They're not just trying to flame the file-sharers.
They're trying to convert the Internet to a broadcast service (with themselves as the gatekeeping broadcasters). A centraized-producers, sea-of-consumers model.
These are the media conglomerates. They became multibillion dollar businesses in an environment where providing "content" had a high capital cost. This let them serve as gatekeepers, collecting a lot of money for the works they distributed (but still out-competing live performances due to economy of scale) and keeping most of it for themselves.
But advancing electronic and computer technology has brought the tools of mass-market "content creation" within the financial reach of the ordinary "starving artist". And the inherently peer-to-peer internet, with high bandwidth, flat rates, and cost-distributing tools like torrents, has brought the distribution costs to near-nothing. So artists without investors can bypass the gatekeeper and go straight to the customers/fans on an industrial scale.
This means that the artists can now out-compete the gatekeepers. So to protect their lucrative business they are taking a number of measures to attack, not just competition from unpaid distribution of their own material, but also this potential competition from artists escaping from their channels.
You see it here in the attempt to brand peer-to-peer as synonymous with crime. But you also see it in other places:
- Torrent throttling.
- Asymmetric connections with high download and low upload speeds (presupposing the ISP's customer consumes more than he produces).
- Bans on "servers".
- Legal moves to make ISPs responsible for the content of files they transfer, store, or serve. (Web sites, netnews, etc. Notice the attack on the.alt hierarchy for example.) This raises the cost of "distributing content" and reintroduces gatekeepers.
- And (of course) propaganda buried in the entertainment they produce. (Just as they once did crime shows villainizing video games, computer bulletin boards, and the like when they were competing with broadcast TV for eyeball time - and before that smearing cable TV operators as a class in the early days when they weren't and conglomerate-owned and were seen as pirates of the off-the-air signals and competitors purveying imported and local content.)
If they can get the general public to avoid peer-to-peer and go back to central producer / distributed (and paying!) consumer, they've won the battle.
Whenever I start contemplating DNA (!), self-reproduction and the utter insanity of how complex the machinery of a single cell is, much less multicellular life, much less an animal, much less a self-aware brain, I just shake my head in wonder.
Doesn't bother me. Evolution is a massively parallel computation and has been going on for a LONG time.
If you skip DNA and just look at RNA it all gets easy:
- RNA caries genetic information and can be copied by an appropriate enzyme. (It's less stable than DNA, but quite stable enough to form the genomes of viruses. At the early stages, with no competition yet, being error-prone is actually good.)
- RNA has enzymatic activity. (It's not as strong or as versatile as protein-based enzymens. But it is quite capable of folding itself up into structures coded by its sequence, sticking together at appropriate places and presenting controlled patterns of charges on outer surfaces of a controlled shape, to become a little molecular machine.
Nucleotides line themselves up on a strand of either RNA or DNA to form the complimentary code sequence. They'll bind themselves into a strand given enough time and jostling. But if you have a RNA strand that also sometimes folds up into a little zipper-tab that runs down the lined-up RNA bases and sticks them together into a fresh strand you're all set:
- You'll eventually have both that and its compliment hanging around in the same container.
- At some point the strand that folds up into a zipper will zip up the new bases stuck to its complimentary strand. Then you have TWO zippers tab strands plus a complimentary strand.
- Now the zipper strand(s) start churning out new zipper strands and complimentary strands.
Slow at first, because rev 0.1 probably doesn't work well and it's completely dependent on randomly occurring bases for "food". But with the exponential under way the errors start to accumulate. Now you get some that are better at zipping than others - and they dominate the regions where they occur. And you get strands with multiple copies and other noise sequences - which can now evolve separately within the strand and evolve new functions.
Whenever a strand evolves one of its "spare" "genes" into a machine to help out, it becomes more successful.
From there you can evolve:
- Machines to make components of the system from other "useless" stuff.
- Machines to string amino-acids into useful structural stuff - and eventually better machines.
- Machines to control a container, creating the "cell" and its division mechanism.
- Machines to make backup copies of the RNA code in more stable DNA and then make more RNA from that. and so on.
There's plenty of suggestions that this is what happened. For instance: Most of the machinery of RNA-directed protein synthesis - both most of the parts of the ribosome (the stringing factory) and all of the transfer RNA (the amino-acid carrier/code reader mechanisms) are RNA enzymes.
So, no, contemplating the current complexity doesn't bother me at all. It can all be explained by evolution from a single, simple, mechanism that could easily be produced in millions of years of random abiotic chemical reactions in a planetary scale vat of solar-irradiated, weather-stirred chemicals.
IBM was merrily outfitting the Nazis with equipment to help them manage their concentration camps (completely ignorant of their application, naturally) while Bletchley park was breaking Nazi codes.
It was also manufacturing M1 rifles for the Allies. (Along with such companies as Rockola - the jukebox maker - and Saginaw Steering Gear. It's handy to convert a factory to guns when it already has equipment for drilling holes the long way down several feet of steel rod and other machines for building small and complicated devices composed of mechanical moving parts.)
(Back when I was buying an old M1 carbine for participating in the Civilian Marksmanship Program training I picked an IBM-branded one just out of nostalgia. The rangemaster was impressed when I qualified with a carbine, rather than a full-length M1, on the first try. Shorter barrels make for less accuracy. B-) But I could have used a Field Engineer: While the steel parts worked fine, the wooden barrel cover kept popping off during recoil. B-( )
The same statistical argument applies to creating DNA databases and using them to hunt for perpetrators.
A DNA mismatch is essentially unambiguous. But a hit is a statistical issue, and hard to quantify because of unknown correlation of markers due to inheritance.
