The big trick with submarine patents doesn't work under the new system. Back in the day, your patent was for 17 years from date of patent grant. So it was a big thing to apply, knowing the patent would be bounced back and forth between your company and the PTO, intentionally delaying the grant date. This allowed for legal submarine patents... you don't have to disclose anything (in the USA) until the patent is granted. In today's revised system, you get 20 years from date of application, so there's little incentive for this kind of foot-dragging.
But here's the other thing... it's quite possible, particularly given the nature of software, for clashes to go undiscovered. Certainly the H.264 people did a search of any patents violations that might be pertinent, but they also solved the problem for commercial concerns by licensing through the MPEG-LA. Some new patent shows up covering any H.264 IP, and the MPEG-LA people deal with it, not Apple or Google (well, at least in theory). But also, H.264 is about as public as a thing can get... no one with video compression IP can claim to be ignorant of it.
This is not the case with Ogg Theora. Yes, the VP3 patents that had been granted to On2 are permanently licensed for VP3 open source use, and more recently, the changes Xiph have done to create Theora. But how about Theora itself... there's no guarantee that it's not stepping on other patents. And just because On2 had VP3 patents, that doesn't guarantee that VP3 itself was in the clear from others' patents, either. Perhaps On2 did all that work and all (probably), but it's up to On2 or Xiph to verify this, not the patent holders, who don't necessarily know a thing about the Theora project.
And now there's VP8... unreleased. Google has work to do, to ensure that it's not patent encumbered by outside patents. They can afford that... but FOSS projects typically can't. So there's always some risk for a big company buying into something like Ogg Theora, simply because there might be unknown patent encumberances that, literally, show up at their door some day soon.
This, of course, is a good example of why software patents should not exist. The original idea of an invention was to give the inventor a short monopoly on the use of that invention in return for public disclosure, to allow others to build on it. But it's rare that software benefits from such public disclosure, and rarer still that there's the kind of invention in software that's all that novel. Add to that the fact that, for the most part, the PTO's definition of "prior art" when they examine a patent is not much broader than "prior patents". They have this "if it's been invented, it's been patented" mentality, which is why so many patents are issued despite volumes and decades of known prior art.
1080p isn't remotely of interest to Apple's current product line, outside of Macs which already handle it. They're not even delivering a 720p screen on the iPad, much less a 1080p output. iPods output CVBS analog, iPads can do that, or VGA, the same 1024x768 they do on-screen.
Maybe for the future... after eveyone else is delivering 1080p over built-in HDMI, Apple will care about it. Or not. Or they'll wait for one of the wireless HD video procotols. Of course, Apple's only concerned with video you can buy from the iTunes store, anyway, and they only sell the low bitrate 720p or the SD stuff, simply because 1080p can't play on any device they currently make. They're in no hurry to change... 1080p doesn't matter on a 3.5" or 8" screen.
But even then, it's not an immediate concern. First of all, H.264 is going to likely be the most popular video CODEC for the next 10 years even if Google has a smash success with VP8 and the FOSS community gets it all exactly their ideal way. Why? Because everyone else will still use H.264. There won't be a revision of Blu-Ray to incorporate VP8, unless it's so good they want it for the next generation 4K players... which are probably at least 10 years off. Cable and satellite have a heavy investment in H.264... satellite moved to H.264 to increase the number of channels, but unless VP8 (or something else) offers profound advantages on bandwidth, they're going to replace 10 million+ receivers, not to mention all the head end gear. US broadcast is still MPEG-2, Europe is H.264 in HD, and they're not going to change anytime soon.
And, back to Apple... they're a closed system. They were using H.264 early on, back when most people thought "MPEG4" was going to just be for cellphones, and even Sony was MPEG-2 only on their Blu-Ray prototypes (with cartridge, some years back). They're still a closed system, even if the world moves onto something better. There's no big win for Apple moving away from H.264, period.
Right. Back in the 1980s, a standard PC couldn't handle MPEG-1 decoding very well. It took about 300+ MHz and MMX before they stopped putting MPEG-2 hardware decoders in the few PCs that actually played DVDs back then. MPEG-2 offer about 1/2-1/3 the coding efficiency of a modern video CODEC like H.264.
On my Q9550 PC (that's four cores at 2.83GHz) and a nVidia 8800GT CPU (118 stream processors), I did some playing around with some H.264 samples at 1080/60p, just to make it interesting. Running the latest VLC under Windows 7 took 50-60% of that system, and it was choppy playback. That's a pretty hefty price to pay for decoding, that means you simply can't play that, in VLC anyway, on a more ordinary 2-core CPU.
Using full video acceleration (under Windows 7's DXVA 2.0 API, which is supported by the H.264 CODEC included in Win7), I got down to 12% CPU. That's making much better use of the graphic card's processor array and other features it may have (color space conversion was the original "video acceleration" available on graphic cards, for example).
