Could it have been the result of a knowledgeable outsider?
If it only happened on the digital tier, this is possible. A hacker could get access to the web GUI of a multiplexer, and take an MPEG-2 program from one output QAM and move it to another with PID remapping.
If it happened on the analog tier as well, that would mean a hack of the baseband video router, which I find more difficult to believe (many plants don't have these commanded over IP) , but I suppose it can't be ruled out.
Reasonable providers (read those that aren't Comcast) will separate adult and non-adult programming onto different QAMs.
Here is where the problem could have occured:
1) Video input problem: MSO Satellite Radio tuned to wrong channel. Doubtful because of authentication / encryption. This is more likely if the east coast/west coast feed of the same provider got swapped. Also if the MSO was using an over-the-air TV receiver, that of course can't happen. If the MSO gets a video fiber from the TV station, that might be a fiber carrier routing screw up.
2) Video router misconfiguration: Ff the satellite radio outputs baseband video, it may go through a video crosspoint swtich (they call them "routers"). The wrong crosspoints between the satellite radios and the modulators (analog tier) or encoders (digital tier)
3) Multiplexer/CherryPicker misconfiguration: On the digital tier, MPEG-2 programs are multiplexed together into a QAM. It is possible there was a brief misconfiguration.
Big question - did this happen on the Comcast SD digital tier, HD digital tier, the analog tier, or some combination? If it was both analog and digital SD, I would suspect a video router crosspoint misconfiguration. If it was just digital SD or digital HD, I would suspect a multiplexer misconfiguration.
do you think the public would accept the idea of a taxpayer-financed but privately-operated school
Public grants for private schools been accepted in the Netherlands, Poland, Sweden, and Finland. At least private school operating costs are granted by the govenment to some private schools in the UK and France.
Of course abolishing the publicly run schools has not happened anywhere yet.
In all this talk of "Gbps Internet", I am interested in what end-user Internet speeds are actually achieved in Korea. As of right now, there have ony been a few million 10 GBe ports shipped, so I am wondering how anything near 1 Gbps line speed can be maintained by any significant number of users at the highest hierarchy of the networks.
It would be interesting to look at TCP speeds on large downloads from a number of servers around the world, with the numbers from Korea and the US compared.
It is possible that "within network" speeds in Korea are very fast, but Internet speeds may or may not be.
I have sub-Mbps Internet speeds at home, but I don't find it a problem for VOIP or videoconferencing or downloading most things short of a whole CD ISO.
.It's not really so much of a "spread out" problem. It's a problem of SCALE. Any time you scale a project up orders of magnitude, you get problems. It's the same problem with large corporations and bureaucracies.
I'm also wondering how many permits, environmental impact statements, and other NIMBYism would be encountered in a massive, country-wide trenching project.
An important lesson for you, a wealthy individual who made their money using illegal business tactics and whose company developed an admitted reputation for deceiving it's customers and who happens to have absolutely no qualifications in the area concerned should not be able to get in the position where the set government policy by the application of their wealth
I think that Gates has on board some people with significant qualifications, such as Dr. Vicki Philips, who was superintendent of Portland Public Schools, secretary of education and chief state school officer for the state of Pennsylvania, superintendent of the Lancaster, Pa., school district, worked in the U.S. Office of Education in Washington, D.C., and has been a middle and high school teacher.
The Gates foundation also announced grants to support research into the impact of teacher-level characteristics on student achievement to ACT Inc., Teach For America, the Educational Testing Service for a research collaboration with the RAND Corporation and the Institute for Social Research at the University of Michigan.
Federal oversight of education is absolutely required
I'm sure that will work just as well as Federal oversight of military spending. Personally, I'd prefer PARENTAL oversight of education spending, i.e. parents making decisions about which schools their children should attend within a free market. Of course, you'd have to end the socialist monopoly of schools first.
Instead maybe it's time to look at schools one at a time and recognize that properly running a school is a management challenge like any other.
I'm sure that government-run schools will continue be as efficient as, say, military purchasing.
In most districts, per-pupil spending is $5000-$10,000 per year. Give that money directly to kids, and take government out of schools. Let the market sort out the winners from the losers, and let parents choose which school is best for their child (perhaps based on their child's optimal learning mechanism).