So using it to eliminate suspects is basic. And using it to point to one of a small number of unrelated suspects identified by other means is plausible. But throw the DNA pattern from a crime scene at a database containing thousands, or millions, of people (but not ALL of them) and you are just asking for false positives.
And as the tests get better and the false positive rate goes down it becomes even more dangerous, because it becomes more tempting to rely on it when only one "hit" shows up. (Consider, for example, identical twins separated at birth, where one is in the database, the other commits a crime, and the poor sap that gets fingered doesn't even know the twin exists.)
And someone brought in on a false positive has an extra handicap: DNA testing won't exonerate him (unless there's enough DNA evidence that more extensive tests can be performed and it happens to find an additional marker that mismatches.)
Whether this is good or bad depends on your point of view to a certain extent, but what is clear is that Hollywood isn't planning on abandoning DRM any time soon. As a result, they're not going to adopt any form of writable media to store digital copies unless it has some kind of DRM system built-in.
And as a second-order result of that they're going to miss the fast nickels while chasing the slow dimes.
But private enterprise is all about being free to make stupid business decisions and taking the resulting consequences on the bottom line. (And losing the fast nickels is NOT one of the consequences they'll be able to lobby their way out of.)
DRM, in this case, is a choice between DRM'd content and no content at all. I'd rather have DRM'd content than none.
I, on the other hand, would rather do with none than with DRM.
I made that choice when the DVDs supplanted videotape and didn't buy DVD movies - or buy or rent any movies at all - until after CSS was cracked and the movie industry gave up on their attempts to stuff that genie back into the bottle. No blu-ray players for me, either. Stopped buying CDs, too, when they started experimenting with the early computer-speaker-blowing "copy protection" that corrupted the data and depended on the error correction on players to recover the music (and thus corrupted it when you got real errors from a dirty disk) and never really got back into purchasing new music after that.
Never actually MADE a backup copy. And never downloaded a "pirated" song or movie, either. I just don't buy encumbered stuff.
Instead I found other ways to amuse myself. (For instance: The amazing number and variety of animals outside the place on the high desert put on a continuous show that's quite entertaining - especially when I flush the well and create a puddle that draws them in from miles around. And there's lots of amusement on the net that is not "pirated" copyrighted content.)
Interestingly, I don't really miss the corporate "content". Either the quality took a nosedive around that time or the product stopped matching my (quite broad) tastes. (Though from what I hear of some local bands it's more the former than the latter.)
... what happens when a DNS record gets updated? With the new behavior, you won't see the change until your cached record expires.... I believe BIND was working as intended to allow updatedrecords to overwrite older ones.
That's what I thought too, at first. It seems right to replace cahced records with newer ones. And perhaps that's why the code is written as it is.
But then I got to thinking: If the server has a cached record already, it wouldn't have asked for and update. So why is the new information there in the first place?
To make matters worse, NVidia has a mound of unsold defective parts that they are going to bleed out into the channel along side of the (hopefully) fixed parts.
This sounds very similar to what finally took down Weitek, back when there were a bunch of graphics chip companies competing hotly and being shaken out if they screwed up.
Weitek had built a very fast and powerful chip. But they had goofed: While it had the mandatory basic VGA mode for acquiring the Microsoft certification, there was a bug in it.
QA told management that the bug was there and would fail them. But Software told them a driver could work around it and people would want the chip because it was so fast on graphics rendering. (Of course it could not - because to get the cert it had to work with the stock bootstrap stuff, before a custom driver could be loaded.)
So they went to production with the bug. And the customers got their prototypes, found the bug, and demanded a fix. Eventually they did a fixed version - but had maybe a couple million of the buggy ones on hand and wouldn't sell the fixed ones unless the customer bought some buggy ones, too. So nobody bought and the company folded.
It's clear they're using a viral capsule for delivery - from a virus that normally targets the cell type the want to transform.
But I'm curious as to what they're delivering:
1) The transcription regulatory proteins?
2) Messenger RNA coding for the proteins (resulting in limited production of the proteins until the RNA "times out").
3) DNA (or RNA plus a RNA-directed DNA polymerase) to add new code to the cell iteslf?
4) A modified virus deleting most of the virus code and adding code for the proteins?
I'd be happy with 1) or 2). Either looks ideal. (2 looks especially easy to generalize, using any appropriate RNA virus to supply a capsule aimed at a suitable target and stringing RNA for the proteins together with the virus' "pack-me-up" tag and some "cut me here" introns). Just make sure you've sorted out real viral DNA (say, by making just the desired proteins in e-coli and assembling things after purification).
3) and 4) have some nasty downsides.
Perhaps someone with access to the article can tell us which it is, please?
If the answer is 1) or 2) I'll do a happy dance. B-)
What people forget is Rome 'fell' for a very, very long time and 'fell' for a number of reasons.
In fact the last stage of the fall was the termination of the reign (and in one case the line) of last two claimants to the title of "Caesar": the Czar of Russia and the Kaiser of the German Empire.
That, and, IIRC a federal law that obliged it to finance research with 50% of its profits in exchange for the monopoly.
As I recall it wasn't a requirement that they do research. It was permission to include the cost of research related to telephony in their cost of doing business - on which they got to set monopoly phone rates so they got a specified rate of return (6% if I recall correctly).
The result was that the more money they spent on research, the more profit they made.
So they set up Bell Labs to spend as much money as possible, on anything even vaguely related to telephony.
