"The desktop market doesn't need a 32-bit processor" - Sir Clive Sinclair, on why the Sinclair QL used an 8 bit processor.
Let us examine the 8-core thing, though. Do you really need eight cores? In essence, 8 cores allows you to run 7 low-latency hard real-time threads, plus everything else you might want on the 8th core.
Is this a realistic scenario? Remember, not many people do much that is hard real-time. You want that when timing must be absolutely perfect. Well, an obvious case where you would want hard real-time is when doing high-quality multimedia. You want the audio and the video to be absolutely in sync, which means that you must have a totally predictable throughput for each. That's only two hard real-time threads, which would require a minimum of 3 cores.
So we're already above Intel's suggested needs for desktops. Can we manufacture - however unlikely - a scenario where we'd need far more?
Let's take that multimedia experience and wrap it around a MMORG. In order to keep the multimedia in phase with the game, the client engine must also be hard real-time. Low latency would make the experience surreal if what you saw had no audible results for several seconds or minutes. You need a guarantee that when action A happens on screen, action B happens over the speakers, within a highly deterministic timeframe. Having it variable and unpredictable would lose the feel. This brings us to 3 real-time threads and therefore 4 cores.
Virtual world servers tend to do a lot more computation than necessary and therefore transmit far more than required. By having a physics engine on the clients handle the microscale change of states, whilst the server calculates macroscale changes, you can reduce the I/O to simply re-synching when cumulative errors exceed some value. Again, to keep the visual and audible effects in time with the engine, you'd need a real-time physics engine on the client. This brings us to 4 real-time cores and 5 total.
MMORGs would be much more interesting if they used virtual reality gear or CAVE systems. The I/O from these, though, is intense. If you could get away with only having one core for input and another for output, you're doing well. This gives us 6 real-time cores and 7 total.
"Now, wait a moment! Why would you need an entire core to do hard real-time?"
Hard real-time is, well, hard. I'm stretching the definition of real-time here to not just mean that there is a guaranteed interval of time per larger block that the process has to run, but also that this interval is of completely fixed size. It can neither shrink nor expand. I'm also stretching it to mean that the interval occurs at a predetermined time, so it cannot occur early or late. In the case of video, events would be at fixed time intervals and would operate at a fixed frequency. In the case of sound, background would be at fixed time intervals and foreground would be controlled by a trigger/timer mechanism.
Obviously, on a single core, you can't provide these kinds of guarantees, even if there is enough CPU power to technically provide the processing time. There are far too many interrupts, locks and priority shuffles to do more than give the crudest of assurances. Games on such machines don't look smooth, because they cannot be smooth. The best they can ever get is fast enough to hide the crudity.
The simplest way to avoid the problems is to farm out the stuff that has very rigid time requirements and can't afford the unpredictability of being swapped out by equal-or-higher priority but ill-determined tasks. You also run into the problem that you do NOT want to swap between real-time tasks if you can help it. This isn't because they'll interfere (although, if the total time they'd need exceeds the total time available, they will), but because you need these threads to occur simultaneously to fool the brain correctly.
eg: Sword hits stone, at a distance of N meters, where N is very small.
GateD used to be under a semi-open license. Then there was MRTD, Zebra and Quagga. XORP is said to be pretty good, too. MIT's Click is probably the most versatile, as you can just about script your own routing elements - very pluggable - with the added capability of routing between physical and simulated (eg: NS-2) networks.
...they buy "world-class support", but having tried to use said support on occasion, I can say that I feel sorry for the world. Sure, it's better than a kick in the head, but not so much that it's worth the cost. I believe the record for longest repair ever was at the University of Manchester, in England, where a Cisco router corrupted the 1518th byte in every packet (thus only corrupting packets with a 1500 byte payload or 1496 bytes over 802.1q). Took them NINE MONTHS to fix. The first three of those, they denied there was even a problem.
...denounce all red-level troubleshooters as traitors to the Computer, on the grounds that there is not a single one of them who has handed in traitors to the Computer, which is highly suspicious.
Well done, friend citizen. You have been promoted to ultra-violet. Please report to the nearest termination center immediately.
You've been reading "Fractal Geometry of Nature" by Benoit Mandelbrot. Very nice illustrations and the section on how fractals all started and another on fractal dimensions were good, but otherwise the book was far too vague and had few proofs. This demonstrates Heisenberg's Writing Principle, which states that you can either know bout a topic or write about it, but not at the same time.
mtDNA is shorter and usually more stable than nucleic DNA, so I wouldn't rule out the possibility of finding some. They don't exactly need the whole molecule, just enough to be able to determine where in the mtDNA strand the segment belongs -and- get a good idea of the extent of the changes.
