battery, as the iPod/iPhone... and it continues their high profit margins. Thing is, a $500 iPad is still only $500... toy money for the well-off. This is very different than the $2,000-$3,000 PC tablets that didn't compare, performance-wise, to a $500 generic PC notebook. The other ARM tablets suggest a similar level of acceptability.
And the PC tabs might too, in a vacuum. Building a tablet from Netbook DNA gets you a $200-$300 tablet, if not for 2011, then certainly for 2012. About all you're doing is taking stuff off the existing Netbook BOM... maybe add $20 for a touchscreen, but some Netbooks already had these, too. Another $20-$25 if you need a cellular LAN.
But there's no vacuum... what does the PC tablet bring that the ARM tablet doesn't? PC apps? Not so much... particularly for Windows. Sure, they'll technically run, but they don't work via a touch UI, they want far more resources than the tablet has available, etc. You're still selling worst-in-class performance. ARM tablets will all be running made-for-touch software evolved out of the smartphone world, and they'll all offer best-in-class performance. Sure, an Atom will compare in performance if not necessarily power savings to a dual core ARM Cortex A9 (the standard for all 2011 tablets), but does the x86 offer any advantage at that point? Is Intel building a whole system that can deliver the things tablet buyers will demand?
I very much doubt this, and here's why. Intel has spent the last 20+ years coming up with more work for the x86 to do. This has been very successful in keeping the PC processor centric, shedding nearly any bit of specialized software that came along (modem chips, audio synthesizers, etc) for another x86 device driver. But now there's the tablet -- it's expected to play HD video for 10+ hours straight, for example. To do most of what it does all day. Most of these are achieved by a basic ARM, small GPU, and a bunch of other dedicated units on an SOC that accelerate these specialty jobs. I have real doubts about Intel playing well here... they have the x86 doing too much work to compete against dedicated units. No matter how power light it gets, the dedicated stuff is going to run less power.
The bottom line will be the real value of the x86 tablet -- does it do anything consumers want, well enough to suggest that it's the correct answer? I don't believe we'll find out in 2011... Intel spent years getting the laptop x86 correct. They got hit occasionally, such as when Transmeta first shipped, but did eventually step up to clobber them. They got caught offguard by AMD's push into 64-bit x86 computing, but fixed that, too. Only this time, they're not playing in x86 land. And in fact, they're going up against the world's most popular 32-bit CPU, on a platform that probably won't benefit from Windows compatibility. This isn't going to be such an easy one.
Yeah... I figure, if I'm going going to be a human billboard, I either better like the product for the product's sake, or get paid for the ad placement. So I'll wear an R.E.M. or Springsteen T-Shirt, music I strongly recommend, or one from Martin Guitars or maybe Panasonic camcorders, since I use both of those products. But Abercrombie? A company largely distinguished by delivering the stupidest looking stores known to mankind, and hiring a couple of shirtless boys to hang out in font? That company... they'd have to pay BIG to place an ad on me.
It's a miracle of modern advertising kung-fu that some people will pay several times more for the same cheap Chinese-made t-shirts, if advertising is also included.
You're also ricking the future based on the status quo. Sure, it's free today. But what happens when the developer goes out of business.. can I still get that paid-for app? What happens if Google decides at some point that all those Chrome users would do wonders to their bottom line if they had to pay ($5? $10? $20?) for all these great web services they've been getting for free?
The personal computer revolution was a revolution because it took power away from corporate central administration and gave each individual control of their own computing power, data, etc. Smart phones have largely followed that model, but still ceeded some of that control, particularly in the case of the iPhone -- you've given away the "I can run any program" capability, even if that never personally matters. I like the web for transparent sync among networked devices, but that's a hard thing to make expensive, or proprietary.. data exchange is pretty easy.
But putting most everything online, you give up all real control. That may not matter now, but it could eventually.
Freedoms are rarely lost all at once. And usually, we get short-sighted and yield them, a tiny bit at a time, in return for some seemingly big deal: apps everywhere, a false sense of security, whatever. As a people, particularly in the USA, we seem less educated about these things than most of the world, despite our country's pivotal role in global democracy over the years.