And it "was an abysmal failure". From year one they made more money on the results of the work (by things like licensing patents) than it cost to run the labs. So basic research was profitable all by itself. B-)
But this counterintuitive effect also has a counterintuitive downside. The rewards for a research project start once it's done and keep coming in for quite a while after it's finished. Most of us would consider this good. But the Harvard Business School approach to management comes into play: The incentive structure on managers is to show as big a profit as possible for a few years and move on, thus looking better than your predecessors and successors and getting progressively better paying positions. So by killing the CURRENT research and just collecting on the results of the previous work they can cut their costs to near nothing while the benefits keep rolling in. For a while. Then they move on. Without new work the revenue gradually dries up and their successors take the rap. (And their successors would have to increase costs while the income was ramping down, which would look even worse, to turn things around.)
Regardless:
Without the guaranteed profit they're in the same boat as every other large cashflow company in the world. Perhaps basic research would continue to be profitable beyond the dreams of avarice. But there are other profitable things to do with the money where the return is more visible in advance, rather than crapshooting on what basic research might come up with. So (like all those other companies), the new generation of management reacts to the new situation by doing the standard thing - which doesn't include basic research.
(And it doesn't help that they already went through the "cut expenses and look good on the return on old work" phase a few years back. IMHO this is the house of cards coming down.)
= = = =
As I understand it, Xerox PARC had something similar going on but for a different reason: a strange accounting system.
One of the first things PARC did was to design a "new control panel" (and brain) for Xerox copiers. This replaced a bunch of relay logic with a microcomputer/early logic chips. And that saved a LOT of money.
PARC got credited with that savings on all the copier products sold from then on (and with similar stuff it did later). So it could spend money hand-over-fist on whatever it wanted and still look profitable.
(This was the same accounting department that, if I've got THIS right, screwed up big time when Xerox went into the mainframe business as the first company to take on IBM's core business, 'way back in the early days of "foreign attachments" opening up the IBM big-iron market. They built a CPU. After a while they decided that they were in the red on it big time, folded the division, and sued IBM on antitrust. In those days equipment was all leased. As a result of the suit IBM got hold of their accounting info and discovered the hadn't really understood how to interpret lease income. They were actually VERY profitable, and had folded the division because of this accounting screwup. Of course this discovery folded the suit.)
Wind power does vary with weather. But at good sites the wind is primarily driven by the "lake effect" and is quite consistent. ("Lake effect" is a bit misleading: At the Altamont Pass wind farm, for instance, the "lake" is the Pacific Ocean and the "island" is California's Central Valley. Similarly, in Texas the "lake" is the Gulf of Mexico and the "island" is the Great Plains - much of the North American continent.)
Lake effect winds highest in the afternoon and evening, a reasonably good match for the peak electrical demand curve.
The multiple mills with significant separation that constitute a large wind farm average out gusts and even storm cells. As to weather changing the speed also affects load because much of load is driven by HVAC (Heating, Ventilation, Air conditioning) and higher wind means insulation is less effective, resulting in more heat gain or loss that must be made up by the HVAC system.
So, though the amount of power generated is not as stable as that from a reservoir full of water, a pile of coal, or a nuclear reactor, the variability of wind power actually REDUCES the amount of standby and peaking power required from the other generation sources on the grid.
(Solar has similar characteristics, with the peak leading that of the demand curve by a bit. Its peak, and variability, track the air conditioning load very well but run counter to the (smaller) heating curve. Wind power is larger for a given investment and goes up with weather that imposes extra heating load. So a mix of the two is even better at evening out the load on the rest of the grid's generation.)
'The' Saturn plant? I guess you don't realize this is a forum that covers more than just your neighborhood.
I'm referring to the original Saturn Corporation plant, in Spring Hill Tennessee, which is decidedly NOT in my back yard.
GM's choice of site was international news - including an editorial cartoon showing a ringed planet crashed into a farmer's field in a mountainous rural setting.
That's just the thing - they don't build new auto plants or aluminum smelters out the hell in the middle of nowhere for just that reason.
I guess you don't remember the Saturn plant, eh? (Just for starters.)
ANY time you build a new plant - unless it's on the site of an old one - you'll need to string a new high line and a bunch of other infrastructure. The sites where the land is cheap enough to make it economically feasible will not normally have it already available. And the cost of installing it is small compared to the difference in land price between building in an existing urban area and building in something more open.
Ditto with building a new power plant, of course. (Even for a new peaking plant in an urban area you still have to string something. The chunk of the grid that powers a warehouse won't handle even a tiny power plant. If you're in the SF Bay area think about the miles of high-line they ran along I-880 to support the plant near the CA 237 intersection, just for a little gas-fired plant.)
Wind farms are just a special case of this. Like hydropower dams, they have to be put where the resource is. And this will be somewhere that the existing power transmission lines aren't big enough to handle the energy harvested.
The summary is a crock and doesn't match the quoted article.
Transporting large and variable amounts of generated power is the dual of feeding large and varying loads. The power grid can handle it just fine.
The problem TFA alludes to is that, while cities and industrial plants already have fat lines to them from the rest of the grid, windfarms are new construction generally sited in rural areas that don't already have a "fat pipe" available. So (for a wind farm bigger than about twice the local load) you have to run some new lines.
Just like you would if you built a new auto plant or aluminum smelter in the same location.
It's a regular line, just like the ones feeding loads. It just happens to be running the power the other way.
Of course some people would love to get the government to pay for the line to their new wind farm, rather than bearing that expense as part of the project. And some people in government would love to have more authority and a bigger budget. So we get FUD like this.
Electrolytic capacitors are used in the power supply filtering of essentially all circuitry currently in use. These components will be present in a number of places on the circuit boards of both disk drives and memory sticks.