Even if they can just get proteins, provided it's a random sample and statistically significant, they may be able to deduce things about the nucleic DNA. They'd know the proportions of the proteins in dinosaurs, the proportions in likely candidates for nearest living relatives now, and the DNA coding for those proteins in those living relatives. From that, they could deduce, using reverse-engineering techniques, what changes would have been needed to go from the ratios of the past to the ratios of the present.
The "red book" that defined the CD standard was the main reason CDs were as interchangable as they were. DVDs, which have no such rigorous standard to meet, tend to be less predictable.
IPv6 hasn't got an "official" centrally-run test suite, but such suites for the purpose of certification do exist.
C follows standards, rather than official regression tests, so you do get a lot more variation in quality and interchangability. In general, though, it demonstrates that simply defining a central standard is adequate for basic stuff.
POSIX used a mix of standards and official testing, and had a big impact on the interoperability of Unix systems - though nowhere near as much as seems to have been hoped.
Standards and tests by committee seem to be a Very Bad Idea. CORBA and SQL followed this road and have resulted in guaranteed inefficiency - apparently for the purpose of promoting the "extended" packages released by the vendors who created the standards in the first place.
If the process is followed - hey, that's great, but is needs to be done right if it is to be worth the doing.
TCP is based on packet acknowledgement and it is very doubtful that spammers have thought to check their software for deadlocks or timeouts. Instead of dumping the data, just have the connection hang after it is fully established, or send deliberately malformed acknowledgement packets. The idea here is to try and crash the zombie by either running it out of resources or giving it replies it can't handle.
Alternatively, if the spammer/zombie computer has port 25 open itself, have a netfilter rule that rewrites the destination address to that of the sender, increases the TTL, and sends the packets back in duplicate. Again, this is a resource-draining scheme. If it's an open relay, it'll get the spam and resend it. I believe the hop count for SMTP is something like 30 and each packet will go two ways along the wire, so it'll take 2^31 as much bandwidth overall, if a sufficiently large number of users set up this kind of loopback. Companies that simply don't care if their machines are zombies will suddenly notice a degradation of their networks but any packet monitoring they do will show all of the packets to have the IP addresses of their machines for both source and destination. At least some will zombie detox to save their sanity.
Hmmm. The only possible overlap is in the fabrication. Designing a good graphics processor is going to be very different from designing a good CPU, so they can't overlap the development teams (which will likely be small anyway). It's very doubtful the chips would be of similar enough size and have similar enough characteristics to do much about packaging or testing. Unless they're planning on unifying the scale at which they're making the chips, it's not clear they could do much about the etching. They could buy the materials jointly, thus increasing bulk orders and reducing costs, but they could do that with a simple agreement.
Management is a fairly big expense, but as the total number of projects wouldn't change significantly, neither would the number of managers required. That just leaves the board of directors. Half the directors could be fired, but it's doubtful either CEO is going to consider their choice of senior management to be the inferior choice. Which means that one board would win and the other board will lose. On the whole, that is. The CEO of the winning board might cherry-pick a few who are really exceptional or who have given him lots of money.
You also need to bear in mind that CPU sales for AMD are (on average) rising but their profit margins are slumping, so if they gain access to another fab plant, it won't be to close it. It'll be for continuing in a price-war against Intel that both companies are losing. (Neither has the resources to continue until the other is completely vanquished AND remain competitive with other CPU manufacturers. Both Intel and AMD are latecomers in both the multi-core and 64-bit arenas, and neither can match the more experienced players on scalability at this time.) However, neither AMD nor Intel can afford to back off - their designs are divergent enough that the market cannot sustain both of them indefinitely. Intel can't even afford to maintain the StrongARM architecture anymore, things are so tight.
I'll add two more, so the list is at least 4: SiS and VIA both make GPUs. Does S3 still exist? Also, the MediaGX (rebadged to the Geode) has a GPU core built into the CPU, so that sort-of counts.
Open Cores, Sourceforge and Slashdot should get together to see if they can jointly buy OpenGL. SGI'll probably take anything at this point, most vendors already have OpenGL implementations of their own and don't need anything SGI still has, and I'd rather trust CowboyNeil with the specs than most of the vendors out there.
(Can you imagine what would happen if Microsoft bought it? Does anyone seriously believe ANY implementation would be safe, MESA included?)
Failing that, Google must have some spare change. Hell, they could probably buy SGI for less than the value of the machines in SGI's inventory, which would seriously boost their server power.
If it's ATi trying to buy out AMD (which is perfectly possible), then they might not have enough money left to stop nVidia doing a hostile takeover of them both. That would eliminate one of nVidia's competitors -and- give them control over the CPU that looks set to take over.
You need to bear in mind that the GPU is the critical component in most systems, but makes almost no money for the vendor and has a relatively low volume. There is precisely no reason whatsoever for AMD to want to merge with ATi or to buy them up. That would be expensive and earn them little. In fact, given how much they've made from their component-neutrality, sacrificing that might mean they'd actually lose money overall.
