Re:Windows mirrors linux mirrors windows.
on
Fedora 13 Is Out
·
· Score: 3, Interesting
Your comment isn't talking about Windows and Linux at all. You are almost entirely talking about Explorer and GNOME/KDE. There are a few underlying system services, like CUPS, grub, kudzu, etc., but you're mostly talking about UI.
If that's all you care about, good for you. The rest of us want much more from an operating system than a vaguely familiar interface. The more extensively you actually use it, the more apparent the differences become.
Now, explain how a few milliliters of ink can cost as much as a fucking 1L bottle full of it?
Easy... Inkjet cartridges (from everyone but Epson) have the print heads and spraying mechanism built-in. Your 1L bottle does not.
Now, you might get lucky and be able to refill one of those cartridges a few times, but if you print low volumes, and/or you live in a harsh climate, you might find the heads clog up about as soon as you run out of ink.
Now, it's more or less a sleazy trick that manufacturers are selling smaller and smaller cartridges, ensuring higher prices per litre, but the fixed cost is there.
Time Warner, at least here in north central NC, has been making a concerted effort for the last several years, and actually has pretty darn good service. Their broadband is almost never down. They almost always show up when they say they will, you can get someone on the phone typically within 5 minutes, and the people on their phone support seem to actually know what they are talking about.
Exactly the same experience with Time Warner here in Southern California.
I fought with Verizon for over a month, trying to get DSL working... No good. Have to call them between 9am and 4pm to get an ISP representative, rather than auto-transfer to some billing moron who will talk to you for 15 minutes, before mentioning you've got the wrong department, they can't possibly help you, and you need to hang up and dial the exact same number you called to get the first idiot. And god help you if you do get through to a rep after a half hour, because they hardly understand what a phone line is, and try desperately to get you to hang-up with some nonsense, without helping in any way. The number of times I've heard "it'll be fixed tomorrow" is astounding. Even after insisting on getting a tech to come out a couple time, nobody ever showed. No calls, no note, nothing.
Meanwhile, about 5 minute hold time with Time Warner, a rep that offered the best deal they had up-front, and had two installers come out the next damn day. Normally 33% more expensive than Verizon DSL, but also 50% faster, and unlike Verizon, I've had all of 1 hour of downtime in the past year+.
The early nuclear experiments with aircraft were just that. The technology has advanced quite a bit. If nothing else, using an Alpha or Beta emitter would instantly solve the shielding problem.
Does anyone else think it is odd that the fastest plane in the world is still the SR-71, which came into service in 1964.
Only as odd as the fact that the fastest STEAM SHIP was built even longer ago...
High speed reconnaissance aircraft are simply no longer needed in the age of satellites. High-speed attack aircraft also aren't terribly important in an age of ICBMs, stealth and UAVs
Still, I'd be surprised if the US doesn't have an even faster aircraft we've never heard about, somewhere in the black projects.
While I fully agree with your comment, there's one important detail you are omitting... There are extreme differences between high bitrate codecs, and low bitrate codecs.
MPEG-2 is undeniably the king of high bitrate lossy video, and even H.264 can barely improve upon it. MPEG-1 Layer II audio is also the highest-quality common audio codec codec out there, but only when used at higher bitrates.
But at low bitrates, the lack of an in-loop deblocking filter absolutely kills MPEG-2's video quality. While temporal domain audio codecs like MPEG-1 Layer 2 are great when used with a high enough bitrate, frequency domain codecs like AAC can undeniably outperform them, and other tricks like SBR used in AAC+ provide great improvements in audible quality as well.
With something like web video, low bitrates are king.
Google clearly isn't afraid of patent trolls - whether because they believe there wouldn't be any, or because they think they can buy out any threats.
Sorry, no. Option C is still very possible, and pretty likely.
Then I don't understaThen I don't understand the point of your original post to which I replied.nd the point of your original post to which I replied.
You've been talking to several different people in this thread. My first comment was a reply to you...
I've already made my point as simple and straight-forward as I can imagine being possible:
[Google] believe[s] any lawsuits (over VP8 encoder patent infringement) will cost less than continuing to license H.264, VP6, H.263, MP3, AAC and Flash.
That doesn't mean there aren't any patents. That doesn't mean they will be able to buy any patents. It means they will pay perpetual license fees to which-ever patent troll(s) happen(s) to win in court.
if you can not can not provide the rights which come with the GPL then the GPL does not grant you the right to distribute.
