Yup... PowerVR has long been the dominant GPU in mobile devices. You'd have to say that just because they're in every iPhone and iPad, even if they weren't also in many other SOCs, like those from TI, Samsung, Marvell, NEC, and Renesas. And Intel.
Yup, they've been in netbooks... the PowerVR SGX535 usually goes by the alias "Intel GMA500" or "Intel GMA600" or whatever, but they've been the low-end default among Intel systems for awhile. That's probably changing, with Intel doing their own actual GPUs here and there.
As mentioned, AMD/ATi sold off their mobile GPU to Qualcomm, which calls it Adreno these days. And there have been some signs Qualcomm may be looking elsewhere, as they have yet to make that technology convincing. nVidia, too, hasn't been their usual desktop selves in the mobile space; the fastest graphics in mobile devices for the last year have been dual-core PowerVR in the Apple products, unless you count the quad-core PowerVR driven Sony Playstation Vita. ARM's MALI is also fairly new... ARM looking to capitalize on their crazy-successful processor core business with a whole soup-to-nuts solution.
While it's partially true that the industry standard pushed hard against the proprietary systems, that's no entirely true. Some of it was simply that the established proprietary system companies didn't try hard enough. It seems that color, video, and animation took Apple by surprise on the Mac.. they were offering what, the Apple ][ GS for that market? And they're the one that survived... only barely, and mostly because Steve Jobs came along and converted them into a high profit luxury CE company, not a personal computer company anymore (Macs are about 18% of Apple's business, and falling every quarter).
Commodore and the Amiga blazed that multimedia trail, with once revolutionary hardware and the best overall OS in the personal computer business at the time. But the bosses paid themselves more than the CEOs of Apple and IBM... combined, and spent way too little on R&D, despite Commodore at the time being more tied to custom chip development (eg, spending lots of money) than any of the others. Commodore couldn't remain competitive that way, but it was really more of a suicide than our being overrun by IBM compatibles.
Atari's management (the ex-Commodore Tramiels) didn't understand the difference between a late 80s/early 90s computer that needed real ongoing OS development at upgrades, and the Commodore 64 era of the OS basically being part of the hardware. Wang was already pretty much of it by then... they didn't really transition well from dedicated word processing gear to general purpose PCs. Rat Shack when totally IBM compatible, and eventually just didn't see any profit in making their own. Sinclair only put their toe into the 16-bit world anyway, with the very restricted QL, and hit enough trouble in 1985 to sell all their personal computer assets to Amstrad. DEC pretty much missed the idea that personal computers would grow more powerful than minicomputers faster than minis could keep up.
It happens... it shouldn't. But check out the notorious CADtrack XOR patent (US Patent 4,197,590). CADtrack was some kind of CAD company from way back, that emerged in the 80s as a patent troll. They had a patent on using the XOR operation to draw and undraw a cursor on a bitmapped screen. This should have completely failed the test of obviousness... I did exactly this as a kid of 17 on my first home computer, an Exidy Sorcerer (it didn't have full bitmapped graphics, but it had programmable character memory that could be used to simulate some game-class bitmapped displays with some cleverness). And at that point, I was mostly just self-taught "in the art", having read many issues of BYTE, Kilobaud, and Creative Computing. There should be a "kid rule"... if a High Schooler or below figured out your thing independently, you lose your patent.
A buddy of mine worked with the lawyers at Commodore on this, as we were being sued. He resolved to patent AND and OR... I don't think he could come up with the cash, though.
Sure looks like some monkey business here. This patent was filed in 2011, but it's actually a filed as a continuation. Here's the critical bit:
This patent application is a continuation of co-pending U.S. patent application Ser. No. 12/632,262, filed Dec. 7, 2009, which is a continuation of U.S. patent application Ser. No. 12/168,578, filed Jul. 7, 2008, and entitled DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/614,363, now U.S. Pat. No. 7,518,615, filed Jul. 12, 2000, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/098,041, now U.S. Pat. No. 6,650,327, filed Jun. 16, 1998, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, the disclosures of which are incorporated herein, in their entireties, by reference.
In short, they get to claim a priority date going back to 1998, as long as at least one patent application in the chain was in pending status when the next was filed, going all the way back to 1998. And that at least one inventor is listed in common. And, of course, that this filing (8144158 is all about the format of their floating point) is really an elaboration on what they did back in 1998, and not something different.
This is a really long time for this kind of thing... and that's just what they were after, of course. At least, under the new patent laws, the filing date and 20 year life are based on the actual filing date. Under the old system, you were given 17 years from the date of grant, so it was common to try to delay the grant date as long as possible, so that the things you did "way back when" were now commonplace. A legal submarine patent. This long string of continuations amounts to much the same kind of thing, though of course, they can't claim something different. For example, the 32-bit or 64-bit floating point units in current GPUs are probably not infringing, unless they derive directly from this design (unless it's determined that the FPU design is in itself obvious, or already covered by prior art). This also means that the 8144158 patent will expire in 2018 along with 6650327, not in 2031, which would be the expiration date of a new patent applied for in 2011.
From the troll company's viewpoint, it could well be that they found a bunch of GPUs out in the market using a design very similar to that of the old SGI GPUs, and decided to file this to enhance their case, which again, is just dandy as long as at least one patent in that long chain was still pending at the time this one was filed. It could be the defeat of the '327 patent against ATi illustrated their weakness going up against other devices.
I wondered who they're specifically after:
Apple iPhone: PowerVR (all. SGX543MP2 in the iPhone 4S) HTC EVO4G: Qualcomm Adreno 200 LG Thrill: PowerVR SGX540 RIM Torch: Qualcomm Adreno 205 Samsung Galaxy S: PowerVR SGX540 Samsung Galaxy SII: ARM MALI-400 Sony Xperia Play: Qualcomm Adreno 205
So that's basically every GPU maker in a mobile device other than nVidia. And not all that interesting that Qualcomm's on the list, since they're now licensing PowerVR anyway, and so Adreno may be relegated to economy devices, or gone completely in short time. This was probably critical to Qualcomm's reputation, with the Krait-based S4 SOCs, they'll otherwise have the fastest per-CPU-core ARM on the market, at least until A15s are out. No need to hamstring that with the weakest GPU... but I digress.
The patent clearly can't cover all uses of floating point, and even the patent authors clearly understood this. But claims, as always, try to suggest they cover everything.