On the other hand, CPUs are high volume, high profit, and AMD is gaining market-share. It is an ideal target for a buy-out, particularly as ATi can't be doing that well in the GPU market. Buying AMD would be like buying a money-printing-machine, as far as ATi were concerned. Better still, AMD is a key player in bus specifications such as HyperTransport, which means that if ATi owned AMD, ATi could heavily influence the busses to suit graphics in general and their chips in particular.
(Mergers are never equal, as you don't have two CEOs, two CFOs, etc. One of them will be ultimately in charge of the other.)
If the rumour is correct, then don't assume AMD is the one instigating things - they have the most to lose and the least to gain - and don't assume either of them will be around when the mergers and buyouts finish.
B is exactly the sort of thing I was thinking of. There's plenty of known markers on human DNA and chimp DNA, as those are completely mapped, and some of those must be usable as targets for the sort of process you're thinking of. It would seem far more logical to do it that way than to map from scratch, knowing nothing.
Since you know the biochem side, you might want to write up a paper on how you'd go about this. At worst, you might easily get published, and journals pay very decently. At best, a biotech company that's a rival to 454 decides the idea would be both faster and more accurate, decides it's worth a shot to slaughter a rival, and buys the idea from you.
LaTeX might well be usable directly by some POD publishers. PDF certainly can, and DVI-to-PDF converters come with most LaTeX distributions. LaTeX has one book format provided as standard and there are MANY other templates from CTAN (a TeX distribution system that inspired Perl's CPAN).
It would be good if LaTeX 3 ever got released, but the mailing list is silent and the website suggests nobody has done any signifcant development in years. If LaTeX 3 progress remains dead, I'd say fork the development tree as it stands and produce a variant with the features desirable in modern desktop publishing. Nothing to stop anyone, and LaTeX 2e came about precisely because people thought the original version 2 was crap and went on to make their own derivatives.
I doubt, for example, any DTP software out there has meaningful support for high-contrast colour formats such as OpenEXR or JPEG2000. Printers are probably not up to it yet, but who is going to add on a complex feature nobody can use? The software is much easier to change, so it is much easier to create the demand first.
(In fact, LaTeX' bitmap support is truly pathetic. True, LaTeX is designed to be totally scalable, which means vetors are always preferable, but there are many illustrations that simply can't be done that way.)
LaTeX' handling of subsections is also very poor and works in absolutes rather than relatives. Sure, not many people want to nest more than, oh, three or four deep. But plenty of technical texts can nest much deeper than that, and LaTeX has no provision for it, because of the way you have to name the specific depth you're working in. (This also makes it much harder to develop things section by section, as you have to be very cunning to make something viewable at an arbritary depth.)
Finally, you've got to process your fonts to work at a specific DPI, at the time of display or printing. "But won't I know this?" Only if you're using a pixel-based printer. What happens when someone produces a laser printer that can handle vectors as well as pixels? What's the DPI of a vector, if the laser has two degrees of freedom and can control both motors simultaneously? For that matter, what if I want to output to a plotter, where I already have this capability? Vector monitors aren't in widespread use any more, but they do exist. If the software is designed to work with such methods, shouldn't it be outputtable to such devices?
(LaTeX editors exist now, which was my big grudge for a long time, although they're nowhere near the point of, say, Ventura Publisher. Yet, anyway. This, despite the fact that TeX is already a very powerful engine.)
You are correct, because they sequence directly. This isn't necessary for Neandertal DNA as we have plenty of reference points from chimp DNA and human DNA, which are already mapped. We only need to sequence those segments that are different from either.
There is an added problem. Most geneticists use chain termination sequencing, which is good for fairly decent lengths of DNA. 454 uses pyrosequencing which is faster but only good for much smaller lengths. When the unknown elements may contain repeats of unknown length, smaller sample sizes are a Bad Idea. If the sample size is smaller than the repeat, you will be incapable of knowing how long the repeat is and will therefore have multiple equally valid solutions to the analysis. This is Bad, since it is these differences the researchers are looking for.
There is an alternative technique (primer walking/chromosome walking) which works on much larger fragments and should be able to deal with the indeterminacy problem completely.
However, this all assumes that we need to use an existing method at all. We have two known points of reference (humans and chimps) and a ready supply of DNA from both. ALL we need to determine is what segments of Neandertal DNA do not match with either, and then sequence just those segments. What would be needed, then, is a genetic version of diff, as opposed to a genetic version of cat. DNA is ideally suited to producing diffs, because you can't connect non-matching bases, but is much harder to cat.
I'd say that it would make far more sense to do the fundamental research needed to develop new techniques that can exploit the fact that we have multiple known related species with mapped DNA, where the exact relationship with Neandertal is also known, but the differences created by that relationship are not. We have a gigantic library of knowns, which we can use to simplify the process. Instead, 454 is going to use an entirely random, utterly unsuitable method, simply because it's faster than developing a new method.