The GPLv2 has no clause that would prevent distribution just because patents happen to apply. That (and digitally signed binaries ala Tivo) is the very reason why GPLv3 was created.
What matters to lawyers and judges are not concepts like "justice", "equity" or "reason". What matters to them is the written rule of the law,
The laws say that's what they are supposed to do. In the US, they are required to follow the law, except in extreme cases, where it is in conflict with the constitution.
What you're spewing isn't an insult of judges, you're condemning the role they fill, which we the people said was necessary.
Frankly, if it were up to judges to decide on justice and equality, you'd be making them emperors rather than judges, with the ability to ignore any laws they don't like, and punish those who have broken no written law.
As to the assertion that they twist the laws "into ways that benefit both the priesthood and its patrons--the wealthy and powerful," you're going to need to cite several cases to make such a claim. It runs DIRECTLY CONTRARY to your previous statement that judges care only about the letter of the law, and lawyers' arguments.
it is reasonable to assume that they would just add those patents to the ones they already hold with regard to VP8, and grant similar permission to use them freely.
Patent trolls don't ask if you'd like to buy the patent, instead of being sued. Google any anyone else can't force someone to sell their patent and/or stop suing people for money. And even if they could, it's very likely any patent troll would want an astronomical amount of money.
Honestly, Google didn't buy all (or any of...) the H.264 patents, there's no reason to believe them any more capable of buying all submarine VP8 patents.
Well, Google certainly seems to be willing to risk it as they're moving YouTube to it.
Hate to burst your bubble, but Google has money. It doesn't have to be 100% free or nothing at all... For them, it's a simple risk/reward thing.
They are going to adopt VP8 because they believe any lawsuits (over VP8 encoder patent infringement) will cost less than continuing to license H.264, VP6, H.263, MP3, AAC and Flash.
They certainly are. I've put together plenty myself, and heard from several others who've done the same. A quick search would find plenty of mailing-list posts from individuals (public mailing lists are often utilized by smaller companies with minimal IT staff in-house).
Honestly, how many hard drives can you fit in a 2U chassis? If you need more than 2TBs of space, that's 4 drives (or it was the minimum until very recently). An non-redundant boot drive isn't an option, so you need at least 2 more drives in the system. Do you have enough power, plugs, SATA ports, and hot-swap drive bays in your 2U servers for 6 hard drives? And if you do (say, in a 3U server), why aren't you using 6 slightly smaller (cheaper) hard drives, or increasing your storage capacity another 50%?
Sure, if you're lucky, and you've got a high-end SATA RAID card, you could define two logical volumes, but again, you're wasting a lot of money just so you can continue to use all your legacy tools, and don't have to handle that 2TB limit (which really only requires a different partition utility than fdisk, and a bootloader change)...
Telling your boss it's too hard to boot-up off of a single logical device larger than 2TBs is a sure way to put yourself out of work... But, of course, you probably don't work in the field.
While an excellent analysis, it unfortunately confirms all the worst fears I've had about VP8:
It was excellent, but I would call it rather pessimistic. Did you get a good look at the stills? It looks better than everything EXCEPT x264. Better than VC-1, far better than Xvid, and vastly better than the latest version of Theora.
As for x264, even baseline looks vastly better, which really doesn't make much sense. It sounds like VP8 was heavily tuned for low bitrates, and simply doesn't have the option to weaken the deblocking filter, and up the sharpness. It could be a rather easy change to get VP8 looking quite a bit better. Additionally, I've always been irritated at x264 adding a strangely distracting level of random noise to the picture, so I'm not even sure the added sharpness is entirely real.
I know, better than most anybody, that still pictures don't reflect the real quality of a video codec (see ratecontrol). But they're at least in the right ballpark to get me very excited to get a hold of it and start hammering away at it.
The quality doesn't match up to H.264, it despite that also can't even match it in speed,
The decoder isn't drastically slower than ffmpeg's H.264. As for the encoded, I think most of us would reluctantly accept a painfully slow encoder for just a slight improvement in quality, so I don't expect that to be a problem. If anything, x264 show demonstrate that open sourcing it is the quickest way to get someone to tweak the speed.
the spec is apparently an unholy abomination, the implementation needs work,
Specs can be written by 3rd parties... That's how it went with VP3. Yes, it's unfortunate, but it's a problem that can be resolved by the community given a bit of time. The implementation is somewhat backhandedly complimented as not needing remotely as much work as VP3. Yeah, it needs work, but anyone that has ever seen proprietary code would probably say that about everything out there.
and most disappointingly of all, it appears there is serious risk of patent issues (largely due to blatantly ripping off various parts of H.264).