In truth, a patent covers a specific way of doing something. So assuming there's no other reason to judge this a bad patent, you still have to look at exactly how they implemented their floating point solution in hardware. For example, some fragment of SGI went after ATi some years back, and the court rejected the '327 patent as not applying, even though ATi was also using floating point rasterization. If your FPU engine uses a similar floating point implementation and architecture, even if it's an improvement, it probably infringes. If you went an entirely different way but still happen to use floating point, probably not.
Obviously, there are edge cases, which is the whole point of having a patent court -- it's not always obvious if one device infringes on anothers' patent. And the courts aren't a panacea, either, they sometimes make the wrong technical decisions. Just the best system that exists for this kind of thing so far (though in terms of digging up prior art, the community as a whole often does this public service much better than the courts). Keep in mind that ultimately, the claims themselves don't have legal weight, other than to illustrate the part of the patent the patentee claims is unique. You have to look at the text of the patent when deciding what infringes, and specially, the very specific preferred and alternate embodiments called out in the main description. Many have tried, but in theory, you can't patent an algorithm.
That's incorrect -- the description is the basis for granting the patent, it's the part that actually matters. The claim matter, too, but in a different way. They're calling out parts of the invention description that the inventor claims are unique. But the patent is granted on the basis of the description, not the claims. The claims are checked, but not with any precision. It's quite common for claims to be judged overly broad, or thrown out entirely. They have to relate directly back to the description of the invention, and they act largely as footnotes -- they can suggest, for example, that deviations from the preferred embodiment of the invention (what's discussed in the description, for any utility patent) are just dandy, but that in itself has no legal power. The PTO will usually eliminate overly broad claims, but not always. It's the body they're primarily concerned with -- that IS the invention.
The claims are for the courts, and very much left open to interpretation. In fact, when you write a patent, you KNOW that the PTO is not overly careful in reviewing claims, and so you try to make the claim sound like it's covering anything you can get away with. But, in fact, it doesn't. A patent covers a very specific thing, not a general case of all similar things. This is one reason that companies, particularly patent trolls like this fragment of the once great SGI, don't like to go to court. Once in court, a claim can be judged to not apply. But it can also be tossed out entirely, either based on being overly broad, or in light of prior art. When that happens, you can't use it again, even to scare someone else into licensing. If significant enough, it may even call the PTO to re-examine the entire patent.
Actually, the claims only have a certain level of legal weight. The actual patent is granted on the "Description of Invention" -- the main body of the patent. That's the part the PTO reads and analyzes against prior art. The claims are generally only checked against the body, to ensure they're supported by the body, and I'm sure, not written too outlandishly.
Now, once you're in court, it's the claims that are used to prosecute infringement. But they'll be challenged, or even tossed out, if they aren't properly backed up by the body of the patent, once you get to trial. It's true that you can't claim infringement other than on claims, but the claims are not the final answer. They're more like an annotated index into your description of what you believe the actual "invention" to be.
I'm not a lawyer. I do write patents, and analyze the technical merits for lawyers, from time to time. Not my finest work, but I'm good at it.
Actually, if you're running 4K video over HDMI, that's HDMI 1.4 by spec, though it wouldn't be a shock if some 1.3 devices met this spec level, The 1.3 standard maxes out at the usual PC graphics card cap of 2560x1600 or thereabouts.
Apple fanbois will complain about any technology that Apple's not using, with the dogmatic belief that, if Apple's not using it, there must be a flaw.
Anything you do with your portable device eats battery. LTE is not all that significant in more recent LTE devices. Thing is, no one shipped an LTE device based otherwise on 2009-2010 phone sensibilities and performance. Well, ok, Nokia has, but who's paying attention to Windows Phone?
So when I got my LTE-based Galaxy Nexus last December, it didn't just have LTE, it had a much faster dual core CPU (dual-memory bus TI OMAP 4460 at 1.2GHz), much higher resolution screen, four times the memory of my old phone, etc. All of that draws more power, peak, than an older device like my retired O. G. Droid. I look at my phone now, it's been on battery 8h 22m, the battery's at 80%, the screen accounts for 32%, Android for 22%, the cellular modem just for 10% (3G at home, 4G at work).
The early G5s were faster than the Pentium 4s of the day. But not the Opterons or AMD64s... the very first Apple G5 machine, Apple's self-professed "supercomputer", was outperformed a few months after it shipped by similarly packaged media-editing systems based on the Opteron from Boxx Technologies.
Apple's also extrapolating here... and not well. They're claiming a linear speed improvement based on using four PowerVR cores rather than two. But they haven't scaled their CPUs... they're still running a dual core Cortex A9, and given that they aren't mentioning the clock speed, it's probably more or less 1GHz. The Tegra 3 has four A9 cores, and may run as fast as 1.6GHz. That does affect graphics performance, particularly one you leave the world of artifical benchmarks, where the CPUs are actually calculating on the fly, rather than just feeding pre-calculated coordinates to the GPU.
And of course, all of these are based on a unit screen -- 1280x720, with display disabled ("offscreen test"). So yeah, it really doesn't matter how well they do on this test; what matters is comparing actual screen displays, one to another. Or old to new... a 4x boost over the iPad 2 is going to be needed just to break even on the perception of performance.
It's quite true that the tile-based rendering in the PowerVR chips is very efficient, but that comes at the cost of accuracy. For the early small displays on portable devices, these were seen as a reasonable alternative to the desktop GPU. But once you have this kind of resolution, the nVidia is simply going to render better looking 3D than the PowerVR.
There are other performance issues here, too. For one, Apple has traditionally supported the ARM "NEON" SIMD instruction set on iOS. It's been more of evolving thing in the Android NDK... and not all Android-supported SOCs (for example, the Tegra 2) support NEON. The Tegra 3 does. Another performance difference, at least in real world code, that may not be factored well into artificial benchmarks.
Open source solutions for media content creation do exist, but most of them aren't much better than the $50-$100 entry level programs from these same media companies. Same goes for CAD tools. The $10,000-or-so copy of Altium Designer my company was nice enough to buy for me is so significantly better than any open source replacement, it's not funny. And I really do hope all our competitors are trying to do 8-12 layer impedance controlled PCBs with open source tools.
