Who remembers the Rolls-Royce HOTOL proposal? It must be over 25 years ago now. It was to be a reuseable airbreathing horizontal take-off thing. It looked like an aircraft, but it had no crew, and it was aimed at the space bulk freight market. This would have saved a bunch on all the life support and pressurization stuff for the early models. If it had been found reliable after, say, 50 flights, then there was the option to add and extend a pressurized cabin, toilets, lemon-scented towels in individual sachets, and other comforts.
Okay, Britian has a long history of telling people what they ought to have built without actually putting very much together themselves. But it still strikes me as the right solution.
The second-generation plasma panel displays used to have up to four layers. The plasma panel cells are like little neon lamps - they are only stably either on or off, so to get greylevels (becuse they didn't switch very fast), the makers stacked up several layers, with a 50% grey filter between them. The most significant bit plane was at the front, the next one was behind it, and so on. There was a subtle 3-D effect too, but it was hard to see a real use for it.
I also remember another device where a mono LCD used a colour CRT as a backlight. At the time (about 1985) this offered high black-and-white resolution, and the ability to display CMYK (inverse RGB, and black), which was quite interesting at the time. The CRT had a thick front plate, so the LCD was clearly 'floating' some way in front of the CRT image.
A holodeck, it ain't. Even quite modest volumes contain an awful lot of voxels. Think how many little cubes you get in a bag of sugar.
I am sceptical, but I remember reading an article about sonoluminescence many years ago, and wondering whether this would be able to do what lasers and microbaloons couldn't.
What people are trying to do is to focus a shockwave. It's sort-of like cracking a whip with rotational symmetry. You start with the sort of energy densities you can generate by conventional processes, such as lasers or pulsed power or plastic explosive, and you focus the energy and momentum into a small amount of matter. If all the stuff collides in the centre and the momentum all cancels itself out, you can end up with tremendous pressures and energy densities. For example, as a light-bulb filmament burns out, the current runs through a smaller and smaller area, until it gives out a brief burst of thermal X-rays before turning to plasma. The pop you hear comes a bit later, when the plasma has thinned from its near-solid densities and avalanching becomes possible.
The problem is getting a symmetric implosion. People tried this with microballoons - little glass spheres containing deuterium-tritium mixtures. Unfortunately, if one bit is a little hotter than the rest, or one bit of the imploding material is a bit softer or lighter than the rest, or even if the surface is slightly curved in the wrong way, then you will get one bit that reaches the centre before the rest, and spoils the fun (Rayleigh-Taylor instabilities, and such stuff).
This experimental work starts with a large bubble for sonouminescence - about 1mm. We suspect that the forces driving the implosion are fairly symmetric, and there are some fairly large forces for surface tension that ought to keep the surface spherical over the first stages of the implosion. It would be nice if all the energy went into a symmetric implosion, but things don't work like that. You will get adiabatic heating at the front of the shock wave, and when the temperature exceeds the triple point of the material you are imploding, then you can't expect surface tension to work. From then on in, you just have to cross your fingers a hope things are symmetric enough.
In these experiments, it is hard to know how symmetric the implosion is, though I would guess it is smoother than the laser microballoon experiments, for a lot less cost. It is hard to know what physics are going on as the bubble smacks together, but we do know it must be something pretty extraordinary if we get bluish light from it. Maybe the Livermore labs people will dust off their code, and give us a good simulation that predicts sonoluminescence.
You might even be able to explain away the lack of neutrons, because you have oxygen and carbon in the implosion.
Okay, it's probably another false alarm. But I am going to have a few pleasurable moments gloating over the possibilities before I tear up my lottery ticket...
I one had a posting from an old steam train worker on Africa somwhere. They used to have pairs of back-to-back steam engines for pulling heavy trains. The drive wheels used to have a random scatter in the diameter, and people in the engine shed used to try and match up sets of driving wheels with the same diameter, because that made the enginges 'run more smoothly'. From his account, it seems that on long, straight sections of track, the two engines would lock into step, but only if the wheels were a near match on both engines.
You get it with piano strings too. Where two or more are tuned to the same note. The delta in tuning has an important on the sustain-decay profile of the notes.