The reviewer is not a lawyer. On2 had several lawyers... On2 has, time and time again, back to VP3, made codecs that sound like patented ones, with just barely enough changes to circumvent patents. That's where all the weird properties of Theora come from. If On2 was smart enough to copy patented techniques with just enough changes to get away with it in the VP3 days, I see no reason to believe they didn't do just as good of a job now.
Actually it's a shame that sector 0 of an GPT disk doesn't contain code to load a boot manager that understands GPT to allow booting from a GPT disk with an old fashioned Bios.
They do just as soon as you install Grub on it (either version 2 or a patched older version).
EFI is uncommon, but raid arrays with 4 or more TB+ HDDs are VERY COMMON. What do you think everyone's been doing with those TB+ enterprise HDDs for the past couple years?
Yes, it is "exactly accurate". It's just misleading...
In countries without software patents, you can implement an H.264 decoder that runs on your PC, free of charge, no problem. HOWEVER, as soon as that code gets flashed into the firmware of a device, it's now a hardware patent, and you're either paying the patent license fees, or having your devices seized by authorities, in just about any country in the world...
The call for resistance to H.264 will make a great deal of sense if and when there is a reasonably reliable basis on which it can be assumed that a format such as Theora and/or VP8 doesn't infringe patents.
No clue about VP8, but as for Theora, On2 specifically went out of their way to avoid patented tech at the time. It's now old enough that patent expiration may have already voided any patent which DID apply, if any ever did. It's been sold commercial by On2 for years. Xiph did their patent research an gave it a clean bill of health. AOL uses VP3 extensively, and paid for VP5/6 (IIRC) licenses, which they also use extensively. Damn near everybody uses/has used VP6 thanks to Adobe including that as the latest greatest codec in Flash 8. etc.
In short, if any patents ARE found, the whole planet is already massively screwed...
This is really just AMD getting beaten to the punch again, and having to try to spin it in some positive way.
I'll have to call you an idiot for falling for Intel's marketing, and believe that, just because they can legally call it by the same name, it remotely resembles what AMD is doing.
I guess that beauty isn't really in the eye of the beholder anymore.
It never was. You're buying into a myth, and claiming we shouldn't do anything to up-end that same myth.
Scientists have known for many years that beauty is directly related to body proportions, and symmetry. While we get examples of overweight people, or those with massive rings in their lips being attractive, that is really an example of wealth being attractive, and unrelated to beauty.
Imagine, for a moment if how someone might feel if their body shape had been *scientifically proven* unattractive.
I fail to see why some teen-aged girl is going to logically evaluate a scientific study, and then have some extremely emotional reaction, stronger than she would about her peers laughing and whispering about her... Your premise makes no sense, and your political correctness/think of the children circuit has gone haywire.
Theoretically, 10 Gb is exactly 1.25 GB, but then you need to account for protocol overhead, packet loss and so on.
You don't have to use TCP/IP over ethernet, you know. AoE & FCoE come to mind.
There are very few ways you might lose packets in a well-built local data-link network. Collisions/congestion are a thing of the past. Modern networking gear is fully capable of forwarding packets at full speed. Packets don't just go flying out of CAT-6A, never to be seen again.
And yes, bonding two 10Gbps ethernet interfaces would give you greater than 1.25GByte/s, though certainly NOT double that...
The question isn't how Flash did perform, it is how it does.
It performs HORRIBLE, unless you're using the very latest version, playing back H.264 video, on a recent Windows system. Otherwise, it's a dog.
but then I've seen nothing else that does what it does but lighter weight, except maybe Silverlight.
Your ignorance, willful or otherwise, is no excuse, and certainly won't change facts.
Everyone seems to assume HTML5 will be a magic fix, that Flash is slow from sucking and not because what it is doing requires a lot of calculations.
HTML5 most certainly WILL be a magic fix, because once it's not locked-up in a proprietary format nobody can read, open source players will spring-up like crazy, with all the features, and hardware acceleration you could want. The fact that the very first few bits of software don't perform amazingly goes without saying. That you expect they should is moronic.
While people love to hate on Flash, it actually performs quite will for video on most systems. It can chat with the video card and use it to accelerate decoding.