That kind of processing will work on PCs for other kinds of video CODECs. But not applicable to devices like smart phones, which do more of this in tuned hardware specific to H.264 and DCT algorithms. Most smart phones don't have reprogrammable processing akin to that GPU. Even the few that do (like the OMAP chips' DSP) are still counting on hardware to do much of the acceleration.
Google already does run multiple encodings of your submission: lower bitrate, HQ, HD... the submissions process decides what kind of encodings to do.
They are not going to support Ogg Theora for the simple reason it'll cost them money. The CODEC is not as efficient as H.264, so they wind up with higher bitrates or lower quality.
But moving to VP8, they could absolutely offer VP8 versions. These could play on PCs after an update, and if VP8 really is of higher coding efficiency than H.264, Google would save money. Their big expense is bandwidth, not CPU cycles or storage. I believe Google would be happy to pay off some bandwidth with a little extra storage and a few more encoding cycles.
There are objections preventing YouTube from using Ogg Theora, but you're a bit off base here.
Yeah... devices. For You-Tube class low bitrate 720p, any old regular PC should have no trouble playing H.264 or Theora or probably even VP8. The problem is precisely the small devices, from Netbooks though handhelds. On a desktop these days, you have a combination of hardware and software in the GPU (via accelerations interfaces like DXVA in Windows) to speed up video. But on a handheld, what you'll probably find are a bunch of hardware units called up by a conventional CPU or perhaps a DSP, but the hardware units tend to be specific to H.264 and very similar video schemes. And also, perhaps not well described in public docs, so even your ability to run other CODECs is poor, at least without broad industry support making it a priority.
Transcoding on YouTube would not be a huge issue. For one, it doesn't have to happen overnight, and for two, there could be a prize at the end. But not for Theora. Theora is a lower efficiency CODEC than H.264, and that's obvious even when you look over the samples that Xiph.org keeps touting... and that's about the best example you'll find anywhere, simply because it's the best rigged demo they could produce without telling outright lies.
The big expense of YouTube isn't video storage or transcoding cycles, it's bandwidth... all the data being schlepped in and out of the site every day. Going to a less efficient CODEC like Theora will result in a lower quality video or much more expense. And given YouTube's continued lack of profitability, don't think Google's going to budget on this for a femtosecond. But reportedly, VP8 offers higher efficiency than H.264, particularly at lower bitrates. Moving to VP8 might save Google serious green, over the years. Enough to say, buy On2? Well, hey, they did... they must have some plan for it.
With the H.264 royalties not set to kick in until 2015 now, there's plenty of time for a transition. If Google open sourced VP8 tomorrow and did the free license thing on any patents On2 might have on the technology, this would be a very good thing, even if it's not yet a perfect solution. They don't have to back-transcode everything right away, but they could start encoding new submissions in H.264 for "device" use and VP8 for the desktop (where the acceleration is not AS needed, and easier to re-program given that much of it's running on stream processors rather than dedicated hardware). If VP8 became THE long term standard for the internet, devices would start supporting it.
Also, keep in mind that many of the cable companies run really fast fiber to the neighborhood node. So you're not really dealing with coax over crazy distances. I don't think anyone's running new copper for this kind of thing, it's only the last mile or less issue, whether you fiber, coax, or well, nothing, to the home.
The Constitution grants the Federal Government the right to pass laws to deal with some things not specifically addressed in the Constitution, and the States rights to deal with others.
Given that radio waves, much less fiber optic internet, had not yet been discovered in 1787, this is a very clear case in which one needs not simply heed the Constitution, but all of the law built on top of it since.
You may now return to drinking that teabagger kool-aid.
Yeah, the "no more copper" issue was pretty much the death DSL expansion. It's a product of competiton... the cable companies were shooting at the telco's main profit... once you have cable TV and decent broadband, taking on voice service is a small deal. So the telcos had to respond by offering the same video, internet, and voice bundle. You don't get that over copper.
It is actually pretty reasonable, if you're talking about peak service. I mean, this is already commonplace in South Korea. Hanaro Telecom rolled it out and in six months, had 200,000 customers on DOCSIS 3.0 at 100Mb/s... back in 2007. Today, most of the country is on 100Mb/s fiber. This all started with a mandate back in 2003.
This year, the UK's already taking about a similar program, to wire the whole country at 100Mb/s by 2017. The USA simply has to say something, or we're going to look more like the hillbillies of internet connectivity than we already do.
Mostly, the telephone and power lines are there because the telcos and power companies got breaks in return for universal coverage. So AT&T got to be a monopoly basically as long as they wanted to be one. They actually subsidized everyone's local service by overcharging for other things, like long distance, some kind of business services, etc. The actual cost over time of the additional coverage was very small, given that they got the whole pie to themselves.