If people always stuck to the tools they already had, to avoid developing something more appropriate to the problem, we'd still be using stone tools and the closest to blogging anyone could get would be to find a really, really big wall to paint on.
I didn't say it wasn't used, I said it makes it more complicated. Which, if you read it, is exactly what the Wikipedia article says. If you'd read on, you'd also know that 454 uses pyrosequencing for the components, which vastly complicates genome assembly as it uses considerably shorter stands than chain termination.
Twelve times coverage was used for human DNA, to deal with repeats which are damn-near impossible to place using the shotgun method and chain termination in a single pass. With the added problems of pyrosequencing, you might easily need four to five times as much coverage to get a reliable measurement. Neandertals had a lot longer to develop such repeats, so there's an excellent chance that far greater coverage again will be required to accurately determine where such repeats are and how long they are. Humans have had one tenth of the time to evolve distinctly from chimps than Neandertals, so the complexity created by repeats could easily be ten times as great. Throw in the contamination (which apparently requires 20 times the number of samples to eliminate) and you're talking 12,000 runs. Where, exactly, are these researchers going to find 12,000 Neandertal bones?
Also, bear in mind that the human genome project largely worked off the DNA of one person (one of the project directors, IIRC), so all DNA samples would produce the same repeats. Here, they're talking about using multiple sources of indeterminately different origins - both in geography and in time. The odds of the repeats being identical is almost (but not quite) nil.
In consequence, there will be multiple possible solutions, unless they use a sample size comparable to the entire Genographic Project's database (and they only looked at 12 STRs, not the entire nucleic DNA sequence). I'm not sure there are that many distinct examples of Neanderthals known, never mind available for crushing.
Sure, it worked on cave bears and mammoths to a degree (but we've no idea what that degree is), but I'd be willing to bet that the complexity of junk DNA for cave bears is far lower than for Neandertals anyway, and that their practice runs covered a geographically and temporally smaller region, vastly simplifying the process. Also, a lot of the really good samples of mammths in Siberia are frozen, so contamination and natural disintegration would have been far less significant.
Primer walking (also on Wikipedia, and a perfectly good alternative method to Shotgun Sequencing) would seem to be more promising. We know where common elements are in human and chimpanzee DNA, so we can very easily walk the chromosomes consecutively because we have a notion of what consecutive means. Random sampling eliminates all of the useful information we already have that could eliminate noise introduced by the method.
Besides which, what fool would limit themselves to existing technologies? Here is a perfectly good opportunity to produce entirely new methods of sequencing DNA when similar (but not identical) DNA exists. Existing methods can't exploit existing data, and we don't need to specifically know actual values, all that is of interest is the diffs. Existing methods, therefore, give us the wrong information from which we then have to calculate the information that is of actual interest. It follows that we should have no interest in the limitations of the shotgun method because that is not the method we should be using.
With the original sequencing of human and chimpanzee DNA, there was no existing data and therefore no known quantity to meaningfully diff against. You could also obtain a very large number of samples with highly predictable variations. It was therefore not only possible to use the shotgun method, it was the optimal method to use.
Here, we don't have that. We have a very small number of samples with no predictability in the variations, but which are likely to be extremely similar to very well-known sequences. It is insane to start entirely from scratch, in
Definitely interesting, highly contradictory though. The blog directly linked to claims that the neandertal DNA is being found in the bacteria - that the bacteria had somehow made it a part of its own DNA. This seems highly improbable. Bacterial DNA can do strange things, but absorbing large chunks of neandertal DNA is almost certainly not one of them.
The other descriptions imply that it's contamination through questionable extraction techniques - they're grinding up the fossils, so ALL the DNA in the sample will be mixed together, and strands may well end up getting broken, making it much harder to sequence correctly.
Sequencing fossil DNA is certainly possible, and is extremely desirable, but the approach seems... odd. The BBC article, for example, claims that they're going to look for the genes that differentiate modern humans from neandertals, such as mental capacity. Given that we don't fully understand what "mental capacity" actually means, or indeed what mental capacity neandertals actually had, they would need to be looking for an unknown difference to identify an entirely theoretical and totally unquantifiable distinction. That's not good science.
Lastly, we know from studies of neandertal mtDNA that there was a large genetic diversity. Far larger than had been suspected, prior to that study. If these scientists are taking neandertal nucleic DNA from significantly different regions and/or times, they cannot be certain that the nucleic DNA had not evolved or otherwise differed to the point where direct comparison or simple in-lining of the genes would make no sense whatsoever.
This is a good research project, but I am highly uncertain of their methods and am not convinced it will yield meaningful results. Because repeat studies will be difficult to do, this is an area where those involved HAVE to take extra care to put their results beyond question. This care is NOT being taken, based on what I'm seeing.