Gates made the claim that the schools that were able to change took "radical steps to change the culture, such as allowing the principal to pick the team of teachers or change the curriculum."
I'm not sure how "good" a teacher needs to be. A teacher needs to be "good" in terms of maintaining class discipline. A teacher needs to not be totally ignorant of their subject. But short of that, a teacher needs to teach a properly constructed curriculum that allows students to take appropriate steps forward in learning, and the results need to be measured.
But perhaps letting the principal pick the team of teachers, rather than a faceless civil service bureaucracy, is a good idea.
Also Gates claimed his best results were at "charter schools that have significantly longer school days than other schools." Thus, total time spent learning may play a role in optimal performance as well.
This is in contrast to Japan where the Japanese government did nearly nothing in similar situation a few years before
Regarding Japan's Lost Decade, I suggest you review the data. Here is what actually happened:
1) Bad initial monetary policy, the Bank of Japan actually boosted the discount rate from 4.75 percent to 6 percent in August 1990 and held it at that level until June 1991. This was a similar mistake to the monetary contraction by the US Fed from 1929 to 1933.
2) Japan engaged in large stimulus packages - outlays on big projects from 6.5% of GDP in 1990 to 8.3% in 1996. Most went towards largely toward unproductive public works projects and credits to small businesses that were no longer economically viable ("zombie companies"). Sound familiar? These did not do much to help the situation evidently.
3) In 1997 the Japanese government raised its consumption tax from 3 to 5 percent, which lead to an increase in deflation in 1998.
4) To end the "Lost Decade", the Bank of Japan switched during the spring of 2001 to a policy of quantitative easing and PM Junichiro Koizumi was undertook aggressive measures to recapitalize Japan's banks, which were still heavily burdened by nonperforming loans. (The latter is what TARP was supposed to do).
How in the heck would you sync the two? Doubt any two converter boxes have the same delay, just factoring in the correct delay in the encoding stage would be a total PITA
A subcarrier could be used to digitally (but narrowband) key PTS values. However, since most STBs seem to ignore PTS/DTS anyway, having a STB knob on top to tweak A/V sync might be good even under ATSC!
The UK found an increased energy cost for consumers (because they'd need a set-top box), and a much reduced energy cost for transmitters.
The UK uses DVB-T (COFDM) rather than ATSC (8VSB) for DTV, although truth be told 8VSB should be cheaper than COFDM because 8VSB has a much lower peak-to-average power ratio. 8VSB was supposed to be a computationally less demanding demodulation than COFDM, but once people realized how multipath killed 8VSB they now have serious equalizers on the front so the demodulation is much tougher now. The demod is probably small potatos compared with the MPEG-2 decode, especially for HD.
By converting the wasteful analog transmissions to more efficient digital, they reclaimed a resource
I'm really not sure about this. Yes you get HDTV capability in the 6 MHz channels in DTV rather than ananlog NTSC SD, but in terms of gathering the channels together into the "in-core" channels of 51 and below, you probably could have done this even without going digital by putting tighter transmitter emission masks on analog transmitters (the DTV emission mask is -44dB adjacent channel interference)
I'll add the caveat that I expect we will be surprised by some adjacent & next adjacent channel interference and intermodulation interference in DTV especially once all DTV powers are maxed out and final DTV channel changes occur, especially for distant stations. It may turn out we are not able to truly push together all these channels without enlarging areas of non-reception due to interference.
When they finally do it, instead of shutting off all analog signal they need to make every station in the country broadcast a repeating message for a week explaining what happened and giving instructions plus a phone # to call for more details. That's about the only way to limit the number of angry phone calls that everyone from the electric companies to the stores that sold the remote controls will get.
There is supposed to be a 30-day analog "night light", along these lines at least according to the Feb. 17 turn-off plan. Not sure what the deal is if they push the cut-off to June.
I work for PBS. No check will be written. Our money comes from viewers, sponsorships and endowments.
As someone who used to work for PBS and aided in the plan for the $120 million taxpayer-funded satellite upgrade, I think you may be missing the $50-$100 million per year that regularly comes from the CPB (from Federal tax dollars) or directly from Federal grants to PBS.
Then you need to consider that CPB gets around $400 million from the Federal Government each year, and they turn around and grant $200 million of that directly to public television stations as Community Service Grants, who then turn around and pay $150-$200 million to PBS as member assessments .
