Tapes with a proven shelf life of many, many years, or DVDs where a single scratch can render 4GB of data worthless.
Yeah, magnetic tapes NEVER get damaged. EVER!
And with DVDs, a single small scratch makes the entire disc unreadable. After all, the data is stored right at the surface, on the plastic, and scratches can't be repaired, or polished away. And DVDs are seriously damaged by static electricity...
It's a shame nobody has ever come up with some sort of tool that would calculate parity information to allow small damaged portions to be trivially easily be recovered.
-- Face it, tapes are used because of their size, not their reliability. Splitting hundreds of GBs across DVDs would be a tricky task. Possibly a labor-intensive nightmare, if not when backing-up, then at least when restoring.
Nobody has ever thrown away papers that were actually needed...
Paper is an inexpensive and compact way to store terabytes of information...
Paper is trivially easy to instantly duplicate on a large scale...
Paper is trivially easy to haul off-site and store...
People constantly generate diffs between the most recently archived paper copy, and all work they have done every day since. They don't just make undocumented changes, willy-nilly, requiring just as much effort to backup daily changes as it is to backup full copies of everything...
No question, paper is superior. The data retention problems we always hear about are in every way caused by digital storage methods, and have nothing to do with the policies and people running the organizations...
(No I will not pay for any damaged caused by this post overloading your sarcasm meter.)
With hard drives, data doesn't just go away. Sure, it may not be recoverable with simple "undelete" software, but data recovery experts will charge far less than $200,000 to pull important files off of a wiped hard drive.
The same goes for tapes. There is no mention in the article of why they were "unreadable" what level of damage there was to the data, etc.
We all make mistakes, but 3 layers of backup data storage all failing suggests a horrifically poor system in-place. Not JUST "very bad," that's hard to believe, without some massive natural disaster causing it.
Why stop at farm subsidies? Lets get rid of all corporate subsidies
Because all evidence has shown that every $1 the government spends on subsidies translates into $2-3 of additional wealth across the economy.
I think the basis for libertarianism is that there are lots of people who know nothing about economics, and will buy into whatever idiotic system sounds good at first glance.
They were ~$2/bushel, but they have gone up a dollar or more since the bush admin enacted the fucking ethanol mandates.
Corn was so cheap because of the vast subsidies already in place.
I can see more benefits than drawbacks to raising the price of corn. For one, high fructose corn syrup is going to go away. For another, the farmer in the article will have to feed livestock something else, now that corn is going back to a reasonable price.
Ethanol is highly inefficient when mixed with gas, so you lose efficiency in your MPG, so that causes you to buy more fuel, so it is a nasty little cycle.
That's beyond idiotic.
Ethanol doesn't give you as much energy as gasoline in an unmodified engine, but the difference isn't huge, and it sure as hell won't make the 90%+ of gasoline in your tank less efficient. Maybe every 10% of ethanol in fuel only translates into 8% less gasoline consuption, but that's the OPPOSITE of "a nasty little cycle". Quite the opposite really.
Additionally, more ethanol means higher octane, which means newer cars which depend on ethanol in gasoline can be made with higher compression ratios, and get better gas mileage with an ethanol mixture, than they could with pure gasoline.
I also do my best to brush off my shoes and jacket when entering my house.
This is not remotely comparable. I'm sure you don't lock yourself in an air-tight room for an hour, with multi-million dollar cleaning equipment.
You could vacuum it up before turning on the air so that none of the dust will get into the air.. but of course you need air for the vacuum to do anything
I didn't suggest a vacuum, I suggested washing.
Are you saying Moon Dust is too heavy for normal air?
It would be better to simply grant every company ever a patent on everything possi -- oh, wait, that is the USPTO's actual strategy.
Which keeps the big/old players from suing each other (IBM/Microsoft), but allows any one of the big players to crush any smaller, newer players they would simply prefer not to try to compete with.
The price on the patent would be part of a delayed mark-to-market capital gains tax accounting system that would encourage companies to either monetize or sell patents because they would be paying taxes on those patents.
Sounds to me like you're going to be forcing patent trolling...
If someone has a patent on X, and 4 companies need to use X, a law firm buys the patent for 4X as much as any of the individual firms can afford, and proceeds to either raise the price of patent licenses, or otherwise sues the 4 companies for all they're worth.
Frivolous patents on Y (a basic/essential concept) will not be worth anything to any of the companies that use Y. However, the patent Y will be invaluable to a patent-law firm, who can then use it to extort money from each company involved.
etc.