The insulating layer between the plates in electrolytic capacitors is formed by electricity-driven chemical action. It gradually degrades over a period of several years. If the capacitor is operated occasionally the operating voltage across it will rebuild the insulating layer. But if it's left unused for too long the layer will degrade to the point that, when power is finally applied, the capacitor will short and (because most of them are hooked across a power supply) fail catastrophically. Like by blowing open and jetting chemical fumes, while shorting out the supply and damaging other components in the current path.
Mechanical moving parts may lose lubrication and spot-weld with time. This also makes storing entire drives problematic.
Recordible CDs usually record on a die layer that will degrade with time.
Some types of flash memory store data as stored charges, which will leak away with time.
So IMHO degradation of the medium itself is likely to be a killer problem. Much more than readability with future devices. (After only 25 years there should still be some working players available for currently widely-deployed standards, even if no new ones are being manufactured. Once you've go the bits read you can transfer them to new media.)
Things I'd consider:
- Integrated circuit memory devices using a technology like fusible link or a crystalline/amorphous transition. (Replace any electrolytic capacitors with ceramic types - which will greatly increase the size of the assembly.)
- CD masters involving actual removal of material - a material inert enough that it will not corrode away with time.
- If you want to store drives for removable media (and convert the caps), check with the manufacturers about what the bearings are like and talk with a mechanical engineer with applicable experience (like mil-spec or space-rated). I'd avoid sealed hard drives, especially those that don't lift the heads off the platters when parked.
Also: Use a coding scheme that has industrial-grade error correction. B-)
One downside to your task is that, with only a 25-year storage time, you'll probably still be alive to be blamed for failures when they open it. B-)
Got to admit, but it's amazing to think that what would normally be a hindrance "the shimmering in the heat" could become an asset.
OFDM (Orthogonal Frequency Division Modulation) schemes (which include WiFi and WiMax) do the same thing with multipath interference: Take advantage of the extra signal brought in by the multiple paths and add it all up to improve the signal-to-noise ratio.
My own limited experimentation with synthetic aperture photography strongly suggests that the "diffraction limit" is overrated.
Synthetic aperture techniques produce an analytical solution based on phase information which is partly discarded in classical optics "cameras". Their resolution doesn't have the diffraction limit - at least not in its usual form - and is mainly limited by noise - including sampling and quantization "noise" in digital techniques.
This new hack is a classical optics approach and is still subject to the diffraction limit. It beats the diffraction limit of the objective lens aperture for a different reason - the ACTUAL objective lens is the portion of the distorting region which feeds light into the device. This region is (sometimes) larger than the first lens aperture of the device, at least for part of the image. So by picking chunks of the image that are currently effectively focussed by a larger "lens" you get a larger aperture to diffraction-limit your selected chunk. Iterate until you have enough bonus-resolution chunks to cover the image plane.
I've worried about the possibility of BGP poisoning attacks myself, but only because we have a virtual monoculture...
BGP is a near-monoculture because the standards document is not adequate for buidling a working implementation, while a broken one breaks not just the box running it but much of the net (and essentially all of it that has traffic that should go to or through that box).
The protocol has negative feedback and delay, which results in oscillation ("route flapping") if nothing is done about it. The current workarounds for this are arcane and not well known.
Because deploying a broken BGP implementation breaks things so badly, ISPs are very loath to switch to boxes that don't have a proven BGP implementation. So one of the biggest barriers to entry for an equipment manufacturer is licensing customer-trusted BGP code or finding the rare talent with the occult knowledge, building their own, and then convincing the ISPs that the implementation is safe.
As a result there are VERY few BGP implementations deployed. Last time I looked - a few years back - there were three: The original (don't recall who did that one), Cisco's, and Redback's. Getting its BGP accepted was a significant part of why Redback's Smartedge product line was slow in breaking into the market.
This near monoculture also means that, in addition to the exploits inherent in the protocol, any implementation-specific security flaws will have a very large target population for exploits.
Presuming this is true: The court remedy might require unbundling the hardware and software sales.
Apple could set the cost of each wherever it wants. But it will need to keep the sum affordable to keep selling the "out of the box experience".
- If it sets the SW price so high that the HW is underpriced you'll see people buying bare Apple machines and loading other OSes on them. This also leads to people being used to running other OSes on Apple hardware.
- If it sets the HW price so high that the SW is underpriced you get people like Psystar, selling Apple-like hardware running Apple's OS.
- If it sets them both proportionally to their actual costs, but at a premium, you get both at once. B-)
It gets even more interesting if the court requires them to split into two business units.
(And of course Apple could win. Or Apple could lose but the court prescribe other remedies.)
Back in the '50s I knew a guy with a sports car who had five bat-handle toggle switches he'd mounted at the front of the console between the seats. They did things like disable the brake lights and popeye the tail and headlights. When a cop was after him he'd disable the brake lights, make a turn, switch his car's light-show appearance, and slow 'way down...
He also used aircraft landing lights for his high-beam-only lamps for a while - again with the switches. They'd light up the trees on the horizon bright blue. (Don't shine your high beams in HIS eyes!) He said he quit that after getting caught in a roadside inspection traffic stop. When they put the brightness sensor caps over his highbeams and had him turn on the lamps the sensors caught fire. B-)
Slashdot Mirror
it's a very big LED.
Oh. Of COURSE. (Blinks very slowly, right?)
It's on top of the triple-pronged antenna platform on the hill over by the Castro district.
They probably can't find it because they're re-hacking the antennas for digital TV and are moving all the equipment around.
Why do they insist on calling p2p itself illegal?
They're not just trying to flame the file-sharers.
They're trying to convert the Internet to a broadcast service (with themselves as the gatekeeping broadcasters). A centraized-producers, sea-of-consumers model.