Huygens figured out the general principle. If you have two things that are matched in frequency, and capable of influencing each other, then any influence, however tiny, will eventually drag them into some preferred phase relationship. If there is some difference in frequency, then this may destroy the coupling effect, if it is too small. You get it with piano strings too. Where two or more are tuned to the same note. The delta in tuning has an important on the sustain-decay profile of the notes. Arguing that entrainment must exist to some degree between two clocks is easy. Showing exactly what causes it is a lot harder. That is what the recent paper was about.
If you read the article, you might be left with the thought that there has been a glitch in the rollout of digital cinema technology.
There's all sorts of problems with film, from the way Monday's print bath is different to the others, through scratches, and sparkle, and the fact that the colours change with age, and the truly monster costs of actually printing all the copies of a blockbuster film. Bit film is still a tough act to follow.
On the digital size of things, TI spent something like 2 billion dollars and 20 years in research in the DLP elements in many projection systems before they even began to roll our the early digital OHP projectors. However, the DLP is an odd gadget: it is the most ambitious piece of MEMS technology, but it is made (for historical reasons) from alumina rather than silicon. Now, there are lots of people who claim that a more 'traditional' MEMS solution using silicon and silicon dioxide might give better results for lest cost. Other people claim you can get use a TV-like display with phosphors that work like lasers. So what do you do , if you own a chain of cinemas? Well, what they have done is to put a digital projector or two in a few high-profile places (we have a few in London), and sit and wait to see how the technology goes. Those digital projectors cost hundreds of thousands of dollars, so they could get seriously burned if it turns out they backed the wrong technology, or the customers don't like it. I doubt if the cinema chains ever had any plans to put in the numbers of digital projectors suggested by the article.
The next problem is the data format. Film doesn't have pixels, so you can run a film with 2K pixels per line or 4K pixels per line on the same projector. You can stick on anamorphic lenses to get widescreen. If you haven't got an IMAX projector, you can probably get a print on 35mm. There are a bag of mature technogies.
With digital, you have lots of similar decision to make, but none of the tecnologies are mature. You will probably have 2K pixels per line in the near future, with the promise of 3K or 4K sometime in the future. Digital projector A may have a better colour gamut than B, but projector B may be brighte and have less flicker, which makes the colours look better. If you are producing 'Monsters, Inc', then you can get your computer to make a version that fits the best projector resolution available, but this doesn't quite work with conventional films.
Lastly, there is the politics of fixing a data standard. Once, the data standard used to be decided by the company that had the best working solution. Nowdays, most data standards such as MPEG, JPEG, DICOM, TIFF, and ICC are an uneasy alliance of lots of companies, each wanting to get their patented process as part of the standard, and shake down the rest of the world for the licencing rights.
Eventually, digital to film will surely be like CD's to vinyl. But it's going to take a technical breakthrough or at least a decade of slow toil to shake out a few clear leaders from the contending technolgies, and get a digital standard that pleases most parties. Remember, the CD took a long time to actually get from the lab into our homes.
Canon made a point to point near IR laser link, called 'CanoBeam' in 1999. I have tried their website to see whether it works in fog, but I can't get an answer. Anyone?
The story of electronics for the last 50 years has mostly been more silicon, smaller, faster, more on a chip. To the unaided eye, it might seem that this is a bit more of the same. We have a clock, a CPU, and address bus, and memory, and all of these have beend designed and stuck in the right place. It couldn't be a computer if it wasn't like that, could it?
Not so. If we go on making smaller and smaller conventional semiconductor circuits, then it gets harder to make a simple conducting wire. Tunnelling effects make your wire appear thicker then it actually is. If your conducting layer is 2-3 atoms of Al thick, then your signal may get reflected from the steps betwen regions of 2 atoms and regions of 3 atoms. If you have a III-V semiconductor, the edge of your track may be charged depending on whether the edge lies on an III or a V. And there are all sorts of quantum funnies, I haven't mentioned. Strangely enough, making something like a transistor may get easier. So, when designing at the nanoscale, everything is topsy-turvy: you have to concentrate on the wires, and forget about the devices.