$10,000 sounds like a great deal to a non-professional. But if you consider that I'd get that for two PCB layouts, if I was working along (or possibly one complex one), it doesn't really sound that expensive for a professional. Now, figure if I could do three PCB layouts per month with the pro tool, and only two with the free one, I'm going to be very far ahead. And honestly, while it's been a little while since I looked at the freebies, there's really nothing of professional quality for CAD work in the FOSS world. Nothing much for audio or video, either. The best FOSS tool I've played around with for any media content work is Cinelerra, and while you might actually do professional-level editing with it, at least on some projects, it's still way behind tools like Vegas Pro, Media Composer, and Premiere Pro.
You can repeat a thing a billion times, doesn't make it true. Using pirated software is stealing. It's a different kind of theft than stealing a car or a television, but it's just as much a theft. And if you steal a $10,000 program, you'll find the law doesn't have a much different opinion about this than if you stole a $10,000 car. And I'm certain that the guy stealing the $10,000 car will have as many self-rationalizations about his crime as you do yours. Doesn't make either of them "not stealing".
At $10,000, they're certainly a niche in the video market. Most software-only professional video editing tools run around $1,000 or less (Adobe, Sony, Avid, etc). Sure, you can spend much more for niche stuff... $40K+ for a fully featured professional Blu-ray authoring suite like Blu-print or Scenarist HD, similar levels for some of the hardware accelerated versions of Avid... HW included.
The bottom line for this class of software: yeah, like all other bits of software, it's going to get pirated. You actually should not care so much about someone stealing your software who's never even going to shell out $1,000 or $100 for a similar product. They're not serious users, and not potential customers. You're selling a tool at that price only to people who make money from it. So all you really have to worry about is whether your actual customers are stealing extra copies or whatever, not that some 16 year old kid in Germany got a copy and is trying to figure it out without a manual.
The worst thing you can do is make the program unusable to regular users in an effort to stop piracy. You won't.. it's sad, but it's well demonstrated that better copy protection only sets up a more lucrative target for crackers. And that's another thing... there are people out there who break copy protection just to break it. They have no intention of ever using the program, they simply get a thrill out of breaking your protection. Yeah, it's a typical case of those who can't create being destroyers. Good news is that they may grow up... I knew a bunch of alleged C64 software pirates who become competitive demo writers and eventually real programmers... some of these crackers are kids, still in school. Bad news.. the next generation is already learning to code:-)
As a dedicated user of video editing tools (Sony Vegas Pro, Boris BCC and FX, Cineform, etc) I know you're not targeting this product at the semi-pro or small operations like event videographers... you're going to be selling this to a fairly small number of larger operations, I suspect (without knowing the details, obviously... but the only $10K program I use is a CAD program, paid for by my business, and yes, very heavily pirated). As others mentioned, you have many other ways of keeping them honest. Some of the pro-level CAD tools I've used will "phone home", so the developers can track how many copies any given customer runs at a given time. Some are node-locked, so the average user isn't going to be able to run it on their PC, but a computer wiz can work around that. Others watch the LAN and refuse to run if they see another with the same license running, but that's easily stopped by a Firewall. Others still require a floating license server, so you can run it on any PC you like, but only as many instances as you have licenses.
It's also notable that, in many higher-end pro markets, most of the software developer's income is from support contracts. The high initial price of my CAD program is hard for a small company to swallow, but you get a year's real upgrades, and the continued support license is relatively low, $1K-$2K, with some guarantee of new versions being delivered, and of course, real and effective phone support. I'm not going to get much of any useful support out of Sony when Vegas malfunctions; but the Altium people will answer my emails or phone calls ASAP. That's another thing you get for the 10x-20x price difference. You have to be that guy, too.
Yeah.. the Catholic Church official position is this: "Methodical research in all branches of knowledge, provided it is carried out in a truly scientific manner and does not override moral laws, can never conflict with the faith, because the things of the world and the things the of the faith derive from the same God. The humble and persevering investigator of the secrets of nature is being led, as it were, by the hand of God in spite of himself, for it is God, the conserver of all things, who made them what they are".
In short, Science can't conflict with the Church's teachings. Either the scientists got it wrong, or the Church is interpreting the Bible incorrectly. This came directly from Pope Leo XIII... dude even had a cool name. But Santorum is far more of the flavor of the Nutty Far Right Evangelicals... the folks he's counting on for support.
Indeed. Any religion really concerned just about abortion would be handing out condoms on the street. It's been kind of funny to watch some of the political talk shows lately, now that the Radical Right has launched their latest brilliant idea: the War on Birth Control. The Radical Right Representative du jour in the hotseat typically just implodes then the show host points out that making birth control readily available always lowers the abortion rate, and the things they're trying to do will increase it.
Funny, too, how these guys are yelling from the one side of the mouth about government getting up in everyone's business (even when it's not true), and from the other side of the mouth, advocating government control of the most intrusive possible kinds, pretty much all in the bedroom.
The difference between being an atheist and being a Christian or Muslim is this: 1. As a Christan or Muslim, you reject every other god, every other religion but your own... so you're about 5,000 against, one pro. I'm 5,001 against. Not all that different... I just reject one more magical big-daddy in the sky than you do; we agree pretty much on the other 5,000... though I rekon, if you're a Muslim or various flavors of Christianity, you'll want to do more harm to the followers of those other 5,000 magical mythical mojo-hobos than I will. Despite your religion's teaching to the contrary....
Yeah, this guy wants to bring on the American Taliban... he thinks government and religion (his, of course) should be the same thing.
But he's not as stupid that suggests, and he's clearly a follower of Karl Rove's strategies, if not as capable at it. The major innovation that Rove brought to political warefare was this: find your weakest point, and attack that first in your opponent. Who cares if the attack it true or fictional, the simple fact that you attacked first makes the counterattack weak.
Look at the 2004 election... John Kerry was a decorated war hero, George Bush served in the National Guard, and even there got out of actually doing much serving thanks to his family's political connections. In the past, the Bush people would have done anything possible to avoid even discussing military service. Rove orchestrates the "Swift Boat Veterans for Truth", sells a bunch of complete lies about Kerry, and pretty much destroys that avenue of attack against Bush. Evil but frickin' brilliant... didn't help that Kerry didn't have a clue what to do about it.