I don't know why this is so hard to communicate, but nobody seems to get it. It's the UHF spectrum where the DTV signals are now that is being freed up, not the current VHF channels.
Because this is wrong. The only channels being "evacuated" are the "out-of-core" UHF channels 52 and higher. Below that, it is only a matter of reshuffling. Some analog VHF stations are staying VHF in DTV, others have UHF DTV assignments. Some analog stations in UHF are staying UHF in DTV, or moving to VHF in DTV.
For example, in Washington, DC, all analog TV stations regardless of VHF or UHF have UHF DTV channels.
In Las Vegas, five analog VHF stations (KVBC, KVVU, KLAS, KLVX, KTNV) have VHF DTV channels.
But then there are UHF analog channels that have VHF DTV channels, such as Albany WXXA which is analog UHF 23 but digital VHF 7.
The shuffling is even more bizarre for stations with out-of-core DTV assignments. Los Angeles KCOP is on analog VHF 13 and currently out-of-core DTV UHF 66, which will move DTV to in-core VHF 13 after the analog shut-off.
"It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. We address ourselves, not to their humanity, but to their self-love, and never talk to them of our necessities but of their advantages."
-Adam Smith, The Wealth of Nations, Book I Chapter II
That was the reason for all the various alphabet soup agencies -- to do something about the surplus labor.
Of course the NRA did not suck up surplus labor, it cartelized industry and labor to try to artificially raise wages and prices (which pretty much failed, and was also found unconstitutional). Perhaps investors were scared of the regime uncertainty of the rapidly expanding role of government in the market, reducing their confidence in the durability of private property rights.
So, yes, it was Roosevelt that turned around the Great Depression.
The question is what did Roosevelt do that "turned around the Great Depression"? As you mentioned GDP growth returned in 1934, one year after taking office. What most people think of the "New Deal" in terms of NRA, CCC, WPA did not ramp up for years. What did happen in 1933 was the Gold Clause ban and the dollar devaluation to reverse the highly contractionary money supply changes the Fed had been engaging in since 1929. That said, while GDP recovered to pre-120 levels by 1937, unemployment definitely did not until WWII (when many people were put to work at the "point of a gun").
You don't want people up on their rooftops in mid-February adjusting their antennas after the switchover.
And you really don't want to get the job of going up a mountain-top 500 foot tower during the winter to move/adjust a TV transmitter antenna/feedline!
While most stations have a DTV signal up, it is usually on a different antenna from the analog signal, and many stations are planning to remove their analog antenna and/or move their DTV antenna to a better position post-cut-off.
On the other hand, John Maynard Keynes was right. Recessions are caused by too little spending. Right now, consumers are (on average) overextended, so cannot increase spending. Businesses see no economic returns on additional spending. So they can't increase spending. That leaves government.
Or we could wait until the economy re-organizes itself (less finance and builiding, more health care & flying cars or whatever works) so that the economy can go back to creating wealth, enabling spending. By not spending tax dollars during that time, we save wasting wealth (current taxes or future taxes to pay down debt) on government boondoggles.
The economy can best re-organize when there are few inappropriate regulations to slow down the re-organization.
If you really are worried about short-term effects (like unemployment rises) during this period of re-organization, then perhaps reduce the tax on employment (payroll taxes, for example) and you can even offset it with a tax on something we don't want (such as carbon).
I do agree that if we had 10% unemployment for more than a year, it would hurt the human capital stock of the US. Longer term unemployment is linked to significantly reduced future personal earnings.
The Government has done this because private corporations are not willing to pay for something you just give away free to the public
The one counter-example are private Universities, which do spend their own money on publicly available basic research.
(source) total basic research spending in universities in 2001 was $20.8 billion. $12.9 billion came from the Federal government, and $7.8 billion came from non-Federal sources.
Institutional funds (e.g. university endowments) are the largest source of non-Federal university basic research spending, followed by industry and state/local funding.
Basic research, of course, pales in comparison with the $250 billion total amount of US R&D done (source).
Slashdot Mirror
Could it have been the result of a knowledgeable outsider?
If it only happened on the digital tier, this is possible. A hacker could get access to the web GUI of a multiplexer, and take an MPEG-2 program from one output QAM and move it to another with PID remapping.