Your idea is a great way to drive patent prices through the roof, based on ridiculous speculation, much like the current stock market.
My house has its own floor, but I still track dirt in, and it travels well enough even with 1 whole g pulling it down.
A) Your house has a floor, but it doesn't have a re-pressurization chamber, where you could wash off and/or filter out any dust/dirt.
B) Dust travels well on earth because of the heavy atmosphere. Fish "travel well enough" in water, but take away their heavy water atmosphere, and suddenly they fall like a rock.
You are expressing opinion though, opinion that it's the US's job to do what it's determined as correct around the world, whether the external interference is desired or not.
No, I'm simply explaining where the military money is going, and why. I never said if I think it's right or wrong, whether there's any way to do things better, pros and cons, etc.
But that shouldn't mean you go 'It's doing a bad job, pull money out of that loss pit', rather it should be a case of 'How can the money better be used, and can extra money help even more?'
We should find out how existing money can be better used, of course. I would NEVER consider giving them more money, though, until AFTER the system is fully fixed. The current amount of money is huge, and for that we're getting almost nothing out of it. We could cut school from 12 to 6 years, and still be educating kids just as well as we currently do.
At that rate, to spend our way out of the problem would require basically spending all tax money in the US on schools, and nothing else.
Education is the foundation stone of EVERYTHING else in a country, it HAS to be if you want a developed country.
Of course it does, but when things don't work, you can't just keep throwing more money at it, and expect it to get better. Public schools should be fixed. However, the problem has been around for a couple decades now, it's only getting worse, and nobody is doing anything about it. In such a situation, I'd rather see money taken out, and used for other things where it might accomplish some good. And if/when the schools are fixed, they'll already have much more money than they can spend, so the cuts really wouldn't hurt.
By diverting funds from the defense budget, to an extent where they probably wouldn't even notice the missing amounts, to education you can make a difference,
You have yet to prove that any amount of the US military budget is in-fact wasted, and/or that any (reasonable) amount of money would improve the school system.
That the military wastes millions of dollars, and schools are underfunded, is pure left-wing propaganda.
No. "Lossless" absolutely DOES NOT imply compression, it merely means a bit-perfect copy, and is used in many places where compression isn't even possible.
If there's no load, it shouldn't be connected to the mains at all.
Obviously, there are problems with that. Most devices use a wall-wart, where there is no way for the device to shut-off/disconnect the transformer. Integrating that transformer into the device instead, would make it more expensive, and both less flexible, and less reliable. I've got many devices that I kept using after the original wall-wart burnt-up. My computer speakers, with their built-in transformer, however, had to be trashed. The small energy savings probably doesn't put you out ahead...
The same has to be said of the mechanical switches needed to shut-off the transformer. They corrode, oxidize, deform, arc, etc. On anything large (like a CRT), I think the higher chances of failure will offset any energy savings.
For many other devices like VCRs, Computers, etc., you need a trivially small amount of current when "off" to keep the clock running, save settings, and being ready to power itself on. If you shut-off the transformer, you're going to have to add a battery and additional circuitry, which will probably waste more energy and money, as well as making the device less reliable.
But there are still many ways in which energy is wasted, as you yourself have just explained rather well.
Home insulation and ducting refrigerators are the only examples I believe are really practical. Even then, it can be very expensive to rip a house apart to improve the insulation, for somewhat small gains.
The rest are either time consuming, don't do the job as well, or otherwise less than ideal.
I believe clothes driers, dish washers, etc. could be made more energy efficient, but not overwhelmingly so. There becomes a point of diminishing returns, where any more energy efficiency becomes very expensive, and possibly energy intensive to produce.
Thinks like video cards for instance waste an incredible amount of energy, and most desktop CPUs do, too. [...] I have to wonder if it's really necessary to waste 400 watts just for writing a letter?
I have to agree about high-end video cards. I'm sure they too will come around, when cooling them becomes just too challenging.
CPUs, however, are in pretty good shape now. SpeedStep/PowerNow/CoolnQuiet/etc have become standard on PCs now, thanks to AMD, so an idle CPU is using less power than ever, despite being able to go up to 2GHz when performance is needed.
Intel, IMHO, is somewhat sabotaging this, though. When they rebrand their Core chips as "Celerons" one of the things they disable is SpeedStep. Add to that the needlessly hot northbridges on all their boards. Buy a new computer today, with an AMD CPU and a Seasonic PSU, and it will probably use half as much power as your old one.