Yup.. the market is currently free enough to ensure I get no possible solution for wired broadband where I live. A totally free market will offer a half dozen solutions to the most populated areas, and none to the less populated areas. After all, winning 25% of a city of 100,000 is far more profitable than winning 100% of a city of 1,000.
And like the man says, that's how power and telephone worked. I'm quite sure I'd be off-grid entirely, if telcos and power carriers had not been forced to hook me up. Or, at the minimum, I would have had to pay for the power run, transformer, etc.
That's an easy one to solve... I'd deal with that in a heatbeat. Find a reasonable neighbor, offer to pay for their interent access if they'll let you set up a wireless link. Plain old 802.11g with a couple of Yagi or "coffee can" directional antennas, and you're good for hundreds of feet. Better with 802.11n, but only if you're wiring for MIMO (2 or 3 antennas at either end, and issues with where they're placed if you're optimizing it).
I actually design radios in my day job, and one such device is a mesh router that can run up to about six miles. I've been really tempted to tap real broadband in neighboring towns... the frequencies used, illegal as hell, unless your're police or the military... but tempting anyway.
Mandating higher speeds isn't even really necessary. Most places with any kind of decent internet service usually have competition, like FiOS or DSL vs. Cable, whatever. Natural competition has kept those guys one-upping each other for more than a decade already, and that's not going to stop. Accelerating that will just offer them an excuse for not expanding coverage out to those of us living in not-so-highly-profitable areas. Hell, at this point, I'd happily pony up $120 a month for FiOS... I'm paying that now for relatively crappy satellite.
Fiber to the house, and you're "out in sticks". Are you sure? Can you see any neighbors from any point on the ground around your house?
I'm jealous.. I'm out in the sticks. Or maybe it's the boondocks, hard to tell. But I'm not talkin' Kansas or Idaho here, this is South Jersey. There's a town 3 miles West, another 5 miles to the East... not exactly the mountains of New Mexico.
I can't wired-anything to my house, unless you count POTS. POTS works pretty well here for POTS, largely because there's a DSL-capable local node across the street... I can see it from my mailbox. But no telco is willing to drop as couple of DSL boards in there (yes, it's compatible with DSL). So I'm paying $120 a month for 1.5Mb satellite, with a 500MB per day download cap.
Meanwhile, the folks who had 784kb/s DSL, then 3Mb/s DSL, then 5Mb/s cable, then 12Mb/s cable, are now getting 25Mb/s FiOS upgrades.
But my forest IS really nice in the warmer months...
Same is true in South Jersey, where we did get a whole lot of snow this year.
It's kind of obvious that global warming will lead initially to more snow. That's independent of the origins of this year's snows.. whether global warming was involved or not. The colder you get, the less water vapor in the air. More heat means more energy in the system which means more potential snow. Based on the studies of average temperature, the global average from the 1850s throught to 2005 is 1.36F. That's not the difference between snow and now snow, except in rare conditions. On the other hand, a fraction of that warming the tropical Pacific can cause an El Niño episode, which has a very dramatic effect on weather for the whole USA. More energy in the system leads to more variation and more severe weather, not anything you'd easily call out as "warming".
And just a little extra energy alone isn't the reason we went from a usual 10-15" of snow for a whole season to over 70"... so far. The weather is definitely weird this year. We usually get a cold snap in late December maybe through mid January, with temps down to the 10F's or so... this year, not so much. It's hovered in the upper 20Fs and lower 30Fs. But the storms! The three huge blizzards were due to nor'easters... big circulating storms, kind of like a small hurricane, that basically hover around the coast and bring all kind of precipitation inland a bit. That's why, for the first time ever in my recollection, we got much more snow than most of Eastern PA. Thing is, a nor'easter is usually a fall and spring storm, not so much a winter thing. Is that the end result of more energy in the system?
I'm sure people who actually know this stuff, actual scientists, not just we computer geeks, have their opinions. "Near-normal amounts of precipitation are expected over the eastern third of the country, as well as over the Pacific Northwest and Northern Plains, while drier-than-normal conditions are forecast to occur over the Southwest and the Upper Midwest/Great Lakes."... wait, that was the Farmer's Almanac, for the 2009-2010 winter season. So here's an actual climate scientist: http://www.democracynow.org/2010/2/12/climate_scientist_record_setting_mid_atlantic
And hey, why trust me... I'm just a computer engineer with a guitar. But man, if you can't trust Bill Nye the Science Guy, there's just something wrong about you. Here's Bill, speaking about the snow storms, and questioning the patriotism of the "science deniers": http://www.youtube.com/watch?v=sm05Mcah0i8
The point is, unless you have the "world is black and white" vision of a radical right teabagger, or maybe you've lived in California your whole life and have never seen snow beyond "the ski slopes", you understand that snow != cold. Sure, you need a little cold to get snow, but if you have lots of cold, no snow. Sure, some northern climes look very snowy this time of year, but that's the effect of many tiny bits of snow without a melt. Here, we usually melt out a few days after any snowstorm... this year, with 30" in the same week, it's taking JUST a bit longer. Up at my brother-in-law's place in up-state New York, the snow they get in December is probably going to hang around until March. So they don't have much falling at any given time, but it accumulates.