If it happened on the analog tier as well, that would mean a hack of the baseband video router, which I find more difficult to believe (many plants don't have these commanded over IP) , but I suppose it can't be ruled out.
Reasonable providers (read those that aren't Comcast) will separate adult and non-adult programming onto different QAMs.
Here is where the problem could have occured:
1) Video input problem: MSO Satellite Radio tuned to wrong channel. Doubtful because of authentication / encryption. This is more likely if the east coast/west coast feed of the same provider got swapped. Also if the MSO was using an over-the-air TV receiver, that of course can't happen. If the MSO gets a video fiber from the TV station, that might be a fiber carrier routing screw up.
2) Video router misconfiguration: Ff the satellite radio outputs baseband video, it may go through a video crosspoint swtich (they call them "routers"). The wrong crosspoints between the satellite radios and the modulators (analog tier) or encoders (digital tier)
3) Multiplexer/CherryPicker misconfiguration: On the digital tier, MPEG-2 programs are multiplexed together into a QAM. It is possible there was a brief misconfiguration.
Big question - did this happen on the Comcast SD digital tier, HD digital tier, the analog tier, or some combination? If it was both analog and digital SD, I would suspect a video router crosspoint misconfiguration. If it was just digital SD or digital HD, I would suspect a multiplexer misconfiguration.
do you think the public would accept the idea of a taxpayer-financed but privately-operated school
Public grants for private schools been accepted in the Netherlands, Poland, Sweden, and Finland. At least private school operating costs are granted by the govenment to some private schools in the UK and France.
Of course abolishing the publicly run schools has not happened anywhere yet.
In all this talk of "Gbps Internet", I am interested in what end-user Internet speeds are actually achieved in Korea. As of right now, there have ony been a few million 10 GBe ports shipped, so I am wondering how anything near 1 Gbps line speed can be maintained by any significant number of users at the highest hierarchy of the networks.
It would be interesting to look at TCP speeds on large downloads from a number of servers around the world, with the numbers from Korea and the US compared.
It is possible that "within network" speeds in Korea are very fast, but Internet speeds may or may not be.
I have sub-Mbps Internet speeds at home, but I don't find it a problem for VOIP or videoconferencing or downloading most things short of a whole CD ISO.
.It's not really so much of a "spread out" problem. It's a problem of SCALE. Any time you scale a project up orders of magnitude, you get problems. It's the same problem with large corporations and bureaucracies.
I'm also wondering how many permits, environmental impact statements, and other NIMBYism would be encountered in a massive, country-wide trenching project.
An important lesson for you, a wealthy individual who made their money using illegal business tactics and whose company developed an admitted reputation for deceiving it's customers and who happens to have absolutely no qualifications in the area concerned should not be able to get in the position where the set government policy by the application of their wealth
I think that Gates has on board some people with significant qualifications, such as Dr. Vicki Philips, who was superintendent of Portland Public Schools, secretary of education and chief state school officer for the state of Pennsylvania, superintendent of the Lancaster, Pa., school district, worked in the U.S. Office of Education in Washington, D.C., and has been a middle and high school teacher.
The Gates foundation also announced grants to support research into the impact of teacher-level characteristics on student achievement to ACT Inc., Teach For America, the Educational Testing Service for a research collaboration with the RAND Corporation and the Institute for Social Research at the University of Michigan.
Federal oversight of education is absolutely required
I'm sure that will work just as well as Federal oversight of military spending. Personally, I'd prefer PARENTAL oversight of education spending, i.e. parents making decisions about which schools their children should attend within a free market. Of course, you'd have to end the socialist monopoly of schools first.
Instead maybe it's time to look at schools one at a time and recognize that properly running a school is a management challenge like any other.
I'm sure that government-run schools will continue be as efficient as, say, military purchasing.
In most districts, per-pupil spending is $5000-$10,000 per year. Give that money directly to kids, and take government out of schools. Let the market sort out the winners from the losers, and let parents choose which school is best for their child (perhaps based on their child's optimal learning mechanism).
You don't *need* money, you need good teachers.
Gates made the claim that the schools that were able to change took "radical steps to change the culture, such as allowing the principal to pick the team of teachers or change the curriculum."
I'm not sure how "good" a teacher needs to be. A teacher needs to be "good" in terms of maintaining class discipline. A teacher needs to not be totally ignorant of their subject. But short of that, a teacher needs to teach a properly constructed curriculum that allows students to take appropriate steps forward in learning, and the results need to be measured.