My point is... No doubt there is waste. However, it's not a substantial portion of the overall power use, and it would be either vastly inconvenient (driving a moped in winter) or expensive/unreliable enough that it might eliminate any real benefit to that savings.
Performance doesn't hardly matter. I have one system that does all the video encoding, game playing, etc., and the rest of my machines are just the cheapest hardware I can find on pricewatch when the old one finally breaks...
Mostly, I'd be happiest with the quietest, lowest power equipment I can get. A 25W Turion CPU, 80%+ efficient Seasonic PSU, etc. Not sure about GPUs, but something low powered, but has full-featured open source 3D drivers, and TV-out. Integrated chips need not apply.
Incidentally, this system is actually pretty cheap to put together. Newegg has a 2GHz Turion for $70, and Socket 754 mobos are only $50. Throw in $40 for the PSU, $40 for the GPU, and maybe $40 for RAM. Throw in the largest capacity HDD available, since I'm not paying for it...
The case really seems the most important part of a computer these days. Front ports for lots of frequently used connectors like audio, USB, Firewire, etc. I'd even like front RS-232 and IR. Importantly, good and quiet cooling is almost entirely a function of a well-designed case layout. Not to mention ease of maintenance is almost always hampered by cramped, poorly located components, and the usual jerry-rigged fans to handle hotspots that shouldn't even exist, in cheap/junk cases.
As for my multimedia machine, mostly the same as above, but lowest power (fast) Opteron available, RAID, and importantly, ECC RAM to guarantee stability/integrity.
For a notebook, a more expensive version of the OLPC, redesigned for adults, would be pretty close. Large HDD, sharper screen (in color mode), DVD burner, decent keyboard, etc. .
What I'd really like to see in computers, is the various components working together like never before. How about if the monitor's power button actually just sends a signal to the PC, which can decide whether to shut off the monitor, suspend the system, start an idle timer, etc. It wouldn't hurt to have external speakers receive a signal, telling them to power-off when the system does. And networked information sharing, so I can quickly change the esound host for all the machines at once. More useful LEDs (HDD but no CPU activity LED?). A small, 2-line LCD for status/info output while the monitor is off would be good as well.
robots.txt is meant to be a mutually beneficial standard for the site owner, and the spider operator alike. It started as a method to prevent scanning CGI and other processed or infinite-loop pages from being scanned, and just morphed into a form of restrictions.
If you want enforcement, you password-protect a section of your website, or at least write an.htaccess file that will deny access to known spiders user-agent strings.
There's nothing technical requiring spiders to honor it, presumably there's no legal system to honor it, it's all just trust.
That's how it always works. When any industry is new, and there aren't yet any laws on the books, everyone works together, and defines their own. Most of the time, when it becomes an issue, lawmakers take these accepted practices and make them straight into laws, with few changes.
Before established as a law, it's commonly called "industry standards," which, while not a law, is quite a good argument in court. For instance, if a commercial jet crashes, and it's found that the airline was doing the FAA required maintenance, but not doing the level of maintenance that every other airline is doing, they're likely to be held liable, despite their complete compliance with the applicable laws.
one registrar that charges $35/year for every second-level domain. [...] Shazam! No more domain squatting. It's not longer profitable.
Except for the fact that domain squatting started, and was the most prevalent, back in the old days, when that was EXACTLY the situation. More money means fewer typo-domains, and the like, but squatting any common name/brand URL is guaranteed to return tens of thousands of dollars.
nobody cares whether its on JoeBlow.com or JoeBlow.CheapISP.com.
Except you're now at the mercy of CheapISP, and have no way to transfer to another service provider.
And you know where that's going to lead... Somebody buys com.com, and starts selling names more cheaply, until they get lots of users locked in, then raises their price through the roof.
An open and inexpensive DNS system is simply far better in many ways than any of the alternatives. I could understand making.com expensive, but only if also keeping less desirable TLDs, like.net, very cheap. It's funny how they approve numerous new TLDs, but keep the prices higher than the standard ones, eliminating any possibility of the new TLDs becoming popular/standard/etc., and therefore at all valuable.
And remember that when it comes to registering your domain, you get what you pay for.
Except for the fact that expensive registrars can be incredibly crappy, and cheap registrars can be damn good...
Having to learn new programs (via learning new steps) scares people. It makes them unhappy. And if they've been doing a set of steps for a few years, those steps have become habitual. So you not only have to teach them the new steps, you have to break them of the old ones. Breaking habits is unhappy work.