Simple enough, when you actually live it every year.
Actually, it's the cows. Over 50% of all the corn we grow freed cows, chickens, and other livestock. Before anyone can complain about ethanol taking food away from people (regardless of the relative efficiency of corn sugar to ethanol... about as good as oil sands to petrol), you had better shut down the beef and poultry industries.
Microsoft did just what Palm did.. they added phone APIs to an existing PDA OS. Doesn't matter how long you've been at it... it matters how well you do it.
Of course, this is only temporary... if you want to support iPhone and iPad, you'll have to worry about resolution. You already have to worry about performance, given that the iPhone 3GS has about twice the CPU and 2x-4x the GPU of previous iPhones. And if Apple's planning any followup to the iPhone 3GS, they're going to need a nominally 800x480 screen, just like all of the current higher-end smart phones today.
So it's an issue... Apple's just putting it off, and regular users aren't thinking about it.
And of course, the other view: these are the one or two models you get to chose. They are perfect for every single human being on the planet, because Steve says they are. Sorry... that's a fail. Which is why others are leading the future of the smartphone now, just as Apple's played catch-up on the PC front.
It's not that Apple didn't aid Adobe in porting Flash to the iPhone, it's that they proactively prohibited Adobe from releasing Flash for the iPhone. Big difference.
But yeah, Microsoft is working with Adobe toward a Flash release. Adobe already officially announced Flash and AIR for Android. I don't they'd have any problem putting it on Palm's WebOS devices, either. And it's already on some Nokia devices. So like Flash or not, Apple will be the only mobile device voluntarily setting itself up as a lesser web experience.
Because Jobs doesn't like Flash, doesn't like multitasking, and wants to ensure Apple has complete control of the iPhone applications world. There are no technical reasons to not have Flash on your device.
App-switching, at least old PalmOS style (eg, you come back to exactly where you left off) works about 50% of the time. Of course, it worked well on a Palm because the PalmOS put everything in RAM, and just left it there. Having to reload every app, that's going to slow things down.
When you have real multitaking, you get every advantage of task switching, and no disadvantages. If your app is waiting on user I/O, and you lose focus, it's there in memory (unless it needs to be kicked out for space reasons), but it's not consuming cycles. So it's functionally identical to app-switching. But when you need a background process, that works, too. Apple has these on the iPhoneOS, they just don't allow anyone but Apple to write them.
Here are some more... so I can run a locale aware daemon, to turn off my Wifi when I'm out of the house or office, swap out the "office inappropriate" background when I'm at the office, kick on Bluetooth when I'm in the car, etc. Not necessary, but it saves on battery power.
Another GPS daemon... I can track where I go. Run this, and it keeps a record of my travels, no disruption of any kind to other uses of the phone/computer. This is extremely useful for geotagging photos and video shot with my various GPS-lacking cameras.
How about listening to any media player in the "background", not just the one my phone manufacturer deems suitable for that task (I like Pandora for web audio, Museek for regular MP3 player functionality). Maybe I'm in the car, and want to run Navigator at the same time... no problems.
How about jumping between the web, the message I'm composing, the calculator I brought up to make a few calculations, and my project notes for all the various projects I'm working on, all without having to stop the Mahjong game I was playing before I got that email (also read without stopping).
In short... ever use a real computer? It's just like that.
It's interesting how you can slice data and get the results you want. If you compare Apple to Apple, year to year, they look great in 4Q09. But if you compare Apple 3Q09 to Apple 4Q09, Apple actually lost market share (18% down to 16%), even though that was their best quarter ever. They did not keep up with the growth of the smartphone market, quarter to quarter. http://blogs.wsj.com/digits/2010/02/01/iphone-loses-market-share-in-fourth-quarter/tab/article/
That's more interesting than looking year to year, as it reflects the current way Apple is trending. Given their three-something years in the market, one whole year is just too long to spot a trend.
Actually, Nokia had been slipping for awhile, but seems to have bounced back a bit in 4Q09. They were managing about 35% of the global smart phone market in 2009, but in the last quarter, they boosted their share to 40%. Great, but not the 50% so many often claim. http://www.ft.com/cms/s/0/5af7cd20-0c05-11df-96b9-00144feabdc0.html?nclick_check=1
Keep in mind that the smartphone market itself is actually growing pretty fast, too.
Slashdot Mirror
That's correct. But the anti-H.264 people like to say "proprietary" because it's an effective scare-word.