Flash is a horrible dog. It's only VERY, VERY recently started performing well on Windows, because they gave up decoding video, and handed it off to the OS, which can accelerate it via the videocard. For fuck's sake, Flash didn't even do the most basic hardware overlay until maybe a year ago, available in the mid 90s on damn near every video card, and standalone video player.
And while Flash may have gotten lucky on Windows, no such luck on other platforms. Flash on Linux is as big of a dog now as it ever was. Jumpy, flickery, tearing mess. And don't claim they can't do better, VADPU support on Linux has been in MPlayer for many months. Besides, I shouldn't need hardware acceleration just so stupid 480kbps 400x300 Hulu videos don't bring my 2GHz+ CPU to a grinding halt...
Another case in point is determining how long to buffer and what bitrate to use (change dynamically). Does HTML5 video offer these options?
The video tag is just that, a video tag. HTML5 however, happens to contain, wait for it..., JAVASCRIPT, which could quite easily be used to download a file of a given size, time how long it took, and then dynamically insert the VIDEO element into the page, based on the previously estimated bandwidth.
my great grandmother used to tell the story of the time ball lightning broke the living room window, did a circle around the room and went back out, leaving scorch marks on the ceiling.
Slashdot Mirror
Your comment isn't talking about Windows and Linux at all. You are almost entirely talking about Explorer and GNOME/KDE. There are a few underlying system services, like CUPS, grub, kudzu, etc., but you're mostly talking about UI.
If that's all you care about, good for you. The rest of us want much more from an operating system than a vaguely familiar interface. The more extensively you actually use it, the more apparent the differences become.
Easy... Inkjet cartridges (from everyone but Epson) have the print heads and spraying mechanism built-in. Your 1L bottle does not.
Now, you might get lucky and be able to refill one of those cartridges a few times, but if you print low volumes, and/or you live in a harsh climate, you might find the heads clog up about as soon as you run out of ink.
Now, it's more or less a sleazy trick that manufacturers are selling smaller and smaller cartridges, ensuring higher prices per litre, but the fixed cost is there.
Exactly the same experience with Time Warner here in Southern California.
I fought with Verizon for over a month, trying to get DSL working... No good. Have to call them between 9am and 4pm to get an ISP representative, rather than auto-transfer to some billing moron who will talk to you for 15 minutes, before mentioning you've got the wrong department, they can't possibly help you, and you need to hang up and dial the exact same number you called to get the first idiot. And god help you if you do get through to a rep after a half hour, because they hardly understand what a phone line is, and try desperately to get you to hang-up with some nonsense, without helping in any way. The number of times I've heard "it'll be fixed tomorrow" is astounding. Even after insisting on getting a tech to come out a couple time, nobody ever showed. No calls, no note, nothing.
Meanwhile, about 5 minute hold time with Time Warner, a rep that offered the best deal they had up-front, and had two installers come out the next damn day. Normally 33% more expensive than Verizon DSL, but also 50% faster, and unlike Verizon, I've had all of 1 hour of downtime in the past year+.
The early nuclear experiments with aircraft were just that. The technology has advanced quite a bit. If nothing else, using an Alpha or Beta emitter would instantly solve the shielding problem.
Only as odd as the fact that the fastest STEAM SHIP was built even longer ago...
High speed reconnaissance aircraft are simply no longer needed in the age of satellites. High-speed attack aircraft also aren't terribly important in an age of ICBMs, stealth and UAVs
Still, I'd be surprised if the US doesn't have an even faster aircraft we've never heard about, somewhere in the black projects.
While I fully agree with your comment, there's one important detail you are omitting... There are extreme differences between high bitrate codecs, and low bitrate codecs.
MPEG-2 is undeniably the king of high bitrate lossy video, and even H.264 can barely improve upon it. MPEG-1 Layer II audio is also the highest-quality common audio codec codec out there, but only when used at higher bitrates.
But at low bitrates, the lack of an in-loop deblocking filter absolutely kills MPEG-2's video quality. While temporal domain audio codecs like MPEG-1 Layer 2 are great when used with a high enough bitrate, frequency domain codecs like AAC can undeniably outperform them, and other tricks like SBR used in AAC+ provide great improvements in audible quality as well.
With something like web video, low bitrates are king.
Sorry, no. Option C is still very possible, and pretty likely.
You've been talking to several different people in this thread. My first comment was a reply to you...