First they colonised Europe. Then they colonised Russia. They left America to last. (The Irish beat them by 500 years, though - Brenden the Navigator was the first European in America. Well, aside from the guy who left that fossilized skull...)
Getting to Mars, though - that was easy. You load up some berserkers with drugs until they're sky high, then explode some distilled mead to launch them across the void.
"If you've not commited any crimes with your friend, you won't have anything to worry about if I ask your friend if he's commited any crimes with you", which does reduce to the grandparent post's phrasing. Basically, the judge is daring the Government to either let the case through (and risk disclosure) -or- be found guilty of lying.
Since the Government isn't a defendent, and as the US has no meaningful concept of "contempt of court" or perjury, the court can't do anything about it if the Government is found guilty of lying. On the other hand, this is election year, which is not a good year to be found guilty of anything, even if there is nothing the courts can do.
My guess is that the Government will do anything and everything to stall proceedings, such that if there is a trial, there's absolutely no risk of anything embarassing being said before polling day. If they're in power, they can clean things up afterwards. If they're not, it's no longer their problem.
Processes are being developed that could make titanium as cheap as aluminium. It's a very common metal, just very hard to extract at the moment. Gold - maybe, although it has lost a lot of value over the past few centuries and shows no real signs of improving in value.
Now, Osmium (roughly 6 times as valuable as gold) is a definite candidate for a precious metal, but couldn't be used ornamentally as it is highly toxic. Oil, as it is being consumed many millions of times faster than it is being generated, could potentially be a candidate as a precious substance.
Some of the freakier minerals are also candidates. Probably the freakiest I know of is a blue feldspar called "Blue John" that only naturally occurs in about a mile radius of the town of Castleton in the northwest of England. Because it is highly chaotic in nature (its absolute non-uniformity is part of its uniqueness and is what people like about it) it would be hard to reproduce in a laboratory. As such, it is pretty much guaranteed to remain extremely rare.
Speculating on future tastes, then, my guess is that non-uniformity and asymmetry will play a major role in the future.
Slashdot Mirror
It's two to the power of the number of ghosts the Senator snagged before losing his last life.
Let us examine the 8-core thing, though. Do you really need eight cores? In essence, 8 cores allows you to run 7 low-latency hard real-time threads, plus everything else you might want on the 8th core.
Is this a realistic scenario? Remember, not many people do much that is hard real-time. You want that when timing must be absolutely perfect. Well, an obvious case where you would want hard real-time is when doing high-quality multimedia. You want the audio and the video to be absolutely in sync, which means that you must have a totally predictable throughput for each. That's only two hard real-time threads, which would require a minimum of 3 cores.
So we're already above Intel's suggested needs for desktops. Can we manufacture - however unlikely - a scenario where we'd need far more?
Let's take that multimedia experience and wrap it around a MMORG. In order to keep the multimedia in phase with the game, the client engine must also be hard real-time. Low latency would make the experience surreal if what you saw had no audible results for several seconds or minutes. You need a guarantee that when action A happens on screen, action B happens over the speakers, within a highly deterministic timeframe. Having it variable and unpredictable would lose the feel. This brings us to 3 real-time threads and therefore 4 cores.
Virtual world servers tend to do a lot more computation than necessary and therefore transmit far more than required. By having a physics engine on the clients handle the microscale change of states, whilst the server calculates macroscale changes, you can reduce the I/O to simply re-synching when cumulative errors exceed some value. Again, to keep the visual and audible effects in time with the engine, you'd need a real-time physics engine on the client. This brings us to 4 real-time cores and 5 total.
MMORGs would be much more interesting if they used virtual reality gear or CAVE systems. The I/O from these, though, is intense. If you could get away with only having one core for input and another for output, you're doing well. This gives us 6 real-time cores and 7 total.
"Now, wait a moment! Why would you need an entire core to do hard real-time?"
Hard real-time is, well, hard. I'm stretching the definition of real-time here to not just mean that there is a guaranteed interval of time per larger block that the process has to run, but also that this interval is of completely fixed size. It can neither shrink nor expand. I'm also stretching it to mean that the interval occurs at a predetermined time, so it cannot occur early or late. In the case of video, events would be at fixed time intervals and would operate at a fixed frequency. In the case of sound, background would be at fixed time intervals and foreground would be controlled by a trigger/timer mechanism.
Obviously, on a single core, you can't provide these kinds of guarantees, even if there is enough CPU power to technically provide the processing time. There are far too many interrupts, locks and priority shuffles to do more than give the crudest of assurances. Games on such machines don't look smooth, because they cannot be smooth. The best they can ever get is fast enough to hide the crudity.