And the Republican Party has been repeatedly anti-science. Sure, they'll claim as much the Democrats to want the business piece of science, but when they get right down to it, their policies have sent research elsewhere. Their energy policy directly lead to China so dominating solar, that more than half of the US solar companies have gone under in the last five years. Their wacko-religious policies on stem cell research all but halted it in the USA, while it's flourishing in Europe. Research when successful becomes business; neither of those are businesses easily won back. And that's not even getting into Creationism, Intelligent Design, or whatever new name they try next for the same basic goal: teaching the Christian creation myth to children as a credible scientific theory, rather than the fairy story it is. Only 22% of Republicans believe that global warming/climate change even exists, much less that it's man made... and they're sure to let you know, every time it snows... even when the increased snowfall actually is due to a warmer weather pattern.
The Dems generally fail on nuclear energy (most of the scientific community is in favor of building new, modern reactors, particular interesting are Thorium reactors), irradiated foods, and even a growing faction is anti-vaccination. And animal research, though that's objection is based on moral, nor scientific grounds.
But don't forget, it was the Republicans who put through the "Noah's Ark" version of the formation of the Grand C
Radios can kill the battery, just like leaving anything else on. But they can also save battery, depending. Bluetooth running take up less power than your audio amplifier, so BTing at Class 3 power levels (0dBm) to Bluetooth speakers will actually use less power than a headphone. Some phones don't even support Class 1 (20dBm = 100mW), which is where you're going to see the power drain. Wi-Fi is usually maxxed out at 100mW as well, while the cellular radio will usually to at least 1/2W peak. The cell tower sets your uplink power level, so operating in buildings, particularly on higher frequencies, can drain the battery much faster than when you have a strong signal back to the tower. This is why 4G on lower frequencies can actually drain the battery less, depending on the real-world circumstances.
I see 10-15Mb/s down, 8-12Mb/s up, on a typical day in Philly. It's nice to be on an unlimited plan... you certainly can chew through that much faster than the typical 800Kb/s down on Verizon EvDO 3G (much slower than HSPA or HSPA+ flavors of 3G, but it does have some advantages, like only needed 2.5MHz bandwidth).
No, T-Mobile uses HSPA+ because they don't have any spectrum for 4G. If they had 4G spectrum, and the cash to build out a 4G network, they'd be trumpeting it just like everyone else. They also suffer due to higher frequencies... less free air propagation at 1700/2100MHz, more attenuation through buildings and foliage, etc.
Huh... the Galaxy Nexus I mentioned is very much a 4G LTE phone (well, as 4G as things get today... by the time we have LTE Advanced on our phones, the radio chips will have had a generation or two of power reductions -- it's the RF over-the-air stuff you'll never improve, the digital parts will get better with each generation). And yeah, you can kill the battery... but that's true of 3G devices as well. Radio isn't a huge factor most of the time... only when the signal is very weak and the RF needs to go out at 1/2W or whatever your particularly phone peaks at. And given the frequency advantage of LTE on 700MHz, this can be a substantial win, particularly over 3G at 850/1900 or 1700/2100, not to mention WiMax at 2500MHz.
I have one of those devices.. a Samsung Galaxy Nexus. And yeah, you can go through the battery.. even the extended battery.. in a short run.
Of course, that was also very true of the unit my Nexus replaced.. the original Verizon Droid. Great device, at least until the touchscreen fell out of calibration and Mot didn't offer, well, a calibration tool. Doing intensive network stuff on a weak connection, I could trash that Droid's battery in 2-3 hours. Good thing I had 5 replacement cells. But that had nothing specific to do with the 3G connection.
Yes, the 4G/LTE hardware can suck some power. The good news is that the RF uplink part is actually lower power than 3G. The digital part, right now, definitely uses more power than the 3G original.
And 4G isn't always even a problem. I work in Philly, in a 100-year-old former car factory (Marketplace Design Center, 2400 Market Street). If I move very far away from my windowed office there on the fourth floor, my 3G connection fades fast. That's not just a problem of connectivtiy,but one of battery. As my 3G signal gets weak, the cell tower will tell my phone to pump up the volume. The peak on a smartphone is on the order of 500mW-1W total output. In short, bad news for your battery. When I go on 4G in the same locales, I have a far, far better signal. That's just 700MHz through brick, morter, and steel. So in the building, 4G is actually saving me power.
You get the radio limiting life on the edge. Every summer, I spend a week on an island in a big lake in New Hampshire. Cell links didn't used to reach us at all, now, you can at least get a voice quality link on most of the island. But you pay a price... I have seen folks' phones lasting less than two days on the island. They're being told to, er,. pump up the volume, right now. The 4G vs. EvDo vs. HPSA vs. CDMA2000 vs. GSM issues are not even relevant.
Thing is, unlike mainstream articles designed to scare children and/or non-iOS users, this isn't even important. When I look at my phone's power use (and yeah, pretty easy to do), that beautiful 1280x720 AM-OLED screen is still sucking doen 40%+ of the power, even if it's in all ways more spectacular than an iPhone 4/4S display.. and maybe even sligthly lower power. Deep down on the menu is the radio modem, 3G or 4G. That's usually about 4% of the power draw. If 4G were actually twice as power hungry, you wouldn't notice the run-time change based on this extra power draw.
The real issue is that we all got 2011 phones release in 2011... all kinds of nice new hardware, all kinds of things that want power, like hi-rez screens. Apple was very conservative, so maybe they're doing ok on a 1400mAhr cell, but nearly everything else about that phone is lower-end than at least most of the higher end Androids. And I'd like to hear from an avid gamer... the 4S is well suited to games, but what does THAT do to one's battery life?
Regardless, 4G is a straw man for iPhone fanbois. If they could get it on an iPhone, you'd never hear an ill word about it. This is, as always, trying to sell backward iPhone tech against those moving much faster. Any time you hear an Apple guy going down on the tech, it's not real... it's an attempt to frame an argument that probably hasn't even happened yet.
Slashdot Mirror
Yup... PowerVR has long been the dominant GPU in mobile devices. You'd have to say that just because they're in every iPhone and iPad, even if they weren't also in many other SOCs, like those from TI, Samsung, Marvell, NEC, and Renesas. And Intel.
Yup, they've been in netbooks... the PowerVR SGX535 usually goes by the alias "Intel GMA500" or "Intel GMA600" or whatever, but they've been the low-end default among Intel systems for awhile. That's probably changing, with Intel doing their own actual GPUs here and there.