These are the media conglomerates. They became multibillion dollar businesses in an environment where providing "content" had a high capital cost. This let them serve as gatekeepers, collecting a lot of money for the works they distributed (but still out-competing live performances due to economy of scale) and keeping most of it for themselves.
But advancing electronic and computer technology has brought the tools of mass-market "content creation" within the financial reach of the ordinary "starving artist". And the inherently peer-to-peer internet, with high bandwidth, flat rates, and cost-distributing tools like torrents, has brought the distribution costs to near-nothing. So artists without investors can bypass the gatekeeper and go straight to the customers/fans on an industrial scale.
This means that the artists can now out-compete the gatekeepers. So to protect their lucrative business they are taking a number of measures to attack, not just competition from unpaid distribution of their own material, but also this potential competition from artists escaping from their channels.
You see it here in the attempt to brand peer-to-peer as synonymous with crime. But you also see it in other places: .alt hierarchy for example.) This raises the cost of "distributing content" and reintroduces gatekeepers.
- Torrent throttling.
- Asymmetric connections with high download and low upload speeds (presupposing the ISP's customer consumes more than he produces).
- Bans on "servers".
- Legal moves to make ISPs responsible for the content of files they transfer, store, or serve. (Web sites, netnews, etc. Notice the attack on the
- And (of course) propaganda buried in the entertainment they produce. (Just as they once did crime shows villainizing video games, computer bulletin boards, and the like when they were competing with broadcast TV for eyeball time - and before that smearing cable TV operators as a class in the early days when they weren't and conglomerate-owned and were seen as pirates of the off-the-air signals and competitors purveying imported and local content.)
If they can get the general public to avoid peer-to-peer and go back to central producer / distributed (and paying!) consumer, they've won the battle.
Whenever I start contemplating DNA (!), self-reproduction and the utter insanity of how complex the machinery of a single cell is, much less multicellular life, much less an animal, much less a self-aware brain, I just shake my head in wonder.
Doesn't bother me. Evolution is a massively parallel computation and has been going on for a LONG time.
If you skip DNA and just look at RNA it all gets easy:
- RNA caries genetic information and can be copied by an appropriate enzyme. (It's less stable than DNA, but quite stable enough to form the genomes of viruses. At the early stages, with no competition yet, being error-prone is actually good.)
- RNA has enzymatic activity. (It's not as strong or as versatile as protein-based enzymens. But it is quite capable of folding itself up into structures coded by its sequence, sticking together at appropriate places and presenting controlled patterns of charges on outer surfaces of a controlled shape, to become a little molecular machine.
Nucleotides line themselves up on a strand of either RNA or DNA to form the complimentary code sequence. They'll bind themselves into a strand given enough time and jostling. But if you have a RNA strand that also sometimes folds up into a little zipper-tab that runs down the lined-up RNA bases and sticks them together into a fresh strand you're all set:
- You'll eventually have both that and its compliment hanging around in the same container.
- At some point the strand that folds up into a zipper will zip up the new bases stuck to its complimentary strand. Then you have TWO zippers tab strands plus a complimentary strand.
- Now the zipper strand(s) start churning out new zipper strands and complimentary strands.
Slow at first, because rev 0.1 probably doesn't work well and it's completely dependent on randomly occurring bases for "food". But with the exponential under way the errors start to accumulate. Now you get some that are better at zipping than others - and they dominate the regions where they occur. And you get strands with multiple copies and other noise sequences - which can now evolve separately within the strand and evolve new functions.
Whenever a strand evolves one of its "spare" "genes" into a machine to help out, it becomes more successful.
From there you can evolve:
- Machines to make components of the system from other "useless" stuff.
- Machines to string amino-acids into useful structural stuff - and eventually better machines.
- Machines to control a container, creating the "cell" and its division mechanism.
- Machines to make backup copies of the RNA code in more stable DNA and then make more RNA from that.
and so on.
There's plenty of suggestions that this is what happened. For instance: Most of the machinery of RNA-directed protein synthesis - both most of the parts of the ribosome (the stringing factory) and all of the transfer RNA (the amino-acid carrier/code reader mechanisms) are RNA enzymes.
So, no, contemplating the current complexity doesn't bother me at all. It can all be explained by evolution from a single, simple, mechanism that could easily be produced in millions of years of random abiotic chemical reactions in a planetary scale vat of solar-irradiated, weather-stirred chemicals.
IBM was merrily outfitting the Nazis with equipment to help them manage their concentration camps (completely ignorant of their application, naturally) while Bletchley park was breaking Nazi codes.
It was also manufacturing M1 rifles for the Allies. (Along with such companies as Rockola - the jukebox maker - and Saginaw Steering Gear. It's handy to convert a factory to guns when it already has equipment for drilling holes the long way down several feet of steel rod and other machines for building small and complicated devices composed of mechanical moving parts.)
(Back when I was buying an old M1 carbine for participating in the Civilian Marksmanship Program training I picked an IBM-branded one just out of nostalgia. The rangemaster was impressed when I qualified with a carbine, rather than a full-length M1, on the first try. Shorter barrels make for less accuracy. B-) But I could have used a Field Engineer: While the steel parts worked fine, the wooden barrel cover kept popping off during recoil. B-( )
The same statistical argument applies to creating DNA databases and using them to hunt for perpetrators.
A DNA mismatch is essentially unambiguous. But a hit is a statistical issue, and hard to quantify because of unknown correlation of markers due to inheritance.
So using it to eliminate suspects is basic. And using it to point to one of a small number of unrelated suspects identified by other means is plausible. But throw the DNA pattern from a crime scene at a database containing thousands, or millions, of people (but not ALL of them) and you are just asking for false positives.