But perhaps letting the principal pick the team of teachers, rather than a faceless civil service bureaucracy, is a good idea.
Also Gates claimed his best results were at "charter schools that have significantly longer school days than other schools." Thus, total time spent learning may play a role in optimal performance as well.
This is in contrast to Japan where the Japanese government did nearly nothing in similar situation a few years before
Regarding Japan's Lost Decade, I suggest you review the data. Here is what actually happened:
1) Bad initial monetary policy, the Bank of Japan actually boosted the discount rate from 4.75 percent to 6 percent in August 1990 and held it at that level until June 1991. This was a similar mistake to the monetary contraction by the US Fed from 1929 to 1933.
2) Japan engaged in large stimulus packages - outlays on big projects from 6.5% of GDP in 1990 to 8.3% in 1996. Most went towards largely toward unproductive public works projects and credits to small businesses that were no longer economically viable ("zombie companies"). Sound familiar? These did not do much to help the situation evidently.
3) In 1997 the Japanese government raised its consumption tax from 3 to 5 percent, which lead to an increase in deflation in 1998.
4) To end the "Lost Decade", the Bank of Japan switched during the spring of 2001 to a policy of quantitative easing and PM Junichiro Koizumi was undertook aggressive measures to recapitalize Japan's banks, which were still heavily burdened by nonperforming loans. (The latter is what TARP was supposed to do).
How in the heck would you sync the two? Doubt any two converter boxes have the same delay, just factoring in the correct delay in the encoding stage would be a total PITA
A subcarrier could be used to digitally (but narrowband) key PTS values. However, since most STBs seem to ignore PTS/DTS anyway, having a STB knob on top to tweak A/V sync might be good even under ATSC!
At 220V, that means you'd need 1300 A of current!?!
It's not just an electric car, it's also an arc welder!
The UK found an increased energy cost for consumers (because they'd need a set-top box), and a much reduced energy cost for transmitters.
The UK uses DVB-T (COFDM) rather than ATSC (8VSB) for DTV, although truth be told 8VSB should be cheaper than COFDM because 8VSB has a much lower peak-to-average power ratio. 8VSB was supposed to be a computationally less demanding demodulation than COFDM, but once people realized how multipath killed 8VSB they now have serious equalizers on the front so the demodulation is much tougher now. The demod is probably small potatos compared with the MPEG-2 decode, especially for HD.
By converting the wasteful analog transmissions to more efficient digital, they reclaimed a resource
I'm really not sure about this. Yes you get HDTV capability in the 6 MHz channels in DTV rather than ananlog NTSC SD, but in terms of gathering the channels together into the "in-core" channels of 51 and below, you probably could have done this even without going digital by putting tighter transmitter emission masks on analog transmitters (the DTV emission mask is -44dB adjacent channel interference)
I'll add the caveat that I expect we will be surprised by some adjacent & next adjacent channel interference and intermodulation interference in DTV especially once all DTV powers are maxed out and final DTV channel changes occur, especially for distant stations. It may turn out we are not able to truly push together all these channels without enlarging areas of non-reception due to interference.
When they finally do it, instead of shutting off all analog signal they need to make every station in the country broadcast a repeating message for a week explaining what happened and giving instructions plus a phone # to call for more details. That's about the only way to limit the number of angry phone calls that everyone from the electric companies to the stores that sold the remote controls will get.
There is supposed to be a 30-day analog "night light", along these lines at least according to the Feb. 17 turn-off plan. Not sure what the deal is if they push the cut-off to June.
I work for PBS. No check will be written. Our money comes from viewers, sponsorships and endowments.
As someone who used to work for PBS and aided in the plan for the $120 million taxpayer-funded satellite upgrade, I think you may be missing the $50-$100 million per year that regularly comes from the CPB (from Federal tax dollars) or directly from Federal grants to PBS.
Then you need to consider that CPB gets around $400 million from the Federal Government each year, and they turn around and grant $200 million of that directly to public television stations as Community Service Grants, who then turn around and pay $150-$200 million to PBS as member assessments .
Some European countries have been on DTV for years. They said "this is the date" and switched. No coupon programs, no hand-outs, no endless delays.