As mentioned, AMD/ATi sold off their mobile GPU to Qualcomm, which calls it Adreno these days. And there have been some signs Qualcomm may be looking elsewhere, as they have yet to make that technology convincing. nVidia, too, hasn't been their usual desktop selves in the mobile space; the fastest graphics in mobile devices for the last year have been dual-core PowerVR in the Apple products, unless you count the quad-core PowerVR driven Sony Playstation Vita. ARM's MALI is also fairly new... ARM looking to capitalize on their crazy-successful processor core business with a whole soup-to-nuts solution.
While it's partially true that the industry standard pushed hard against the proprietary systems, that's no entirely true. Some of it was simply that the established proprietary system companies didn't try hard enough. It seems that color, video, and animation took Apple by surprise on the Mac.. they were offering what, the Apple ][ GS for that market? And they're the one that survived... only barely, and mostly because Steve Jobs came along and converted them into a high profit luxury CE company, not a personal computer company anymore (Macs are about 18% of Apple's business, and falling every quarter).
Commodore and the Amiga blazed that multimedia trail, with once revolutionary hardware and the best overall OS in the personal computer business at the time. But the bosses paid themselves more than the CEOs of Apple and IBM... combined, and spent way too little on R&D, despite Commodore at the time being more tied to custom chip development (eg, spending lots of money) than any of the others. Commodore couldn't remain competitive that way, but it was really more of a suicide than our being overrun by IBM compatibles.
Atari's management (the ex-Commodore Tramiels) didn't understand the difference between a late 80s/early 90s computer that needed real ongoing OS development at upgrades, and the Commodore 64 era of the OS basically being part of the hardware. Wang was already pretty much of it by then... they didn't really transition well from dedicated word processing gear to general purpose PCs. Rat Shack when totally IBM compatible, and eventually just didn't see any profit in making their own. Sinclair only put their toe into the 16-bit world anyway, with the very restricted QL, and hit enough trouble in 1985 to sell all their personal computer assets to Amstrad. DEC pretty much missed the idea that personal computers would grow more powerful than minicomputers faster than minis could keep up.
It happens... it shouldn't. But check out the notorious CADtrack XOR patent (US Patent 4,197,590). CADtrack was some kind of CAD company from way back, that emerged in the 80s as a patent troll. They had a patent on using the XOR operation to draw and undraw a cursor on a bitmapped screen. This should have completely failed the test of obviousness... I did exactly this as a kid of 17 on my first home computer, an Exidy Sorcerer (it didn't have full bitmapped graphics, but it had programmable character memory that could be used to simulate some game-class bitmapped displays with some cleverness). And at that point, I was mostly just self-taught "in the art", having read many issues of BYTE, Kilobaud, and Creative Computing. There should be a "kid rule"... if a High Schooler or below figured out your thing independently, you lose your patent.
A buddy of mine worked with the lawyers at Commodore on this, as we were being sued. He resolved to patent AND and OR... I don't think he could come up with the cash, though.
Sure looks like some monkey business here. This patent was filed in 2011, but it's actually a filed as a continuation. Here's the critical bit:
This patent application is a continuation of co-pending U.S. patent application Ser. No. 12/632,262, filed Dec. 7, 2009, which is a continuation of U.S. patent application Ser. No. 12/168,578, filed Jul. 7, 2008, and entitled DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/614,363, now U.S. Pat. No. 7,518,615, filed Jul. 12, 2000, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/098,041, now U.S. Pat. No. 6,650,327, filed Jun. 16, 1998, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, the disclosures of which are incorporated herein, in their entireties, by reference.
In short, they get to claim a priority date going back to 1998, as long as at least one patent application in the chain was in pending status when the next was filed, going all the way back to 1998. And that at least one inventor is listed in common. And, of course, that this filing (8144158 is all about the format of their floating point) is really an elaboration on what they did back in 1998, and not something different.
This is a really long time for this kind of thing... and that's just what they were after, of course. At least, under the new patent laws, the filing date and 20 year life are based on the actual filing date. Under the old system, you were given 17 years from the date of grant, so it was common to try to delay the grant date as long as possible, so that the things you did "way back when" were now commonplace. A legal submarine patent. This long string of continuations amounts to much the same kind of thing, though of course, they can't claim something different. For example, the 32-bit or 64-bit floating point units in current GPUs are probably not infringing, unless they derive directly from this design (unless it's determined that the FPU design is in itself obvious, or already covered by prior art). This also means that the 8144158 patent will expire in 2018 along with 6650327, not in 2031, which would be the expiration date of a new patent applied for in 2011.
From the troll company's viewpoint, it could well be that they found a bunch of GPUs out in the market using a design very similar to that of the old SGI GPUs, and decided to file this to enhance their case, which again, is just dandy as long as at least one patent in that long chain was still pending at the time this one was filed. It could be the defeat of the '327 patent against ATi illustrated their weakness going up against other devices.
I wondered who they're specifically after:
Apple iPhone: PowerVR (all. SGX543MP2 in the iPhone 4S)
HTC EVO4G: Qualcomm Adreno 200
LG Thrill: PowerVR SGX540
RIM Torch: Qualcomm Adreno 205
Samsung Galaxy S: PowerVR SGX540
Samsung Galaxy SII: ARM MALI-400
Sony Xperia Play: Qualcomm Adreno 205
So that's basically every GPU maker in a mobile device other than nVidia. And not all that interesting that Qualcomm's on the list, since they're now licensing PowerVR anyway, and so Adreno may be relegated to economy devices, or gone completely in short time. This was probably critical to Qualcomm's reputation, with the Krait-based S4 SOCs, they'll otherwise have the fastest per-CPU-core ARM on the market, at least until A15s are out. No need to hamstring that with the weakest GPU... but I digress.
The patent clearly can't cover all uses of floating point, and even the patent authors clearly understood this. But claims, as always, try to suggest they cover everything.
In truth, a patent covers a specific way of doing something. So assuming there's no other reason to judge this a bad patent, you still have to look at exactly how they implemented their floating point solution in hardware. For example, some fragment of SGI went after ATi some years back, and the court rejected the '327 patent as not applying, even though ATi was also using floating point rasterization. If your FPU engine uses a similar floating point implementation and architecture, even if it's an improvement, it probably infringes. If you went an entirely different way but still happen to use floating point, probably not.