The problem with that theory, is that Microsoft makes significant changes to the Windows user interface every release. XP is hardly anything like 98.
Applications don't usually change as dramatically, but they certainly do have numerous smaller changes from version to version, and yet people will still upgrade.
People are stupid, but not nearly as mindless as you give them credit for.
The year of Linux arrives the year Adobe ports their software to Linux.
I certainly don't believe 90% of computers run Adobe products (Flash/AcroRead not withstanding). That might make a dent in certain sectors, but not many.
The Unichrome drivers are buggy as hell. No idea about Intel's work, since I don't have an Intel board, and certainly wouldn't buy one just to get a video chipset anyhow. If Intel made PCI Express cards, then maybe.
Additionally, I wouldn't recommend integrated chipsets to anyone. Even on 2D, integrated video will drag your computer down with it. With 3D, you've got to be some sort of masochist.
I'd expect to see better performance from an old ATI Radeon card, with the open source Mesa DRI drivers, than a brand new Intel/Via video chip.
Slashdot Mirror
That's not expensive at all. What makes you think some other storage method would make manpower and storage facilities cheaper?
Yeah, magnetic tapes NEVER get damaged. EVER!
And with DVDs, a single small scratch makes the entire disc unreadable. After all, the data is stored right at the surface, on the plastic, and scratches can't be repaired, or polished away. And DVDs are seriously damaged by static electricity...
It's a shame nobody has ever come up with some sort of tool that would calculate parity information to allow small damaged portions to be trivially easily be recovered.
--
Face it, tapes are used because of their size, not their reliability. Splitting hundreds of GBs across DVDs would be a tricky task. Possibly a labor-intensive nightmare, if not when backing-up, then at least when restoring.
Right, because:
Nobody has ever thrown away papers that were actually needed...
Paper is an inexpensive and compact way to store terabytes of information...
Paper is trivially easy to instantly duplicate on a large scale...
Paper is trivially easy to haul off-site and store...
People constantly generate diffs between the most recently archived paper copy, and all work they have done every day since. They don't just make undocumented changes, willy-nilly, requiring just as much effort to backup daily changes as it is to backup full copies of everything...
No question, paper is superior. The data retention problems we always hear about are in every way caused by digital storage methods, and have nothing to do with the policies and people running the organizations...
(No I will not pay for any damaged caused by this post overloading your sarcasm meter.)
The article says nothing of the sort, and you have provided no sources to back-up your claim.
Yeah, it's far too hard to read the HEADLINE and/or the SECOND SENTENCE of the article, which both explain EXACTLY where the figure comes from...
I'm sure it was much faster to post a comment to
With hard drives, data doesn't just go away. Sure, it may not be recoverable with simple "undelete" software, but data recovery experts will charge far less than $200,000 to pull important files off of a wiped hard drive.
The same goes for tapes. There is no mention in the article of why they were "unreadable" what level of damage there was to the data, etc.
We all make mistakes, but 3 layers of backup data storage all failing suggests a horrifically poor system in-place. Not JUST "very bad," that's hard to believe, without some massive natural disaster causing it.
Deja Vu!
This time it's an OS instead of an Office Suite, and for some reason, overzealous CNet reporters are nowhere to be found...
Because all evidence has shown that every $1 the government spends on subsidies translates into $2-3 of additional wealth across the economy.
I think the basis for libertarianism is that there are lots of people who know nothing about economics, and will buy into whatever idiotic system sounds good at first glance.
Corn was so cheap because of the vast subsidies already in place.
I can see more benefits than drawbacks to raising the price of corn. For one, high fructose corn syrup is going to go away. For another, the farmer in the article will have to feed livestock something else, now that corn is going back to a reasonable price.
That's beyond idiotic.
Ethanol doesn't give you as much energy as gasoline in an unmodified engine, but the difference isn't huge, and it sure as hell won't make the 90%+ of gasoline in your tank less efficient. Maybe every 10% of ethanol in fuel only translates into 8% less gasoline consuption, but that's the OPPOSITE of "a nasty little cycle". Quite the opposite really.
Additionally, more ethanol means higher octane, which means newer cars which depend on ethanol in gasoline can be made with higher compression ratios, and get better gas mileage with an ethanol mixture, than they could with pure gasoline.
The RIAA has very little to do with this. It's ASCAP who collect royalties.
This is not remotely comparable. I'm sure you don't lock yourself in an air-tight room for an hour, with multi-million dollar cleaning equipment.