And as the tests get better and the false positive rate goes down it becomes even more dangerous, because it becomes more tempting to rely on it when only one "hit" shows up. (Consider, for example, identical twins separated at birth, where one is in the database, the other commits a crime, and the poor sap that gets fingered doesn't even know the twin exists.)
And someone brought in on a false positive has an extra handicap: DNA testing won't exonerate him (unless there's enough DNA evidence that more extensive tests can be performed and it happens to find an additional marker that mismatches.)
(That was SUPPOSED to be a preview while I got the links right. Here it is with the links fixed. View them in order...)
You know, that IS impressive but... Can it figure out how to brew the 'perfect' cup of coffee?
I think you meant tea.
No, I think Xaedalus meant the 'perfect' cup of coffee.
You know, that IS impressive but... Can it figure out how to brew the 'perfect' cup of coffee?
I think you meant tea.
No, I think Xaedalus meant the perfect cup of coffee.
Whether this is good or bad depends on your point of view to a certain extent, but what is clear is that Hollywood isn't planning on abandoning DRM any time soon. As a result, they're not going to adopt any form of writable media to store digital copies unless it has some kind of DRM system built-in.
And as a second-order result of that they're going to miss the fast nickels while chasing the slow dimes.
But private enterprise is all about being free to make stupid business decisions and taking the resulting consequences on the bottom line. (And losing the fast nickels is NOT one of the consequences they'll be able to lobby their way out of.)
DRM, in this case, is a choice between DRM'd content and no content at all. I'd rather have DRM'd content than none.
I, on the other hand, would rather do with none than with DRM.
I made that choice when the DVDs supplanted videotape and didn't buy DVD movies - or buy or rent any movies at all - until after CSS was cracked and the movie industry gave up on their attempts to stuff that genie back into the bottle. No blu-ray players for me, either. Stopped buying CDs, too, when they started experimenting with the early computer-speaker-blowing "copy protection" that corrupted the data and depended on the error correction on players to recover the music (and thus corrupted it when you got real errors from a dirty disk) and never really got back into purchasing new music after that.
Never actually MADE a backup copy. And never downloaded a "pirated" song or movie, either. I just don't buy encumbered stuff.
Instead I found other ways to amuse myself. (For instance: The amazing number and variety of animals outside the place on the high desert put on a continuous show that's quite entertaining - especially when I flush the well and create a puddle that draws them in from miles around. And there's lots of amusement on the net that is not "pirated" copyrighted content.)
Interestingly, I don't really miss the corporate "content". Either the quality took a nosedive around that time or the product stopped matching my (quite broad) tastes. (Though from what I hear of some local bands it's more the former than the latter.)
We all make our choices.
... what happens when a DNS record gets updated? With the new behavior, you won't see the change until your cached record expires. ... I believe BIND was working as intended to allow updatedrecords to overwrite older ones.
That's what I thought too, at first. It seems right to replace cahced records with newer ones. And perhaps that's why the code is written as it is.
But then I got to thinking: If the server has a cached record already, it wouldn't have asked for and update. So why is the new information there in the first place?
To make matters worse, NVidia has a mound of unsold defective parts that they are going to bleed out into the channel along side of the (hopefully) fixed parts.
This sounds very similar to what finally took down Weitek, back when there were a bunch of graphics chip companies competing hotly and being shaken out if they screwed up.
Weitek had built a very fast and powerful chip. But they had goofed: While it had the mandatory basic VGA mode for acquiring the Microsoft certification, there was a bug in it.
QA told management that the bug was there and would fail them. But Software told them a driver could work around it and people would want the chip because it was so fast on graphics rendering. (Of course it could not - because to get the cert it had to work with the stock bootstrap stuff, before a custom driver could be loaded.)
So they went to production with the bug. And the customers got their prototypes, found the bug, and demanded a fix. Eventually they did a fixed version - but had maybe a couple million of the buggy ones on hand and wouldn't sell the fixed ones unless the customer bought some buggy ones, too. So nobody bought and the company folded.
It's clear they're using a viral capsule for delivery - from a virus that normally targets the cell type the want to transform.
But I'm curious as to what they're delivering:
1) The transcription regulatory proteins?
2) Messenger RNA coding for the proteins (resulting in limited production of the proteins until the RNA "times out").
3) DNA (or RNA plus a RNA-directed DNA polymerase) to add new code to the cell iteslf?
4) A modified virus deleting most of the virus code and adding code for the proteins?
I'd be happy with 1) or 2). Either looks ideal. (2 looks especially easy to generalize, using any appropriate RNA virus to supply a capsule aimed at a suitable target and stringing RNA for the proteins together with the virus' "pack-me-up" tag and some "cut me here" introns). Just make sure you've sorted out real viral DNA (say, by making just the desired proteins in e-coli and assembling things after purification).
3) and 4) have some nasty downsides.
Perhaps someone with access to the article can tell us which it is, please?
If the answer is 1) or 2) I'll do a happy dance. B-)
What people forget is Rome 'fell' for a very, very long time and 'fell' for a number of reasons.
In fact the last stage of the fall was the termination of the reign (and in one case the line) of last two claimants to the title of "Caesar": the Czar of Russia and the Kaiser of the German Empire.
That, and, IIRC a federal law that obliged it to finance research with 50% of its profits in exchange for the monopoly.
As I recall it wasn't a requirement that they do research. It was permission to include the cost of research related to telephony in their cost of doing business - on which they got to set monopoly phone rates so they got a specified rate of return (6% if I recall correctly).
The result was that the more money they spent on research, the more profit they made.