Patents, once granted, are publically known.
The big trick with submarine patents doesn't work under the new system. Back in the day, your patent was for 17 years from date of patent grant. So it was a big thing to apply, knowing the patent would be bounced back and forth between your company and the PTO, intentionally delaying the grant date. This allowed for legal submarine patents... you don't have to disclose anything (in the USA) until the patent is granted. In today's revised system, you get 20 years from date of application, so there's little incentive for this kind of foot-dragging.
But here's the other thing... it's quite possible, particularly given the nature of software, for clashes to go undiscovered. Certainly the H.264 people did a search of any patents violations that might be pertinent, but they also solved the problem for commercial concerns by licensing through the MPEG-LA. Some new patent shows up covering any H.264 IP, and the MPEG-LA people deal with it, not Apple or Google (well, at least in theory). But also, H.264 is about as public as a thing can get... no one with video compression IP can claim to be ignorant of it.
This is not the case with Ogg Theora. Yes, the VP3 patents that had been granted to On2 are permanently licensed for VP3 open source use, and more recently, the changes Xiph have done to create Theora. But how about Theora itself... there's no guarantee that it's not stepping on other patents. And just because On2 had VP3 patents, that doesn't guarantee that VP3 itself was in the clear from others' patents, either. Perhaps On2 did all that work and all (probably), but it's up to On2 or Xiph to verify this, not the patent holders, who don't necessarily know a thing about the Theora project.
And now there's VP8... unreleased. Google has work to do, to ensure that it's not patent encumbered by outside patents. They can afford that... but FOSS projects typically can't. So there's always some risk for a big company buying into something like Ogg Theora, simply because there might be unknown patent encumberances that, literally, show up at their door some day soon.
This, of course, is a good example of why software patents should not exist. The original idea of an invention was to give the inventor a short monopoly on the use of that invention in return for public disclosure, to allow others to build on it. But it's rare that software benefits from such public disclosure, and rarer still that there's the kind of invention in software that's all that novel. Add to that the fact that, for the most part, the PTO's definition of "prior art" when they examine a patent is not much broader than "prior patents". They have this "if it's been invented, it's been patented" mentality, which is why so many patents are issued despite volumes and decades of known prior art.
1080p isn't remotely of interest to Apple's current product line, outside of Macs which already handle it. They're not even delivering a 720p screen on the iPad, much less a 1080p output. iPods output CVBS analog, iPads can do that, or VGA, the same 1024x768 they do on-screen.
Maybe for the future... after eveyone else is delivering 1080p over built-in HDMI, Apple will care about it. Or not. Or they'll wait for one of the wireless HD video procotols. Of course, Apple's only concerned with video you can buy from the iTunes store, anyway, and they only sell the low bitrate 720p or the SD stuff, simply because 1080p can't play on any device they currently make. They're in no hurry to change... 1080p doesn't matter on a 3.5" or 8" screen.
But even then, it's not an immediate concern. First of all, H.264 is going to likely be the most popular video CODEC for the next 10 years even if Google has a smash success with VP8 and the FOSS community gets it all exactly their ideal way. Why? Because everyone else will still use H.264. There won't be a revision of Blu-Ray to incorporate VP8, unless it's so good they want it for the next generation 4K players... which are probably at least 10 years off. Cable and satellite have a heavy investment in H.264... satellite moved to H.264 to increase the number of channels, but unless VP8 (or something else) offers profound advantages on bandwidth, they're going to replace 10 million+ receivers, not to mention all the head end gear. US broadcast is still MPEG-2, Europe is H.264 in HD, and they're not going to change anytime soon.
And, back to Apple... they're a closed system. They were using H.264 early on, back when most people thought "MPEG4" was going to just be for cellphones, and even Sony was MPEG-2 only on their Blu-Ray prototypes (with cartridge, some years back). They're still a closed system, even if the world moves onto something better. There's no big win for Apple moving away from H.264, period.
Right. Back in the 1980s, a standard PC couldn't handle MPEG-1 decoding very well. It took about 300+ MHz and MMX before they stopped putting MPEG-2 hardware decoders in the few PCs that actually played DVDs back then. MPEG-2 offer about 1/2-1/3 the coding efficiency of a modern video CODEC like H.264.
On my Q9550 PC (that's four cores at 2.83GHz) and a nVidia 8800GT CPU (118 stream processors), I did some playing around with some H.264 samples at 1080/60p, just to make it interesting. Running the latest VLC under Windows 7 took 50-60% of that system, and it was choppy playback. That's a pretty hefty price to pay for decoding, that means you simply can't play that, in VLC anyway, on a more ordinary 2-core CPU.
Using full video acceleration (under Windows 7's DXVA 2.0 API, which is supported by the H.264 CODEC included in Win7), I got down to 12% CPU. That's making much better use of the graphic card's processor array and other features it may have (color space conversion was the original "video acceleration" available on graphic cards, for example).