I didn't suggest a vacuum, I suggested washing.
No, I was referring to dust in a vacuum.
Which keeps the big/old players from suing each other (IBM/Microsoft), but allows any one of the big players to crush any smaller, newer players they would simply prefer not to try to compete with.
Sounds to me like you're going to be forcing patent trolling...
If someone has a patent on X, and 4 companies need to use X, a law firm buys the patent for 4X as much as any of the individual firms can afford, and proceeds to either raise the price of patent licenses, or otherwise sues the 4 companies for all they're worth.
Frivolous patents on Y (a basic/essential concept) will not be worth anything to any of the companies that use Y. However, the patent Y will be invaluable to a patent-law firm, who can then use it to extort money from each company involved.
etc.
Your idea is a great way to drive patent prices through the roof, based on ridiculous speculation, much like the current stock market.
And when they put in the last brick... it all disappears!
A) Your house has a floor, but it doesn't have a re-pressurization chamber, where you could wash off and/or filter out any dust/dirt.
B) Dust travels well on earth because of the heavy atmosphere. Fish "travel well enough" in water, but take away their heavy water atmosphere, and suddenly they fall like a rock.
No, I'm simply explaining where the military money is going, and why. I never said if I think it's right or wrong, whether there's any way to do things better, pros and cons, etc.
We should find out how existing money can be better used, of course. I would NEVER consider giving them more money, though, until AFTER the system is fully fixed. The current amount of money is huge, and for that we're getting almost nothing out of it. We could cut school from 12 to 6 years, and still be educating kids just as well as we currently do.
At that rate, to spend our way out of the problem would require basically spending all tax money in the US on schools, and nothing else.
Of course it does, but when things don't work, you can't just keep throwing more money at it, and expect it to get better. Public schools should be fixed. However, the problem has been around for a couple decades now, it's only getting worse, and nobody is doing anything about it. In such a situation, I'd rather see money taken out, and used for other things where it might accomplish some good. And if/when the schools are fixed, they'll already have much more money than they can spend, so the cuts really wouldn't hurt.
You have yet to prove that any amount of the US military budget is in-fact wasted, and/or that any (reasonable) amount of money would improve the school system.
That the military wastes millions of dollars, and schools are underfunded, is pure left-wing propaganda.
Yes. Any type of modification to the waveform, typically resampling.
No. "Lossless" absolutely DOES NOT imply compression, it merely means a bit-perfect copy, and is used in many places where compression isn't even possible.
Obviously, there are problems with that. Most devices use a wall-wart, where there is no way for the device to shut-off/disconnect the transformer. Integrating that transformer into the device instead, would make it more expensive, and both less flexible, and less reliable. I've got many devices that I kept using after the original wall-wart burnt-up. My computer speakers, with their built-in transformer, however, had to be trashed. The small energy savings probably doesn't put you out ahead...
The same has to be said of the mechanical switches needed to shut-off the transformer. They corrode, oxidize, deform, arc, etc. On anything large (like a CRT), I think the higher chances of failure will offset any energy savings.
For many other devices like VCRs, Computers, etc., you need a trivially small amount of current when "off" to keep the clock running, save settings, and being ready to power itself on. If you shut-off the transformer, you're going to have to add a battery and additional circuitry, which will probably waste more energy and money, as well as making the device less reliable.
Home insulation and ducting refrigerators are the only examples I believe are really practical. Even then, it can be very expensive to rip a house apart to improve the insulation, for somewhat small gains.
The rest are either time consuming, don't do the job as well, or otherwise less than ideal.
I believe clothes driers, dish washers, etc. could be made more energy efficient, but not overwhelmingly so. There becomes a point of diminishing returns, where any more energy efficiency becomes very expensive, and possibly energy intensive to produce.
I have to agree about high-end video cards. I'm sure they too will come around, when cooling them becomes just too challenging.
CPUs, however, are in pretty good shape now. SpeedStep/PowerNow/CoolnQuiet/etc have become standard on PCs now, thanks to AMD, so an idle CPU is using less power than ever, despite being able to go up to 2GHz when performance is needed.
Intel, IMHO, is somewhat sabotaging this, though. When they rebrand their Core chips as "Celerons" one of the things they disable is SpeedStep. Add to that the needlessly hot northbridges on all their boards.
Buy a new computer today, with an AMD CPU and a Seasonic PSU, and it will probably use half as much power as your old one.