The simplest way to avoid the problems is to farm out the stuff that has very rigid time requirements and can't afford the unpredictability of being swapped out by equal-or-higher priority but ill-determined tasks. You also run into the problem that you do NOT want to swap between real-time tasks if you can help it. This isn't because they'll interfere (although, if the total time they'd need exceeds the total time available, they will), but because you need these threads to occur simultaneously to fool the brain correctly.
eg: Sword hits stone, at a distance of N meters, where N is very small.
GateD used to be under a semi-open license. Then there was MRTD, Zebra and Quagga. XORP is said to be pretty good, too. MIT's Click is probably the most versatile, as you can just about script your own routing elements - very pluggable - with the added capability of routing between physical and simulated (eg: NS-2) networks.
...they buy "world-class support", but having tried to use said support on occasion, I can say that I feel sorry for the world. Sure, it's better than a kick in the head, but not so much that it's worth the cost. I believe the record for longest repair ever was at the University of Manchester, in England, where a Cisco router corrupted the 1518th byte in every packet (thus only corrupting packets with a 1500 byte payload or 1496 bytes over 802.1q). Took them NINE MONTHS to fix. The first three of those, they denied there was even a problem.
Well done, friend citizen. You have been promoted to ultra-violet. Please report to the nearest termination center immediately.
Ooops, wrong universe.
You've been reading "Fractal Geometry of Nature" by Benoit Mandelbrot. Very nice illustrations and the section on how fractals all started and another on fractal dimensions were good, but otherwise the book was far too vague and had few proofs. This demonstrates Heisenberg's Writing Principle, which states that you can either know bout a topic or write about it, but not at the same time.
Even if they can just get proteins, provided it's a random sample and statistically significant, they may be able to deduce things about the nucleic DNA. They'd know the proportions of the proteins in dinosaurs, the proportions in likely candidates for nearest living relatives now, and the DNA coding for those proteins in those living relatives. From that, they could deduce, using reverse-engineering techniques, what changes would have been needed to go from the ratios of the past to the ratios of the present.
If the process is followed - hey, that's great, but is needs to be done right if it is to be worth the doing.
Alternatively, if the spammer/zombie computer has port 25 open itself, have a netfilter rule that rewrites the destination address to that of the sender, increases the TTL, and sends the packets back in duplicate. Again, this is a resource-draining scheme. If it's an open relay, it'll get the spam and resend it. I believe the hop count for SMTP is something like 30 and each packet will go two ways along the wire, so it'll take 2^31 as much bandwidth overall, if a sufficiently large number of users set up this kind of loopback. Companies that simply don't care if their machines are zombies will suddenly notice a degradation of their networks but any packet monitoring they do will show all of the packets to have the IP addresses of their machines for both source and destination. At least some will zombie detox to save their sanity.
Management is a fairly big expense, but as the total number of projects wouldn't change significantly, neither would the number of managers required. That just leaves the board of directors. Half the directors could be fired, but it's doubtful either CEO is going to consider their choice of senior management to be the inferior choice. Which means that one board would win and the other board will lose. On the whole, that is. The CEO of the winning board might cherry-pick a few who are really exceptional or who have given him lots of money.
You also need to bear in mind that CPU sales for AMD are (on average) rising but their profit margins are slumping, so if they gain access to another fab plant, it won't be to close it. It'll be for continuing in a price-war against Intel that both companies are losing. (Neither has the resources to continue until the other is completely vanquished AND remain competitive with other CPU manufacturers. Both Intel and AMD are latecomers in both the multi-core and 64-bit arenas, and neither can match the more experienced players on scalability at this time.) However, neither AMD nor Intel can afford to back off - their designs are divergent enough that the market cannot sustain both of them indefinitely. Intel can't even afford to maintain the StrongARM architecture anymore, things are so tight.
Gah. Makes me want to design a graphics processor just so there's a little variety in the market, even if my design was crap.
I'll add two more, so the list is at least 4: SiS and VIA both make GPUs. Does S3 still exist? Also, the MediaGX (rebadged to the Geode) has a GPU core built into the CPU, so that sort-of counts.
(Can you imagine what would happen if Microsoft bought it? Does anyone seriously believe ANY implementation would be safe, MESA included?)
Failing that, Google must have some spare change. Hell, they could probably buy SGI for less than the value of the machines in SGI's inventory, which would seriously boost their server power.
You need to bear in mind that the GPU is the critical component in most systems, but makes almost no money for the vendor and has a relatively low volume. There is precisely no reason whatsoever for AMD to want to merge with ATi or to buy them up. That would be expensive and earn them little. In fact, given how much they've made from their component-neutrality, sacrificing that might mean they'd actually lose money overall.
On the other hand, CPUs are high volume, high profit, and AMD is gaining market-share. It is an ideal target for a buy-out, particularly as ATi can't be doing that well in the GPU market. Buying AMD would be like buying a money-printing-machine, as far as ATi were concerned. Better still, AMD is a key player in bus specifications such as HyperTransport, which means that if ATi owned AMD, ATi could heavily influence the busses to suit graphics in general and their chips in particular.