So they set up Bell Labs to spend as much money as possible, on anything even vaguely related to telephony.
And it "was an abysmal failure". From year one they made more money on the results of the work (by things like licensing patents) than it cost to run the labs. So basic research was profitable all by itself. B-)
But this counterintuitive effect also has a counterintuitive downside. The rewards for a research project start once it's done and keep coming in for quite a while after it's finished. Most of us would consider this good. But the Harvard Business School approach to management comes into play: The incentive structure on managers is to show as big a profit as possible for a few years and move on, thus looking better than your predecessors and successors and getting progressively better paying positions. So by killing the CURRENT research and just collecting on the results of the previous work they can cut their costs to near nothing while the benefits keep rolling in. For a while. Then they move on. Without new work the revenue gradually dries up and their successors take the rap. (And their successors would have to increase costs while the income was ramping down, which would look even worse, to turn things around.)
Regardless:
Without the guaranteed profit they're in the same boat as every other large cashflow company in the world. Perhaps basic research would continue to be profitable beyond the dreams of avarice. But there are other profitable things to do with the money where the return is more visible in advance, rather than crapshooting on what basic research might come up with. So (like all those other companies), the new generation of management reacts to the new situation by doing the standard thing - which doesn't include basic research.
(And it doesn't help that they already went through the "cut expenses and look good on the return on old work" phase a few years back. IMHO this is the house of cards coming down.)
= = = =
As I understand it, Xerox PARC had something similar going on but for a different reason: a strange accounting system.
One of the first things PARC did was to design a "new control panel" (and brain) for Xerox copiers. This replaced a bunch of relay logic with a microcomputer/early logic chips. And that saved a LOT of money.
PARC got credited with that savings on all the copier products sold from then on (and with similar stuff it did later). So it could spend money hand-over-fist on whatever it wanted and still look profitable.
(This was the same accounting department that, if I've got THIS right, screwed up big time when Xerox went into the mainframe business as the first company to take on IBM's core business, 'way back in the early days of "foreign attachments" opening up the IBM big-iron market. They built a CPU. After a while they decided that they were in the red on it big time, folded the division, and sued IBM on antitrust. In those days equipment was all leased. As a result of the suit IBM got hold of their accounting info and discovered the hadn't really understood how to interpret lease income. They were actually VERY profitable, and had folded the division because of this accounting screwup. Of course this discovery folded the suit.)
Wind power does vary with weather. But at good sites the wind is primarily driven by the "lake effect" and is quite consistent. ("Lake effect" is a bit misleading: At the Altamont Pass wind farm, for instance, the "lake" is the Pacific Ocean and the "island" is California's Central Valley. Similarly, in Texas the "lake" is the Gulf of Mexico and the "island" is the Great Plains - much of the North American continent.)
Lake effect winds highest in the afternoon and evening, a reasonably good match for the peak electrical demand curve.
The multiple mills with significant separation that constitute a large wind farm average out gusts and even storm cells. As to weather changing the speed also affects load because much of load is driven by HVAC (Heating, Ventilation, Air conditioning) and higher wind means insulation is less effective, resulting in more heat gain or loss that must be made up by the HVAC system.
So, though the amount of power generated is not as stable as that from a reservoir full of water, a pile of coal, or a nuclear reactor, the variability of wind power actually REDUCES the amount of standby and peaking power required from the other generation sources on the grid.
(Solar has similar characteristics, with the peak leading that of the demand curve by a bit. Its peak, and variability, track the air conditioning load very well but run counter to the (smaller) heating curve. Wind power is larger for a given investment and goes up with weather that imposes extra heating load. So a mix of the two is even better at evening out the load on the rest of the grid's generation.)
'The' Saturn plant? I guess you don't realize this is a forum that covers more than just your neighborhood.
I'm referring to the original Saturn Corporation plant, in Spring Hill Tennessee, which is decidedly NOT in my back yard.
GM's choice of site was international news - including an editorial cartoon showing a ringed planet crashed into a farmer's field in a mountainous rural setting.
That's just the thing - they don't build new auto plants or aluminum smelters out the hell in the middle of nowhere for just that reason.
I guess you don't remember the Saturn plant, eh? (Just for starters.)
ANY time you build a new plant - unless it's on the site of an old one - you'll need to string a new high line and a bunch of other infrastructure. The sites where the land is cheap enough to make it economically feasible will not normally have it already available. And the cost of installing it is small compared to the difference in land price between building in an existing urban area and building in something more open.
Ditto with building a new power plant, of course. (Even for a new peaking plant in an urban area you still have to string something. The chunk of the grid that powers a warehouse won't handle even a tiny power plant. If you're in the SF Bay area think about the miles of high-line they ran along I-880 to support the plant near the CA 237 intersection, just for a little gas-fired plant.)
Wind farms are just a special case of this. Like hydropower dams, they have to be put where the resource is. And this will be somewhere that the existing power transmission lines aren't big enough to handle the energy harvested.
The summary is a crock and doesn't match the quoted article.
Transporting large and variable amounts of generated power is the dual of feeding large and varying loads. The power grid can handle it just fine.
The problem TFA alludes to is that, while cities and industrial plants already have fat lines to them from the rest of the grid, windfarms are new construction generally sited in rural areas that don't already have a "fat pipe" available. So (for a wind farm bigger than about twice the local load) you have to run some new lines.
Just like you would if you built a new auto plant or aluminum smelter in the same location.
It's a regular line, just like the ones feeding loads. It just happens to be running the power the other way.
Of course some people would love to get the government to pay for the line to their new wind farm, rather than bearing that expense as part of the project. And some people in government would love to have more authority and a bigger budget. So we get FUD like this.