That kind of processing will work on PCs for other kinds of video CODECs. But not applicable to devices like smart phones, which do more of this in tuned hardware specific to H.264 and DCT algorithms. Most smart phones don't have reprogrammable processing akin to that GPU. Even the few that do (like the OMAP chips' DSP) are still counting on hardware to do much of the acceleration.
OT8... the Official Video Format of Scientology!
Google already does run multiple encodings of your submission: lower bitrate, HQ, HD... the submissions process decides what kind of encodings to do.
They are not going to support Ogg Theora for the simple reason it'll cost them money. The CODEC is not as efficient as H.264, so they wind up with higher bitrates or lower quality.
But moving to VP8, they could absolutely offer VP8 versions. These could play on PCs after an update, and if VP8 really is of higher coding efficiency than H.264, Google would save money. Their big expense is bandwidth, not CPU cycles or storage. I believe Google would be happy to pay off some bandwidth with a little extra storage and a few more encoding cycles.
There are objections preventing YouTube from using Ogg Theora, but you're a bit off base here.
Yeah... devices. For You-Tube class low bitrate 720p, any old regular PC should have no trouble playing H.264 or Theora or probably even VP8. The problem is precisely the small devices, from Netbooks though handhelds. On a desktop these days, you have a combination of hardware and software in the GPU (via accelerations interfaces like DXVA in Windows) to speed up video. But on a handheld, what you'll probably find are a bunch of hardware units called up by a conventional CPU or perhaps a DSP, but the hardware units tend to be specific to H.264 and very similar video schemes. And also, perhaps not well described in public docs, so even your ability to run other CODECs is poor, at least without broad industry support making it a priority.
Transcoding on YouTube would not be a huge issue. For one, it doesn't have to happen overnight, and for two, there could be a prize at the end. But not for Theora. Theora is a lower efficiency CODEC than H.264, and that's obvious even when you look over the samples that Xiph.org keeps touting... and that's about the best example you'll find anywhere, simply because it's the best rigged demo they could produce without telling outright lies.
The big expense of YouTube isn't video storage or transcoding cycles, it's bandwidth... all the data being schlepped in and out of the site every day. Going to a less efficient CODEC like Theora will result in a lower quality video or much more expense. And given YouTube's continued lack of profitability, don't think Google's going to budget on this for a femtosecond. But reportedly, VP8 offers higher efficiency than H.264, particularly at lower bitrates. Moving to VP8 might save Google serious green, over the years. Enough to say, buy On2? Well, hey, they did... they must have some plan for it.
With the H.264 royalties not set to kick in until 2015 now, there's plenty of time for a transition. If Google open sourced VP8 tomorrow and did the free license thing on any patents On2 might have on the technology, this would be a very good thing, even if it's not yet a perfect solution. They don't have to back-transcode everything right away, but they could start encoding new submissions in H.264 for "device" use and VP8 for the desktop (where the acceleration is not AS needed, and easier to re-program given that much of it's running on stream processors rather than dedicated hardware). If VP8 became THE long term standard for the internet, devices would start supporting it.
Also, keep in mind that many of the cable companies run really fast fiber to the neighborhood node. So you're not really dealing with coax over crazy distances. I don't think anyone's running new copper for this kind of thing, it's only the last mile or less issue, whether you fiber, coax, or well, nothing, to the home.
You must be new here.
The Constitution grants the Federal Government the right to pass laws to deal with some things not specifically addressed in the Constitution, and the States rights to deal with others.
Given that radio waves, much less fiber optic internet, had not yet been discovered in 1787, this is a very clear case in which one needs not simply heed the Constitution, but all of the law built on top of it since.
You may now return to drinking that teabagger kool-aid.
Yeah, the "no more copper" issue was pretty much the death DSL expansion. It's a product of competiton... the cable companies were shooting at the telco's main profit... once you have cable TV and decent broadband, taking on voice service is a small deal. So the telcos had to respond by offering the same video, internet, and voice bundle. You don't get that over copper.
It is actually pretty reasonable, if you're talking about peak service. I mean, this is already commonplace in South Korea. Hanaro Telecom rolled it out and in six months, had 200,000 customers on DOCSIS 3.0 at 100Mb/s... back in 2007. Today, most of the country is on 100Mb/s fiber. This all started with a mandate back in 2003.
This year, the UK's already taking about a similar program, to wire the whole country at 100Mb/s by 2017. The USA simply has to say something, or we're going to look more like the hillbillies of internet connectivity than we already do.
Mostly, the telephone and power lines are there because the telcos and power companies got breaks in return for universal coverage. So AT&T got to be a monopoly basically as long as they wanted to be one. They actually subsidized everyone's local service by overcharging for other things, like long distance, some kind of business services, etc. The actual cost over time of the additional coverage was very small, given that they got the whole pie to themselves.