Obviously, there are edge cases, which is the whole point of having a patent court -- it's not always obvious if one device infringes on anothers' patent. And the courts aren't a panacea, either, they sometimes make the wrong technical decisions. Just the best system that exists for this kind of thing so far (though in terms of digging up prior art, the community as a whole often does this public service much better than the courts). Keep in mind that ultimately, the claims themselves don't have legal weight, other than to illustrate the part of the patent the patentee claims is unique. You have to look at the text of the patent when deciding what infringes, and specially, the very specific preferred and alternate embodiments called out in the main description. Many have tried, but in theory, you can't patent an algorithm.
That's incorrect -- the description is the basis for granting the patent, it's the part that actually matters. The claim matter, too, but in a different way. They're calling out parts of the invention description that the inventor claims are unique. But the patent is granted on the basis of the description, not the claims. The claims are checked, but not with any precision. It's quite common for claims to be judged overly broad, or thrown out entirely. They have to relate directly back to the description of the invention, and they act largely as footnotes -- they can suggest, for example, that deviations from the preferred embodiment of the invention (what's discussed in the description, for any utility patent) are just dandy, but that in itself has no legal power. The PTO will usually eliminate overly broad claims, but not always. It's the body they're primarily concerned with -- that IS the invention.
The claims are for the courts, and very much left open to interpretation. In fact, when you write a patent, you KNOW that the PTO is not overly careful in reviewing claims, and so you try to make the claim sound like it's covering anything you can get away with. But, in fact, it doesn't. A patent covers a very specific thing, not a general case of all similar things. This is one reason that companies, particularly patent trolls like this fragment of the once great SGI, don't like to go to court. Once in court, a claim can be judged to not apply. But it can also be tossed out entirely, either based on being overly broad, or in light of prior art. When that happens, you can't use it again, even to scare someone else into licensing. If significant enough, it may even call the PTO to re-examine the entire patent.
Actually, the claims only have a certain level of legal weight. The actual patent is granted on the "Description of Invention" -- the main body of the patent. That's the part the PTO reads and analyzes against prior art. The claims are generally only checked against the body, to ensure they're supported by the body, and I'm sure, not written too outlandishly.
Now, once you're in court, it's the claims that are used to prosecute infringement. But they'll be challenged, or even tossed out, if they aren't properly backed up by the body of the patent, once you get to trial. It's true that you can't claim infringement other than on claims, but the claims are not the final answer. They're more like an annotated index into your description of what you believe the actual "invention" to be.
I'm not a lawyer. I do write patents, and analyze the technical merits for lawyers, from time to time. Not my finest work, but I'm good at it.
Actually, if you're running 4K video over HDMI, that's HDMI 1.4 by spec, though it wouldn't be a shock if some 1.3 devices met this spec level, The 1.3 standard maxes out at the usual PC graphics card cap of 2560x1600 or thereabouts.
Apple fanbois will complain about any technology that Apple's not using, with the dogmatic belief that, if Apple's not using it, there must be a flaw.
Anything you do with your portable device eats battery. LTE is not all that significant in more recent LTE devices. Thing is, no one shipped an LTE device based otherwise on 2009-2010 phone sensibilities and performance. Well, ok, Nokia has, but who's paying attention to Windows Phone?
So when I got my LTE-based Galaxy Nexus last December, it didn't just have LTE, it had a much faster dual core CPU (dual-memory bus TI OMAP 4460 at 1.2GHz), much higher resolution screen, four times the memory of my old phone, etc. All of that draws more power, peak, than an older device like my retired O. G. Droid. I look at my phone now, it's been on battery 8h 22m, the battery's at 80%, the screen accounts for 32%, Android for 22%, the cellular modem just for 10% (3G at home, 4G at work).
The early G5s were faster than the Pentium 4s of the day. But not the Opterons or AMD64s... the very first Apple G5 machine, Apple's self-professed "supercomputer", was outperformed a few months after it shipped by similarly packaged media-editing systems based on the Opteron from Boxx Technologies.
Apple's also extrapolating here... and not well. They're claiming a linear speed improvement based on using four PowerVR cores rather than two. But they haven't scaled their CPUs... they're still running a dual core Cortex A9, and given that they aren't mentioning the clock speed, it's probably more or less 1GHz. The Tegra 3 has four A9 cores, and may run as fast as 1.6GHz. That does affect graphics performance, particularly one you leave the world of artifical benchmarks, where the CPUs are actually calculating on the fly, rather than just feeding pre-calculated coordinates to the GPU.
And of course, all of these are based on a unit screen -- 1280x720, with display disabled ("offscreen test"). So yeah, it really doesn't matter how well they do on this test; what matters is comparing actual screen displays, one to another. Or old to new... a 4x boost over the iPad 2 is going to be needed just to break even on the perception of performance.
It's quite true that the tile-based rendering in the PowerVR chips is very efficient, but that comes at the cost of accuracy. For the early small displays on portable devices, these were seen as a reasonable alternative to the desktop GPU. But once you have this kind of resolution, the nVidia is simply going to render better looking 3D than the PowerVR.
There are other performance issues here, too. For one, Apple has traditionally supported the ARM "NEON" SIMD instruction set on iOS. It's been more of evolving thing in the Android NDK... and not all Android-supported SOCs (for example, the Tegra 2) support NEON. The Tegra 3 does. Another performance difference, at least in real world code, that may not be factored well into artificial benchmarks.
Yes, if a programmer is stupid enough to encode their watermark in plain text, it's easy to remove. But no good programmer would do that.
Open source solutions for media content creation do exist, but most of them aren't much better than the $50-$100 entry level programs from these same media companies. Same goes for CAD tools. The $10,000-or-so copy of Altium Designer my company was nice enough to buy for me is so significantly better than any open source replacement, it's not funny. And I really do hope all our competitors are trying to do 8-12 layer impedance controlled PCBs with open source tools.
$10,000 sounds like a great deal to a non-professional. But if you consider that I'd get that for two PCB layouts, if I was working along (or possibly one complex one), it doesn't really sound that expensive for a professional. Now, figure if I could do three PCB layouts per month with the pro tool, and only two with the free one, I'm going to be very far ahead. And honestly, while it's been a little while since I looked at the freebies, there's really nothing of professional quality for CAD work in the FOSS world. Nothing much for audio or video, either. The best FOSS tool I've played around with for any media content work is Cinelerra, and while you might actually do professional-level editing with it, at least on some projects, it's still way behind tools like Vegas Pro, Media Composer, and Premiere Pro.