Electrolytic capacitors are used in the power supply filtering of essentially all circuitry currently in use. These components will be present in a number of places on the circuit boards of both disk drives and memory sticks.
The insulating layer between the plates in electrolytic capacitors is formed by electricity-driven chemical action. It gradually degrades over a period of several years. If the capacitor is operated occasionally the operating voltage across it will rebuild the insulating layer. But if it's left unused for too long the layer will degrade to the point that, when power is finally applied, the capacitor will short and (because most of them are hooked across a power supply) fail catastrophically. Like by blowing open and jetting chemical fumes, while shorting out the supply and damaging other components in the current path.
Mechanical moving parts may lose lubrication and spot-weld with time. This also makes storing entire drives problematic.
Recordible CDs usually record on a die layer that will degrade with time.
Some types of flash memory store data as stored charges, which will leak away with time.
So IMHO degradation of the medium itself is likely to be a killer problem. Much more than readability with future devices. (After only 25 years there should still be some working players available for currently widely-deployed standards, even if no new ones are being manufactured. Once you've go the bits read you can transfer them to new media.)
Things I'd consider:
- Integrated circuit memory devices using a technology like fusible link or a crystalline/amorphous transition. (Replace any electrolytic capacitors with ceramic types - which will greatly increase the size of the assembly.)
- CD masters involving actual removal of material - a material inert enough that it will not corrode away with time.
- If you want to store drives for removable media (and convert the caps), check with the manufacturers about what the bearings are like and talk with a mechanical engineer with applicable experience (like mil-spec or space-rated). I'd avoid sealed hard drives, especially those that don't lift the heads off the platters when parked.
Also: Use a coding scheme that has industrial-grade error correction. B-)
One downside to your task is that, with only a 25-year storage time, you'll probably still be alive to be blamed for failures when they open it. B-)
Got to admit, but it's amazing to think that what would normally be a hindrance "the shimmering in the heat" could become an asset.
OFDM (Orthogonal Frequency Division Modulation) schemes (which include WiFi and WiMax) do the same thing with multipath interference: Take advantage of the extra signal brought in by the multiple paths and add it all up to improve the signal-to-noise ratio.
My own limited experimentation with synthetic aperture photography strongly suggests that the "diffraction limit" is overrated.
Synthetic aperture techniques produce an analytical solution based on phase information which is partly discarded in classical optics "cameras". Their resolution doesn't have the diffraction limit - at least not in its usual form - and is mainly limited by noise - including sampling and quantization "noise" in digital techniques.
This new hack is a classical optics approach and is still subject to the diffraction limit. It beats the diffraction limit of the objective lens aperture for a different reason - the ACTUAL objective lens is the portion of the distorting region which feeds light into the device. This region is (sometimes) larger than the first lens aperture of the device, at least for part of the image. So by picking chunks of the image that are currently effectively focussed by a larger "lens" you get a larger aperture to diffraction-limit your selected chunk. Iterate until you have enough bonus-resolution chunks to cover the image plane.
I've worried about the possibility of BGP poisoning attacks myself, but only because we have a virtual monoculture ...
BGP is a near-monoculture because the standards document is not adequate for buidling a working implementation, while a broken one breaks not just the box running it but much of the net (and essentially all of it that has traffic that should go to or through that box).
The protocol has negative feedback and delay, which results in oscillation ("route flapping") if nothing is done about it. The current workarounds for this are arcane and not well known.
Because deploying a broken BGP implementation breaks things so badly, ISPs are very loath to switch to boxes that don't have a proven BGP implementation. So one of the biggest barriers to entry for an equipment manufacturer is licensing customer-trusted BGP code or finding the rare talent with the occult knowledge, building their own, and then convincing the ISPs that the implementation is safe.
As a result there are VERY few BGP implementations deployed. Last time I looked - a few years back - there were three: The original (don't recall who did that one), Cisco's, and Redback's. Getting its BGP accepted was a significant part of why Redback's Smartedge product line was slow in breaking into the market.
This near monoculture also means that, in addition to the exploits inherent in the protocol, any implementation-specific security flaws will have a very large target population for exploits.
Under what law could a court force Apple or anyone else for that matter to sell a particular product to any customer?
Sherman antitrust act for starters.
OS X is subsidized by hardware sales.
Presuming this is true: The court remedy might require unbundling the hardware and software sales.
Apple could set the cost of each wherever it wants. But it will need to keep the sum affordable to keep selling the "out of the box experience".
- If it sets the SW price so high that the HW is underpriced you'll see people buying bare Apple machines and loading other OSes on them. This also leads to people being used to running other OSes on Apple hardware.
- If it sets the HW price so high that the SW is underpriced you get people like Psystar, selling Apple-like hardware running Apple's OS.
- If it sets them both proportionally to their actual costs, but at a premium, you get both at once. B-)
It gets even more interesting if the court requires them to split into two business units.
(And of course Apple could win. Or Apple could lose but the court prescribe other remedies.)
Back in the '50s I knew a guy with a sports car who had five bat-handle toggle switches he'd mounted at the front of the console between the seats. They did things like disable the brake lights and popeye the tail and headlights. When a cop was after him he'd disable the brake lights, make a turn, switch his car's light-show appearance, and slow 'way down...
He also used aircraft landing lights for his high-beam-only lamps for a while - again with the switches. They'd light up the trees on the horizon bright blue. (Don't shine your high beams in HIS eyes!) He said he quit that after getting caught in a roadside inspection traffic stop. When they put the brightness sensor caps over his highbeams and had him turn on the lamps the sensors caught fire. B-)