My point is... No doubt there is waste. However, it's not a substantial portion of the overall power use, and it would be either vastly inconvenient (driving a moped in winter) or expensive/unreliable enough that it might eliminate any real benefit to that savings.
Performance doesn't hardly matter. I have one system that does all the video encoding, game playing, etc., and the rest of my machines are just the cheapest hardware I can find on pricewatch when the old one finally breaks...
Mostly, I'd be happiest with the quietest, lowest power equipment I can get. A 25W Turion CPU, 80%+ efficient Seasonic PSU, etc. Not sure about GPUs, but something low powered, but has full-featured open source 3D drivers, and TV-out. Integrated chips need not apply.
Incidentally, this system is actually pretty cheap to put together. Newegg has a 2GHz Turion for $70, and Socket 754 mobos are only $50. Throw in $40 for the PSU, $40 for the GPU, and maybe $40 for RAM. Throw in the largest capacity HDD available, since I'm not paying for it...
The case really seems the most important part of a computer these days. Front ports for lots of frequently used connectors like audio, USB, Firewire, etc. I'd even like front RS-232 and IR. Importantly, good and quiet cooling is almost entirely a function of a well-designed case layout. Not to mention ease of maintenance is almost always hampered by cramped, poorly located components, and the usual jerry-rigged fans to handle hotspots that shouldn't even exist, in cheap/junk cases.
As for my multimedia machine, mostly the same as above, but lowest power (fast) Opteron available, RAID, and importantly, ECC RAM to guarantee stability/integrity.
For a notebook, a more expensive version of the OLPC, redesigned for adults, would be pretty close. Large HDD, sharper screen (in color mode), DVD burner, decent keyboard, etc.
.
What I'd really like to see in computers, is the various components working together like never before. How about if the monitor's power button actually just sends a signal to the PC, which can decide whether to shut off the monitor, suspend the system, start an idle timer, etc. It wouldn't hurt to have external speakers receive a signal, telling them to power-off when the system does. And networked information sharing, so I can quickly change the esound host for all the machines at once. More useful LEDs (HDD but no CPU activity LED?). A small, 2-line LCD for status/info output while the monitor is off would be good as well.
robots.txt is meant to be a mutually beneficial standard for the site owner, and the spider operator alike. It started as a method to prevent scanning CGI and other processed or infinite-loop pages from being scanned, and just morphed into a form of restrictions.
If you want enforcement, you password-protect a section of your website, or at least write an
That's how it always works. When any industry is new, and there aren't yet any laws on the books, everyone works together, and defines their own. Most of the time, when it becomes an issue, lawmakers take these accepted practices and make them straight into laws, with few changes.
Before established as a law, it's commonly called "industry standards," which, while not a law, is quite a good argument in court. For instance, if a commercial jet crashes, and it's found that the airline was doing the FAA required maintenance, but not doing the level of maintenance that every other airline is doing, they're likely to be held liable, despite their complete compliance with the applicable laws.
Except for the fact that domain squatting started, and was the most prevalent, back in the old days, when that was EXACTLY the situation. More money means fewer typo-domains, and the like, but squatting any common name/brand URL is guaranteed to return tens of thousands of dollars.
Except you're now at the mercy of CheapISP, and have no way to transfer to another service provider.
And you know where that's going to lead... Somebody buys com.com, and starts selling names more cheaply, until they get lots of users locked in, then raises their price through the roof.
An open and inexpensive DNS system is simply far better in many ways than any of the alternatives. I could understand making
Except for the fact that expensive registrars can be incredibly crappy, and cheap registrars can be damn good...
The problem with that theory, is that Microsoft makes significant changes to the Windows user interface every release. XP is hardly anything like 98.
Applications don't usually change as dramatically, but they certainly do have numerous smaller changes from version to version, and yet people will still upgrade.
People are stupid, but not nearly as mindless as you give them credit for.
I certainly don't believe 90% of computers run Adobe products (Flash/AcroRead not withstanding). That might make a dent in certain sectors, but not many.
The Unichrome drivers are buggy as hell. No idea about Intel's work, since I don't have an Intel board, and certainly wouldn't buy one just to get a video chipset anyhow. If Intel made PCI Express cards, then maybe.
Additionally, I wouldn't recommend integrated chipsets to anyone. Even on 2D, integrated video will drag your computer down with it. With 3D, you've got to be some sort of masochist.
I'd expect to see better performance from an old ATI Radeon card, with the open source Mesa DRI drivers, than a brand new Intel/Via video chip.