Yup.. the market is currently free enough to ensure I get no possible solution for wired broadband where I live. A totally free market will offer a half dozen solutions to the most populated areas, and none to the less populated areas. After all, winning 25% of a city of 100,000 is far more profitable than winning 100% of a city of 1,000.
And like the man says, that's how power and telephone worked. I'm quite sure I'd be off-grid entirely, if telcos and power carriers had not been forced to hook me up. Or, at the minimum, I would have had to pay for the power run, transformer, etc.
That's an easy one to solve... I'd deal with that in a heatbeat. Find a reasonable neighbor, offer to pay for their interent access if they'll let you set up a wireless link. Plain old 802.11g with a couple of Yagi or "coffee can" directional antennas, and you're good for hundreds of feet. Better with 802.11n, but only if you're wiring for MIMO (2 or 3 antennas at either end, and issues with where they're placed if you're optimizing it).
I actually design radios in my day job, and one such device is a mesh router that can run up to about six miles. I've been really tempted to tap real broadband in neighboring towns... the frequencies used, illegal as hell, unless your're police or the military... but tempting anyway.
I totally agree.
Mandating higher speeds isn't even really necessary. Most places with any kind of decent internet service usually have competition, like FiOS or DSL vs. Cable, whatever. Natural competition has kept those guys one-upping each other for more than a decade already, and that's not going to stop. Accelerating that will just offer them an excuse for not expanding coverage out to those of us living in not-so-highly-profitable areas. Hell, at this point, I'd happily pony up $120 a month for FiOS... I'm paying that now for relatively crappy satellite.
Fiber to the house, and you're "out in sticks". Are you sure? Can you see any neighbors from any point on the ground around your house?
I'm jealous.. I'm out in the sticks. Or maybe it's the boondocks, hard to tell. But I'm not talkin' Kansas or Idaho here, this is South Jersey. There's a town 3 miles West, another 5 miles to the East... not exactly the mountains of New Mexico.
I can't wired-anything to my house, unless you count POTS. POTS works pretty well here for POTS, largely because there's a DSL-capable local node across the street... I can see it from my mailbox. But no telco is willing to drop as couple of DSL boards in there (yes, it's compatible with DSL). So I'm paying $120 a month for 1.5Mb satellite, with a 500MB per day download cap.
Meanwhile, the folks who had 784kb/s DSL, then 3Mb/s DSL, then 5Mb/s cable, then 12Mb/s cable, are now getting 25Mb/s FiOS upgrades.
But my forest IS really nice in the warmer months...
Same is true in South Jersey, where we did get a whole lot of snow this year.
It's kind of obvious that global warming will lead initially to more snow. That's independent of the origins of this year's snows.. whether global warming was involved or not. The colder you get, the less water vapor in the air. More heat means more energy in the system which means more potential snow. Based on the studies of average temperature, the global average from the 1850s throught to 2005 is 1.36F. That's not the difference between snow and now snow, except in rare conditions. On the other hand, a fraction of that warming the tropical Pacific can cause an El Niño episode, which has a very dramatic effect on weather for the whole USA. More energy in the system leads to more variation and more severe weather, not anything you'd easily call out as "warming".
And just a little extra energy alone isn't the reason we went from a usual 10-15" of snow for a whole season to over 70"... so far. The weather is definitely weird this year. We usually get a cold snap in late December maybe through mid January, with temps down to the 10F's or so... this year, not so much. It's hovered in the upper 20Fs and lower 30Fs. But the storms! The three huge blizzards were due to nor'easters... big circulating storms, kind of like a small hurricane, that basically hover around the coast and bring all kind of precipitation inland a bit. That's why, for the first time ever in my recollection, we got much more snow than most of Eastern PA. Thing is, a nor'easter is usually a fall and spring storm, not so much a winter thing. Is that the end result of more energy in the system?
I'm sure people who actually know this stuff, actual scientists, not just we computer geeks, have their opinions. "Near-normal amounts of precipitation are expected over the eastern third of the country, as well as over the Pacific Northwest and Northern Plains, while drier-than-normal conditions are forecast to occur over the Southwest and the Upper Midwest/Great Lakes."... wait, that was the Farmer's Almanac, for the 2009-2010 winter season. So here's an actual climate scientist: http://www.democracynow.org/2010/2/12/climate_scientist_record_setting_mid_atlantic
And hey, why trust me... I'm just a computer engineer with a guitar. But man, if you can't trust Bill Nye the Science Guy, there's just something wrong about you. Here's Bill, speaking about the snow storms, and questioning the patriotism of the "science deniers": http://www.youtube.com/watch?v=sm05Mcah0i8
The point is, unless you have the "world is black and white" vision of a radical right teabagger, or maybe you've lived in California your whole life and have never seen snow beyond "the ski slopes", you understand that snow != cold. Sure, you need a little cold to get snow, but if you have lots of cold, no snow. Sure, some northern climes look very snowy this time of year, but that's the effect of many tiny bits of snow without a melt. Here, we usually melt out a few days after any snowstorm... this year, with 30" in the same week, it's taking JUST a bit longer. Up at my brother-in-law's place in up-state New York, the snow they get in December is probably going to hang around until March. So they don't have much falling at any given time, but it accumulates.