You can repeat a thing a billion times, doesn't make it true. Using pirated software is stealing. It's a different kind of theft than stealing a car or a television, but it's just as much a theft. And if you steal a $10,000 program, you'll find the law doesn't have a much different opinion about this than if you stole a $10,000 car. And I'm certain that the guy stealing the $10,000 car will have as many self-rationalizations about his crime as you do yours. Doesn't make either of them "not stealing".
At $10,000, they're certainly a niche in the video market. Most software-only professional video editing tools run around $1,000 or less (Adobe, Sony, Avid, etc). Sure, you can spend much more for niche stuff... $40K+ for a fully featured professional Blu-ray authoring suite like Blu-print or Scenarist HD, similar levels for some of the hardware accelerated versions of Avid... HW included.
The bottom line for this class of software: yeah, like all other bits of software, it's going to get pirated. You actually should not care so much about someone stealing your software who's never even going to shell out $1,000 or $100 for a similar product. They're not serious users, and not potential customers. You're selling a tool at that price only to people who make money from it. So all you really have to worry about is whether your actual customers are stealing extra copies or whatever, not that some 16 year old kid in Germany got a copy and is trying to figure it out without a manual.
The worst thing you can do is make the program unusable to regular users in an effort to stop piracy. You won't.. it's sad, but it's well demonstrated that better copy protection only sets up a more lucrative target for crackers. And that's another thing... there are people out there who break copy protection just to break it. They have no intention of ever using the program, they simply get a thrill out of breaking your protection. Yeah, it's a typical case of those who can't create being destroyers. Good news is that they may grow up... I knew a bunch of alleged C64 software pirates who become competitive demo writers and eventually real programmers... some of these crackers are kids, still in school. Bad news.. the next generation is already learning to code :-)
As a dedicated user of video editing tools (Sony Vegas Pro, Boris BCC and FX, Cineform, etc) I know you're not targeting this product at the semi-pro or small operations like event videographers... you're going to be selling this to a fairly small number of larger operations, I suspect (without knowing the details, obviously... but the only $10K program I use is a CAD program, paid for by my business, and yes, very heavily pirated). As others mentioned, you have many other ways of keeping them honest. Some of the pro-level CAD tools I've used will "phone home", so the developers can track how many copies any given customer runs at a given time. Some are node-locked, so the average user isn't going to be able to run it on their PC, but a computer wiz can work around that. Others watch the LAN and refuse to run if they see another with the same license running, but that's easily stopped by a Firewall. Others still require a floating license server, so you can run it on any PC you like, but only as many instances as you have licenses.
It's also notable that, in many higher-end pro markets, most of the software developer's income is from support contracts. The high initial price of my CAD program is hard for a small company to swallow, but you get a year's real upgrades, and the continued support license is relatively low, $1K-$2K, with some guarantee of new versions being delivered, and of course, real and effective phone support. I'm not going to get much of any useful support out of Sony when Vegas malfunctions; but the Altium people will answer my emails or phone calls ASAP. That's another thing you get for the 10x-20x price difference. You have to be that guy, too.
Yeah.. the Catholic Church official position is this: "Methodical research in all branches of knowledge, provided it is carried out in a truly scientific manner and does not override moral laws, can never conflict with the faith, because the things of the world and the things the of the faith derive from the same God. The humble and persevering investigator of the secrets of nature is being led, as it were, by the hand of God in spite of himself, for it is God, the conserver of all things, who made them what they are".
In short, Science can't conflict with the Church's teachings. Either the scientists got it wrong, or the Church is interpreting the Bible incorrectly. This came directly from Pope Leo XIII... dude even had a cool name. But Santorum is far more of the flavor of the Nutty Far Right Evangelicals... the folks he's counting on for support.
Indeed. Any religion really concerned just about abortion would be handing out condoms on the street. It's been kind of funny to watch some of the political talk shows lately, now that the Radical Right has launched their latest brilliant idea: the War on Birth Control. The Radical Right Representative du jour in the hotseat typically just implodes then the show host points out that making birth control readily available always lowers the abortion rate, and the things they're trying to do will increase it.
Funny, too, how these guys are yelling from the one side of the mouth about government getting up in everyone's business (even when it's not true), and from the other side of the mouth, advocating government control of the most intrusive possible kinds, pretty much all in the bedroom.
The difference between being an atheist and being a Christian or Muslim is this: 1. As a Christan or Muslim, you reject every other god, every other religion but your own... so you're about 5,000 against, one pro. I'm 5,001 against. Not all that different... I just reject one more magical big-daddy in the sky than you do; we agree pretty much on the other 5,000... though I rekon, if you're a Muslim or various flavors of Christianity, you'll want to do more harm to the followers of those other 5,000 magical mythical mojo-hobos than I will. Despite your religion's teaching to the contrary....
Yeah, this guy wants to bring on the American Taliban... he thinks government and religion (his, of course) should be the same thing.
But he's not as stupid that suggests, and he's clearly a follower of Karl Rove's strategies, if not as capable at it. The major innovation that Rove brought to political warefare was this: find your weakest point, and attack that first in your opponent. Who cares if the attack it true or fictional, the simple fact that you attacked first makes the counterattack weak.
Look at the 2004 election... John Kerry was a decorated war hero, George Bush served in the National Guard, and even there got out of actually doing much serving thanks to his family's political connections. In the past, the Bush people would have done anything possible to avoid even discussing military service. Rove orchestrates the "Swift Boat Veterans for Truth", sells a bunch of complete lies about Kerry, and pretty much destroys that avenue of attack against Bush. Evil but frickin' brilliant... didn't help that Kerry didn't have a clue what to do about it.
And the Republican Party has been repeatedly anti-science. Sure, they'll claim as much the Democrats to want the business piece of science, but when they get right down to it, their policies have sent research elsewhere. Their energy policy directly lead to China so dominating solar, that more than half of the US solar companies have gone under in the last five years. Their wacko-religious policies on stem cell research all but halted it in the USA, while it's flourishing in Europe. Research when successful becomes business; neither of those are businesses easily won back. And that's not even getting into Creationism, Intelligent Design, or whatever new name they try next for the same basic goal: teaching the Christian creation myth to children as a credible scientific theory, rather than the fairy story it is. Only 22% of Republicans believe that global warming/climate change even exists, much less that it's man made... and they're sure to let you know, every time it snows... even when the increased snowfall actually is due to a warmer weather pattern.