(Mergers are never equal, as you don't have two CEOs, two CFOs, etc. One of them will be ultimately in charge of the other.)
If the rumour is correct, then don't assume AMD is the one instigating things - they have the most to lose and the least to gain - and don't assume either of them will be around when the mergers and buyouts finish.
Since you know the biochem side, you might want to write up a paper on how you'd go about this. At worst, you might easily get published, and journals pay very decently. At best, a biotech company that's a rival to 454 decides the idea would be both faster and more accurate, decides it's worth a shot to slaughter a rival, and buys the idea from you.
It would be good if LaTeX 3 ever got released, but the mailing list is silent and the website suggests nobody has done any signifcant development in years. If LaTeX 3 progress remains dead, I'd say fork the development tree as it stands and produce a variant with the features desirable in modern desktop publishing. Nothing to stop anyone, and LaTeX 2e came about precisely because people thought the original version 2 was crap and went on to make their own derivatives.
I doubt, for example, any DTP software out there has meaningful support for high-contrast colour formats such as OpenEXR or JPEG2000. Printers are probably not up to it yet, but who is going to add on a complex feature nobody can use? The software is much easier to change, so it is much easier to create the demand first.
(In fact, LaTeX' bitmap support is truly pathetic. True, LaTeX is designed to be totally scalable, which means vetors are always preferable, but there are many illustrations that simply can't be done that way.)
LaTeX' handling of subsections is also very poor and works in absolutes rather than relatives. Sure, not many people want to nest more than, oh, three or four deep. But plenty of technical texts can nest much deeper than that, and LaTeX has no provision for it, because of the way you have to name the specific depth you're working in. (This also makes it much harder to develop things section by section, as you have to be very cunning to make something viewable at an arbritary depth.)
Finally, you've got to process your fonts to work at a specific DPI, at the time of display or printing. "But won't I know this?" Only if you're using a pixel-based printer. What happens when someone produces a laser printer that can handle vectors as well as pixels? What's the DPI of a vector, if the laser has two degrees of freedom and can control both motors simultaneously? For that matter, what if I want to output to a plotter, where I already have this capability? Vector monitors aren't in widespread use any more, but they do exist. If the software is designed to work with such methods, shouldn't it be outputtable to such devices?
(LaTeX editors exist now, which was my big grudge for a long time, although they're nowhere near the point of, say, Ventura Publisher. Yet, anyway. This, despite the fact that TeX is already a very powerful engine.)
There is an added problem. Most geneticists use chain termination sequencing, which is good for fairly decent lengths of DNA. 454 uses pyrosequencing which is faster but only good for much smaller lengths. When the unknown elements may contain repeats of unknown length, smaller sample sizes are a Bad Idea. If the sample size is smaller than the repeat, you will be incapable of knowing how long the repeat is and will therefore have multiple equally valid solutions to the analysis. This is Bad, since it is these differences the researchers are looking for.
There is an alternative technique (primer walking/chromosome walking) which works on much larger fragments and should be able to deal with the indeterminacy problem completely.
However, this all assumes that we need to use an existing method at all. We have two known points of reference (humans and chimps) and a ready supply of DNA from both. ALL we need to determine is what segments of Neandertal DNA do not match with either, and then sequence just those segments. What would be needed, then, is a genetic version of diff, as opposed to a genetic version of cat. DNA is ideally suited to producing diffs, because you can't connect non-matching bases, but is much harder to cat.
I'd say that it would make far more sense to do the fundamental research needed to develop new techniques that can exploit the fact that we have multiple known related species with mapped DNA, where the exact relationship with Neandertal is also known, but the differences created by that relationship are not. We have a gigantic library of knowns, which we can use to simplify the process. Instead, 454 is going to use an entirely random, utterly unsuitable method, simply because it's faster than developing a new method.
If people always stuck to the tools they already had, to avoid developing something more appropriate to the problem, we'd still be using stone tools and the closest to blogging anyone could get would be to find a really, really big wall to paint on.
Twelve times coverage was used for human DNA, to deal with repeats which are damn-near impossible to place using the shotgun method and chain termination in a single pass. With the added problems of pyrosequencing, you might easily need four to five times as much coverage to get a reliable measurement. Neandertals had a lot longer to develop such repeats, so there's an excellent chance that far greater coverage again will be required to accurately determine where such repeats are and how long they are. Humans have had one tenth of the time to evolve distinctly from chimps than Neandertals, so the complexity created by repeats could easily be ten times as great. Throw in the contamination (which apparently requires 20 times the number of samples to eliminate) and you're talking 12,000 runs. Where, exactly, are these researchers going to find 12,000 Neandertal bones?