It should be noted that when Berlin had the analog shutoff, they did have a coupon program.
I don't know why this is so hard to communicate, but nobody seems to get it. It's the UHF spectrum where the DTV signals are now that is being freed up, not the current VHF channels.
Because this is wrong. The only channels being "evacuated" are the "out-of-core" UHF channels 52 and higher. Below that, it is only a matter of reshuffling. Some analog VHF stations are staying VHF in DTV, others have UHF DTV assignments. Some analog stations in UHF are staying UHF in DTV, or moving to VHF in DTV.
For example, in Washington, DC, all analog TV stations regardless of VHF or UHF have UHF DTV channels.
In Las Vegas, five analog VHF stations (KVBC, KVVU, KLAS, KLVX, KTNV) have VHF DTV channels.
But then there are UHF analog channels that have VHF DTV channels, such as Albany WXXA which is analog UHF 23 but digital VHF 7.
The shuffling is even more bizarre for stations with out-of-core DTV assignments. Los Angeles KCOP is on analog VHF 13 and currently out-of-core DTV UHF 66, which will move DTV to in-core VHF 13 after the analog shut-off.
Hardly a charitable act
"It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. We address ourselves, not to their humanity, but to their self-love, and never talk to them of our necessities but of their advantages."
-Adam Smith, The Wealth of Nations, Book I Chapter II
That was the reason for all the various alphabet soup agencies -- to do something about the surplus labor.
Of course the NRA did not suck up surplus labor, it cartelized industry and labor to try to artificially raise wages and prices (which pretty much failed, and was also found unconstitutional). Perhaps investors were scared of the regime uncertainty of the rapidly expanding role of government in the market, reducing their confidence in the durability of private property rights.
So, yes, it was Roosevelt that turned around the Great Depression.
The question is what did Roosevelt do that "turned around the Great Depression"? As you mentioned GDP growth returned in 1934, one year after taking office. What most people think of the "New Deal" in terms of NRA, CCC, WPA did not ramp up for years. What did happen in 1933 was the Gold Clause ban and the dollar devaluation to reverse the highly contractionary money supply changes the Fed had been engaging in since 1929. That said, while GDP recovered to pre-120 levels by 1937, unemployment definitely did not until WWII (when many people were put to work at the "point of a gun").
Because Americans won't work for 3 pieces of rice / day, and require that their employers don't needlessly endanger them?
How many Slashdot users would ever make shoes or plastic toys?
Are any of the Mars Rovers near the methane plume sites?
You don't want people up on their rooftops in mid-February adjusting their antennas after the switchover.
And you really don't want to get the job of going up a mountain-top 500 foot tower during the winter to move/adjust a TV transmitter antenna/feedline!
While most stations have a DTV signal up, it is usually on a different antenna from the analog signal, and many stations are planning to remove their analog antenna and/or move their DTV antenna to a better position post-cut-off.
On the other hand, John Maynard Keynes was right. Recessions are caused by too little spending. Right now, consumers are (on average) overextended, so cannot increase spending. Businesses see no economic returns on additional spending. So they can't increase spending. That leaves government.
Or we could wait until the economy re-organizes itself (less finance and builiding, more health care & flying cars or whatever works) so that the economy can go back to creating wealth, enabling spending. By not spending tax dollars during that time, we save wasting wealth (current taxes or future taxes to pay down debt) on government boondoggles.
The economy can best re-organize when there are few inappropriate regulations to slow down the re-organization.
If you really are worried about short-term effects (like unemployment rises) during this period of re-organization, then perhaps reduce the tax on employment (payroll taxes, for example) and you can even offset it with a tax on something we don't want (such as carbon).
I do agree that if we had 10% unemployment for more than a year, it would hurt the human capital stock of the US. Longer term unemployment is linked to significantly reduced future personal earnings.
The Government has done this because private corporations are not willing to pay for something you just give away free to the public
The one counter-example are private Universities, which do spend their own money on publicly available basic research.
(source) total basic research spending in universities in 2001 was $20.8 billion. $12.9 billion came from the Federal government, and $7.8 billion came from non-Federal sources.
Institutional funds (e.g. university endowments) are the largest source of non-Federal university basic research spending, followed by industry and state/local funding.
Basic research, of course, pales in comparison with the $250 billion total amount of US R&D done (source).