Tech has been an enabler of freedom. PCs give us control of our own computing resources, so no agency (the government, China, Apple) gets to say what apps we can or cannot run. The internet gives us communications that, while not perfect, are very hard for any one entity to thwart. The internet, being the agent of the free flow of information, helped to take down the Soviet Union and create modern Eastern Europe. Together, computers and search give us the first effective "intelligence amplifier"... we can consult thousands of sources of information in the time people from 20 years ago could manage 2-4 in a library.
But the prospect of centralizing ANYTHING is the prospect of controlling that thing. As benevolent at it begins, it's ultimately at the whim of the aggergator.
It's a legit concern.
Actually, there are several. Data stored outside of your control means that, basically, it's outside of your control. If I'm paranoid, I can unplug any PC from the net... can't with this.
Where is it stored? The value of cracking commonly used PCs means that you get all sorts of attacks against Windows; less against less used OSs (MacOS) and/or those with higher levels of expertise and security (Linux).
But if this is all on a big Google server somewhere, there's now one single point of failure for all of that data. That's going to be one big-ass target, if it's at all successful.
Then there are other issues... like apps. Where are all the apps stored? I gather they run, at least in part, on the device, but are they permanent there, or just cached. Does the user have any choice about upgrades, or do they just happen, even if they break things? Are these stored at Google, or are they on the developer's computers? What happens when the server is down and I want to run my app? What happens when Company X goes out of business, but I've paid for that application -- does it still show up forever when I "sync to the cloud", or am I SOL.
There are some interesting things with the ChromeOS. One is simple: I'm sure I can run ChromeOS apps on a PC, and probably at some point on a smartphone or tablet, as well as on a ChromeOS device. This isn't for everything, but there may be a small class of apps that are really better when fully integrated with the net ("cloud", if you must). I was kind of skeptical, but I use a number of smartphone apps that are certainly enhanced based on cloud sync. While I still have POP3 mail and client on my PC, there are some real advantages to Google's GMail (which I also use).. they're the first ones really doing this right.
And certainly, if they're successful, there will be other folks hitting your data. For example, banks. Right now, I open a channel if I want to sync my financial software to my bank... it's on my terms, and if I'm really concerned about what it's doing, I can always fire up WireShark or something and see for myself. But put a Quicken-like app on ChromeOS, and the banks aren't going to worry about synching to my ChromeOS netbook; they're going to want to sync to my account. How do I manage or even know about what they can see and what they can't... and the other 2000 companies wanting the same access?
The PTO only really searches though the patent database for prior art. This is why all patents ought to be on public display for a year (as in Europe) before being granted.
At least, the 1999 filing date is correct, this will expire in 2019, if it's not struck down before then.
No.. I think Lucas is just doing that for experimental purposes. His real motive ... see, he's running out of reasons to re-edit Star Wars. So there's this project, apparently, to produce a fake 3D version, which will no doubt be as unwatchable as all other fake 3D films.
But once the technology is there, he'll be able to resurrect young Harrison Ford, young Carrie Fisher, etc. and remake Star Wars entirely from scratch, in the computer. He'll keep all the old dialog, but none of the original film at all -- all digital.
Laugh now.. laugh while you can, monkey-boy. But after seeing young Jeff Bridges brought back in the new Tron, this is clearly where Lucas wants to go.
And probably every other studio, too. Right now, actors are fairly important, and name-brand actors in particular. But once you can license an actor's likeness, living or not, then you just need a warm body (for a little while) on which to build that digital actor, and some good but non-name voice-over talent. And I'm sure they're working on both of those problems.
There are two things they do.
The "good" part in advancing the state of the iPhone.. new models, that make you WANT a new one (you, not me, I have no use for an iPhone). If you somehow NEED that upgrade, because Jobs and Co. dangle a shiny new carrot in front of you, that's no way you being forced.
On the other hand, if they're obsoleting the older models via software upgrades, you may well be forced to upgrade, or stop having access to available software. They just did that, after all... iOS 4 barely runs on the iPhone 3, doesn't on earlier iPhone and iPod models. AND they pushed developers to upgrade software for this and the iPad.. so the set of available apps that run on older models is drying up. That's how they force an upgrade.
I replaced the main board, CPU, and RAM on my primary system two weeks ago.. ran me about $500. The whole PC is far more expensive than a smart phone: monitors, hard drives, printers, scanners, keyboard, mouse, etc.