Also, bear in mind that the human genome project largely worked off the DNA of one person (one of the project directors, IIRC), so all DNA samples would produce the same repeats. Here, they're talking about using multiple sources of indeterminately different origins - both in geography and in time. The odds of the repeats being identical is almost (but not quite) nil.
In consequence, there will be multiple possible solutions, unless they use a sample size comparable to the entire Genographic Project's database (and they only looked at 12 STRs, not the entire nucleic DNA sequence). I'm not sure there are that many distinct examples of Neanderthals known, never mind available for crushing.
Sure, it worked on cave bears and mammoths to a degree (but we've no idea what that degree is), but I'd be willing to bet that the complexity of junk DNA for cave bears is far lower than for Neandertals anyway, and that their practice runs covered a geographically and temporally smaller region, vastly simplifying the process. Also, a lot of the really good samples of mammths in Siberia are frozen, so contamination and natural disintegration would have been far less significant.
Primer walking (also on Wikipedia, and a perfectly good alternative method to Shotgun Sequencing) would seem to be more promising. We know where common elements are in human and chimpanzee DNA, so we can very easily walk the chromosomes consecutively because we have a notion of what consecutive means. Random sampling eliminates all of the useful information we already have that could eliminate noise introduced by the method.
Besides which, what fool would limit themselves to existing technologies? Here is a perfectly good opportunity to produce entirely new methods of sequencing DNA when similar (but not identical) DNA exists. Existing methods can't exploit existing data, and we don't need to specifically know actual values, all that is of interest is the diffs. Existing methods, therefore, give us the wrong information from which we then have to calculate the information that is of actual interest. It follows that we should have no interest in the limitations of the shotgun method because that is not the method we should be using.
With the original sequencing of human and chimpanzee DNA, there was no existing data and therefore no known quantity to meaningfully diff against. You could also obtain a very large number of samples with highly predictable variations. It was therefore not only possible to use the shotgun method, it was the optimal method to use.
Here, we don't have that. We have a very small number of samples with no predictability in the variations, but which are likely to be extremely similar to very well-known sequences. It is insane to start entirely from scratch, in
If scientists find evidence of chair-flinging as a sport, this would truly explain a lot.
The other descriptions imply that it's contamination through questionable extraction techniques - they're grinding up the fossils, so ALL the DNA in the sample will be mixed together, and strands may well end up getting broken, making it much harder to sequence correctly.
Sequencing fossil DNA is certainly possible, and is extremely desirable, but the approach seems... odd. The BBC article, for example, claims that they're going to look for the genes that differentiate modern humans from neandertals, such as mental capacity. Given that we don't fully understand what "mental capacity" actually means, or indeed what mental capacity neandertals actually had, they would need to be looking for an unknown difference to identify an entirely theoretical and totally unquantifiable distinction. That's not good science.
Lastly, we know from studies of neandertal mtDNA that there was a large genetic diversity. Far larger than had been suspected, prior to that study. If these scientists are taking neandertal nucleic DNA from significantly different regions and/or times, they cannot be certain that the nucleic DNA had not evolved or otherwise differed to the point where direct comparison or simple in-lining of the genes would make no sense whatsoever.
This is a good research project, but I am highly uncertain of their methods and am not convinced it will yield meaningful results. Because repeat studies will be difficult to do, this is an area where those involved HAVE to take extra care to put their results beyond question. This care is NOT being taken, based on what I'm seeing.
Getting to Mars, though - that was easy. You load up some berserkers with drugs until they're sky high, then explode some distilled mead to launch them across the void.
...that said log cabin is in a neutral country and is sufficiently concealed by trees as to not appear on Microsoft's or Google's satellite photos.
Since the Government isn't a defendent, and as the US has no meaningful concept of "contempt of court" or perjury, the court can't do anything about it if the Government is found guilty of lying. On the other hand, this is election year, which is not a good year to be found guilty of anything, even if there is nothing the courts can do.
My guess is that the Government will do anything and everything to stall proceedings, such that if there is a trial, there's absolutely no risk of anything embarassing being said before polling day. If they're in power, they can clean things up afterwards. If they're not, it's no longer their problem.
Now, Osmium (roughly 6 times as valuable as gold) is a definite candidate for a precious metal, but couldn't be used ornamentally as it is highly toxic. Oil, as it is being consumed many millions of times faster than it is being generated, could potentially be a candidate as a precious substance.
Some of the freakier minerals are also candidates. Probably the freakiest I know of is a blue feldspar called "Blue John" that only naturally occurs in about a mile radius of the town of Castleton in the northwest of England. Because it is highly chaotic in nature (its absolute non-uniformity is part of its uniqueness and is what people like about it) it would be hard to reproduce in a laboratory. As such, it is pretty much guaranteed to remain extremely rare.
Speculating on future tastes, then, my guess is that non-uniformity and asymmetry will play a major role in the future.
Assuming it ate large numbers of beans, that is.