Simple enough, when you actually live it every year.
Actually, it's the cows. Over 50% of all the corn we grow freed cows, chickens, and other livestock. Before anyone can complain about ethanol taking food away from people (regardless of the relative efficiency of corn sugar to ethanol... about as good as oil sands to petrol), you had better shut down the beef and poultry industries.
Microsoft did just what Palm did.. they added phone APIs to an existing PDA OS. Doesn't matter how long you've been at it... it matters how well you do it.
Of course, this is only temporary... if you want to support iPhone and iPad, you'll have to worry about resolution. You already have to worry about performance, given that the iPhone 3GS has about twice the CPU and 2x-4x the GPU of previous iPhones. And if Apple's planning any followup to the iPhone 3GS, they're going to need a nominally 800x480 screen, just like all of the current higher-end smart phones today.
So it's an issue... Apple's just putting it off, and regular users aren't thinking about it.
And of course, the other view: these are the one or two models you get to chose. They are perfect for every single human being on the planet, because Steve says they are. Sorry... that's a fail. Which is why others are leading the future of the smartphone now, just as Apple's played catch-up on the PC front.
It's not that Apple didn't aid Adobe in porting Flash to the iPhone, it's that they proactively prohibited Adobe from releasing Flash for the iPhone. Big difference.
But yeah, Microsoft is working with Adobe toward a Flash release. Adobe already officially announced Flash and AIR for Android. I don't they'd have any problem putting it on Palm's WebOS devices, either. And it's already on some Nokia devices. So like Flash or not, Apple will be the only mobile device voluntarily setting itself up as a lesser web experience.
Because Jobs doesn't like Flash, doesn't like multitasking, and wants to ensure Apple has complete control of the iPhone applications world. There are no technical reasons to not have Flash on your device.
App-switching, at least old PalmOS style (eg, you come back to exactly where you left off) works about 50% of the time. Of course, it worked well on a Palm because the PalmOS put everything in RAM, and just left it there. Having to reload every app, that's going to slow things down.
When you have real multitaking, you get every advantage of task switching, and no disadvantages. If your app is waiting on user I/O, and you lose focus, it's there in memory (unless it needs to be kicked out for space reasons), but it's not consuming cycles. So it's functionally identical to app-switching. But when you need a background process, that works, too. Apple has these on the iPhoneOS, they just don't allow anyone but Apple to write them.
Here are some more... so I can run a locale aware daemon, to turn off my Wifi when I'm out of the house or office, swap out the "office inappropriate" background when I'm at the office, kick on Bluetooth when I'm in the car, etc. Not necessary, but it saves on battery power.
Another GPS daemon... I can track where I go. Run this, and it keeps a record of my travels, no disruption of any kind to other uses of the phone/computer. This is extremely useful for geotagging photos and video shot with my various GPS-lacking cameras.
How about listening to any media player in the "background", not just the one my phone manufacturer deems suitable for that task (I like Pandora for web audio, Museek for regular MP3 player functionality). Maybe I'm in the car, and want to run Navigator at the same time... no problems.
How about jumping between the web, the message I'm composing, the calculator I brought up to make a few calculations, and my project notes for all the various projects I'm working on, all without having to stop the Mahjong game I was playing before I got that email (also read without stopping).
In short... ever use a real computer? It's just like that.
It's interesting how you can slice data and get the results you want. If you compare Apple to Apple, year to year, they look great in 4Q09. But if you compare Apple 3Q09 to Apple 4Q09, Apple actually lost market share (18% down to 16%), even though that was their best quarter ever. They did not keep up with the growth of the smartphone market, quarter to quarter.
http://blogs.wsj.com/digits/2010/02/01/iphone-loses-market-share-in-fourth-quarter/tab/article/
That's more interesting than looking year to year, as it reflects the current way Apple is trending. Given their three-something years in the market, one whole year is just too long to spot a trend.
Actually, Nokia had been slipping for awhile, but seems to have bounced back a bit in 4Q09. They were managing about 35% of the global smart phone market in 2009, but in the last quarter, they boosted their share to 40%. Great, but not the 50% so many often claim. http://www.ft.com/cms/s/0/5af7cd20-0c05-11df-96b9-00144feabdc0.html?nclick_check=1
Keep in mind that the smartphone market itself is actually growing pretty fast, too.