I've already made my point as simple and straight-forward as I can imagine being possible:
[Google] believe[s] any lawsuits (over VP8 encoder patent infringement) will cost less than continuing to license H.264, VP6, H.263, MP3, AAC and Flash.
That doesn't mean there aren't any patents. That doesn't mean they will be able to buy any patents. It means they will pay perpetual license fees to which-ever patent troll(s) happen(s) to win in court.
The GPLv2 has no clause that would prevent distribution just because patents happen to apply. That (and digitally signed binaries ala Tivo) is the very reason why GPLv3 was created.
The laws say that's what they are supposed to do. In the US, they are required to follow the law, except in extreme cases, where it is in conflict with the constitution.
What you're spewing isn't an insult of judges, you're condemning the role they fill, which we the people said was necessary.
Frankly, if it were up to judges to decide on justice and equality, you'd be making them emperors rather than judges, with the ability to ignore any laws they don't like, and punish those who have broken no written law.
As to the assertion that they twist the laws "into ways that benefit both the priesthood and its patrons--the wealthy and powerful," you're going to need to cite several cases to make such a claim. It runs DIRECTLY CONTRARY to your previous statement that judges care only about the letter of the law, and lawyers' arguments.
Patent trolls don't ask if you'd like to buy the patent, instead of being sued. Google any anyone else can't force someone to sell their patent and/or stop suing people for money. And even if they could, it's very likely any patent troll would want an astronomical amount of money.
Honestly, Google didn't buy all (or any of...) the H.264 patents, there's no reason to believe them any more capable of buying all submarine VP8 patents.
Hate to burst your bubble, but Google has money. It doesn't have to be 100% free or nothing at all... For them, it's a simple risk/reward thing.
They are going to adopt VP8 because they believe any lawsuits (over VP8 encoder patent infringement) will cost less than continuing to license H.264, VP6, H.263, MP3, AAC and Flash.
They certainly are. I've put together plenty myself, and heard from several others who've done the same. A quick search would find plenty of mailing-list posts from individuals (public mailing lists are often utilized by smaller companies with minimal IT staff in-house).
Honestly, how many hard drives can you fit in a 2U chassis? If you need more than 2TBs of space, that's 4 drives (or it was the minimum until very recently). An non-redundant boot drive isn't an option, so you need at least 2 more drives in the system. Do you have enough power, plugs, SATA ports, and hot-swap drive bays in your 2U servers for 6 hard drives? And if you do (say, in a 3U server), why aren't you using 6 slightly smaller (cheaper) hard drives, or increasing your storage capacity another 50%?
Sure, if you're lucky, and you've got a high-end SATA RAID card, you could define two logical volumes, but again, you're wasting a lot of money just so you can continue to use all your legacy tools, and don't have to handle that 2TB limit (which really only requires a different partition utility than fdisk, and a bootloader change)...
Telling your boss it's too hard to boot-up off of a single logical device larger than 2TBs is a sure way to put yourself out of work... But, of course, you probably don't work in the field.
It was excellent, but I would call it rather pessimistic. Did you get a good look at the stills? It looks better than everything EXCEPT x264. Better than VC-1, far better than Xvid, and vastly better than the latest version of Theora.
As for x264, even baseline looks vastly better, which really doesn't make much sense. It sounds like VP8 was heavily tuned for low bitrates, and simply doesn't have the option to weaken the deblocking filter, and up the sharpness. It could be a rather easy change to get VP8 looking quite a bit better. Additionally, I've always been irritated at x264 adding a strangely distracting level of random noise to the picture, so I'm not even sure the added sharpness is entirely real.
I know, better than most anybody, that still pictures don't reflect the real quality of a video codec (see ratecontrol). But they're at least in the right ballpark to get me very excited to get a hold of it and start hammering away at it.
The decoder isn't drastically slower than ffmpeg's H.264. As for the encoded, I think most of us would reluctantly accept a painfully slow encoder for just a slight improvement in quality, so I don't expect that to be a problem. If anything, x264 show demonstrate that open sourcing it is the quickest way to get someone to tweak the speed.
Specs can be written by 3rd parties... That's how it went with VP3. Yes, it's unfortunate, but it's a problem that can be resolved by the community given a bit of time. The implementation is somewhat backhandedly complimented as not needing remotely as much work as VP3. Yeah, it needs work, but anyone that has ever seen proprietary code would probably say that about everything out there.