The press is full of similar accounts:
http://www.nytimes.com/2011/08/29/opinion/republicans-against-science.html
http://io9.com/5835970/will-the-anti+science-republicans-kill-conservatism-as-americans-know-it
http://www.huffingtonpost.com/leonard-steinhorn/how-the-gop-became-the-an_b_970410.html
http://www.usatoday.com/news/opinion/forum/story/2011-09-20/gop-democrats-science-evolution-vaccine/50482856/1
http://www.waronscience.com/home.php
The Dems, of course, don't have a perfect track record, and on some issues, are closer to the Republicans, or even worse:
http://reason.com/archives/2011/12/27/whos-more-anti-science-republicans-or-de
http://www.usatoday.com/news/opinion/forum/story/2011-09-20/gop-democrats-science-evolution-vaccine/50482856/1
The Dems generally fail on nuclear energy (most of the scientific community is in favor of building new, modern reactors, particular interesting are Thorium reactors), irradiated foods, and even a growing faction is anti-vaccination. And animal research, though that's objection is based on moral, nor scientific grounds.
But don't forget, it was the Republicans who put through the "Noah's Ark" version of the formation of the Grand C
Radios can kill the battery, just like leaving anything else on. But they can also save battery, depending. Bluetooth running take up less power than your audio amplifier, so BTing at Class 3 power levels (0dBm) to Bluetooth speakers will actually use less power than a headphone. Some phones don't even support Class 1 (20dBm = 100mW), which is where you're going to see the power drain. Wi-Fi is usually maxxed out at 100mW as well, while the cellular radio will usually to at least 1/2W peak. The cell tower sets your uplink power level, so operating in buildings, particularly on higher frequencies, can drain the battery much faster than when you have a strong signal back to the tower. This is why 4G on lower frequencies can actually drain the battery less, depending on the real-world circumstances.
I see 10-15Mb/s down, 8-12Mb/s up, on a typical day in Philly. It's nice to be on an unlimited plan... you certainly can chew through that much faster than the typical 800Kb/s down on Verizon EvDO 3G (much slower than HSPA or HSPA+ flavors of 3G, but it does have some advantages, like only needed 2.5MHz bandwidth).
No, T-Mobile uses HSPA+ because they don't have any spectrum for 4G. If they had 4G spectrum, and the cash to build out a 4G network, they'd be trumpeting it just like everyone else. They also suffer due to higher frequencies... less free air propagation at 1700/2100MHz, more attenuation through buildings and foliage, etc.
Huh... the Galaxy Nexus I mentioned is very much a 4G LTE phone (well, as 4G as things get today... by the time we have LTE Advanced on our phones, the radio chips will have had a generation or two of power reductions -- it's the RF over-the-air stuff you'll never improve, the digital parts will get better with each generation). And yeah, you can kill the battery... but that's true of 3G devices as well. Radio isn't a huge factor most of the time... only when the signal is very weak and the RF needs to go out at 1/2W or whatever your particularly phone peaks at. And given the frequency advantage of LTE on 700MHz, this can be a substantial win, particularly over 3G at 850/1900 or 1700/2100, not to mention WiMax at 2500MHz.
I have one of those devices.. a Samsung Galaxy Nexus. And yeah, you can go through the battery.. even the extended battery.. in a short run.
Of course, that was also very true of the unit my Nexus replaced.. the original Verizon Droid. Great device, at least until the touchscreen fell out of calibration and Mot didn't offer, well, a calibration tool. Doing intensive network stuff on a weak connection, I could trash that Droid's battery in 2-3 hours. Good thing I had 5 replacement cells. But that had nothing specific to do with the 3G connection.
Yes, the 4G/LTE hardware can suck some power. The good news is that the RF uplink part is actually lower power than 3G. The digital part, right now, definitely uses more power than the 3G original.
And 4G isn't always even a problem. I work in Philly, in a 100-year-old former car factory (Marketplace Design Center, 2400 Market Street). If I move very far away from my windowed office there on the fourth floor, my 3G connection fades fast. That's not just a problem of connectivtiy,but one of battery. As my 3G signal gets weak, the cell tower will tell my phone to pump up the volume. The peak on a smartphone is on the order of 500mW-1W total output. In short, bad news for your battery. When I go on 4G in the same locales, I have a far, far better signal. That's just 700MHz through brick, morter, and steel. So in the building, 4G is actually saving me power.
You get the radio limiting life on the edge. Every summer, I spend a week on an island in a big lake in New Hampshire. Cell links didn't used to reach us at all, now, you can at least get a voice quality link on most of the island. But you pay a price... I have seen folks' phones lasting less than two days on the island. They're being told to, er,. pump up the volume, right now. The 4G vs. EvDo vs. HPSA vs. CDMA2000 vs. GSM issues are not even relevant.
Thing is, unlike mainstream articles designed to scare children and/or non-iOS users, this isn't even important. When I look at my phone's power use (and yeah, pretty easy to do), that beautiful 1280x720 AM-OLED screen is still sucking doen 40%+ of the power, even if it's in all ways more spectacular than an iPhone 4/4S display.. and maybe even sligthly lower power. Deep down on the menu is the radio modem, 3G or 4G. That's usually about 4% of the power draw. If 4G were actually twice as power hungry, you wouldn't notice the run-time change based on this extra power draw.
The real issue is that we all got 2011 phones release in 2011... all kinds of nice new hardware, all kinds of things that want power, like hi-rez screens. Apple was very conservative, so maybe they're doing ok on a 1400mAhr cell, but nearly everything else about that phone is lower-end than at least most of the higher end Androids. And I'd like to hear from an avid gamer... the 4S is well suited to games, but what does THAT do to one's battery life?
Regardless, 4G is a straw man for iPhone fanbois. If they could get it on an iPhone, you'd never hear an ill word about it. This is, as always, trying to sell backward iPhone tech against those moving much faster. Any time you hear an Apple guy going down on the tech, it's not real... it's an attempt to frame an argument that probably hasn't even happened yet.
Supporting objective truth IS the neutral position. Google and Wikipedia are not the ones in the wrong here.