The reviewer is not a lawyer. On2 had several lawyers... On2 has, time and time again, back to VP3, made codecs that sound like patented ones, with just barely enough changes to circumvent patents. That's where all the weird properties of Theora come from. If On2 was smart enough to copy patented techniques with just enough changes to get away with it in the VP3 days, I see no reason to believe they didn't do just as good of a job now.
They do just as soon as you install Grub on it (either version 2 or a patched older version).
EFI is uncommon, but raid arrays with 4 or more TB+ HDDs are VERY COMMON. What do you think everyone's been doing with those TB+ enterprise HDDs for the past couple years?
Yes, it is "exactly accurate". It's just misleading...
In countries without software patents, you can implement an H.264 decoder that runs on your PC, free of charge, no problem. HOWEVER, as soon as that code gets flashed into the firmware of a device, it's now a hardware patent, and you're either paying the patent license fees, or having your devices seized by authorities, in just about any country in the world...
See:
http://news.bbc.co.uk/2/hi/technology/5312696.stm
http://www.reghardware.co.uk/2006/12/15/german_court_rules_on_sansa_snatch/
No clue about VP8, but as for Theora, On2 specifically went out of their way to avoid patented tech at the time. It's now old enough that patent expiration may have already voided any patent which DID apply, if any ever did. It's been sold commercial by On2 for years. Xiph did their patent research an gave it a clean bill of health. AOL uses VP3 extensively, and paid for VP5/6 (IIRC) licenses, which they also use extensively. Damn near everybody uses/has used VP6 thanks to Adobe including that as the latest greatest codec in Flash 8. etc.
In short, if any patents ARE found, the whole planet is already massively screwed...
And yet, many Linux systems are capable of playing DVDs, even though MPEG-2 is just as patented as H.264.
I'll have to call you an idiot for falling for Intel's marketing, and believe that, just because they can legally call it by the same name, it remotely resembles what AMD is doing.
It never was. You're buying into a myth, and claiming we shouldn't do anything to up-end that same myth.
Scientists have known for many years that beauty is directly related to body proportions, and symmetry. While we get examples of overweight people, or those with massive rings in their lips being attractive, that is really an example of wealth being attractive, and unrelated to beauty.
I fail to see why some teen-aged girl is going to logically evaluate a scientific study, and then have some extremely emotional reaction, stronger than she would about her peers laughing and whispering about her... Your premise makes no sense, and your political correctness/think of the children circuit has gone haywire.
You don't have to use TCP/IP over ethernet, you know. AoE & FCoE come to mind.
There are very few ways you might lose packets in a well-built local data-link network. Collisions/congestion are a thing of the past. Modern networking gear is fully capable of forwarding packets at full speed. Packets don't just go flying out of CAT-6A, never to be seen again.
And yes, bonding two 10Gbps ethernet interfaces would give you greater than 1.25GByte/s, though certainly NOT double that...
It performs HORRIBLE, unless you're using the very latest version, playing back H.264 video, on a recent Windows system. Otherwise, it's a dog.
Your ignorance, willful or otherwise, is no excuse, and certainly won't change facts.
HTML5 most certainly WILL be a magic fix, because once it's not locked-up in a proprietary format nobody can read, open source players will spring-up like crazy, with all the features, and hardware acceleration you could want. The fact that the very first few bits of software don't perform amazingly goes without saying. That you expect they should is moronic.
Flash is a horrible dog. It's only VERY, VERY recently started performing well on Windows, because they gave up decoding video, and handed it off to the OS, which can accelerate it via the videocard. For fuck's sake, Flash didn't even do the most basic hardware overlay until maybe a year ago, available in the mid 90s on damn near every video card, and standalone video player.
And while Flash may have gotten lucky on Windows, no such luck on other platforms. Flash on Linux is as big of a dog now as it ever was. Jumpy, flickery, tearing mess. And don't claim they can't do better, VADPU support on Linux has been in MPlayer for many months. Besides, I shouldn't need hardware acceleration just so stupid 480kbps 400x300 Hulu videos don't bring my 2GHz+ CPU to a grinding halt...
The video tag is just that, a video tag. HTML5 however, happens to contain, wait for it..., JAVASCRIPT, which could quite easily be used to download a file of a given size, time how long it took, and then dynamically insert the VIDEO element into the page, based on the previously estimated bandwidth.
Your grandmother has a Tyler Durden complex?
That's a bit like saying, the lighter the paint color of the car, the cheaper you can get it.
No, it's not easy to understand why.