At some point, we may have to abandon the idea of a clock and a bus because we cannot get them to work reliably at the nanoscale. Instead, we may have a sludge of processing components that have settled out of a solution, with a certain fraction of duff components, and working ones. Okay, there are several good years left in Si, but we must not waste them: it is going to be a massive leap to change over from designed, synchronous circuits in crystalline Si to a semirandom, aynchronous circuits in organic molecules.
Problem number one: how do you program a massively parallel asynchronous array of processors. HP's elegant answer to this one was the Teramac project (see one of the other follow-ups).
Problem number two: how do you connect these processors to real devices? Well, an ordinary chip has tiny components, but real-sized legs. Between the millimetre scale and the micron scale, there are a whole set of technologies to bridge the gap in scale, making the connections and amplifying the signals. If (when?) we go to nanotech circuits, then we are going to need a whole new tier of these connecting technologies to bridge the gap between the micro-scale and the nano-scale. Your nanoprocessor sludge will probably sit on a conventional semiconductor chip.
Again, people at HP (and elsewhere) are trying to address these sorts of things. So, your future nanoprocessor thingy will probably look a lot like a conventional circuit from the outside.
A lot of the fatigue and wear problems we see on macroscopic devices are not problems with the bulk material, but problems with faults, inclusions, grain boundaries, and things like that. Every time you turn a real bike chain, the teeth will scrape off a few atoms, a dislocation may move by a small amount, a fatigue crack may get one atom deeper.
These little gadgets are so small that it is possible to make them out of a single, faultless piece of material. Okay, if you had a dislocation or an inclusion in your bike chain, then it would fail pretty quickly, if it worked at all; but if you get a good one, then it will seem almost immortal when compared to macroscopic objects. So, you make a few spares, and throw away the duds.
We are used to seeing silicon and silicon dioxide as crystalline. However, if you take out the small features that allow a crack to propagate through a crystal, then these materials can seem very tough and flexible. Think of glass fibres and glass. The Sandia site used to have a downloadable video of a minature moving mirror getting trodden on by a flea: it bends but does not crumple, and springs back unharmed.
There are other changes as you get to submicron sizes. Surface tension and other chemical effects seem huge. Water drops seem to have a tough skin on them at this scale, and drops will sit on a surface rather than wet it. This is just as well: a water drop could glue the chain together if it could wet. As things are, these gadgets seem to survive in the open atmosphere just fine.
If you think that is weird, the nanoscale stuff is much weirder. Interesting times, or what?
Yup, you got it. If you have a circular dimension, then you can stick standing waves into it. You could have a clockwise wave with N periods to the universe, and an anticlockwise wave with N periods to the universe. This would mean a wavefunction could have a set of two possible energy conditions for each integer N above zero. See under Kaluza-Klein theory (yeah, a bunch of old guys got there first, tough bananas eh?).
At the moment, the easiest, and lowest-risk way of getting up the resolution seems to be to crank down the wavelength of light. However, if this ever reaches a hard limit, then there are equivalent processes using electrons...
Instead of making a mask, you produce a pattern of raised points on a conducting plate. These points will emit electrons when a field is applied. You can then use reducing electron optics to produduce a much smaller image of your source on your semiconductor target. The whole chip could be imaged at once, so this would be no slower than conventional sources.
You could use a similar trick with ions. These have a bigger mass/charge ratio, so they take more volts to accelerate, but they are less put off by stray electrical and magnetic fields. The better ion beam milling microscopes use sub-nm beams these days.
There are other ways, and perhaps better optimums to be found. The flying wing design - a big hollow wing with hardly any separate fuselage or tail - was tried briefly back in the fifties. On paper, it looked good, but after a couple of crashes, the things were grounded, and nothing like that was produced again. There have also been designs where the wings point foreward from the fuselage, rather than being swept back. They gave lots of lift, but they were difficult to make stiff and stable enough.
These days with intelligent controls, we can make all sorts of unstable things fly. But we don't see flying wings, or wings on backwards. Why? Well, I expect there is a lot more expertise with designing conventional aircraft, so, unless the Dept. of Defense is picking up the tab, they stick to small variations on what they know.
I would like to see hollow wing passsenger aircraft. You would have loads of space to exercise on those long flights. However, we might need the Cold War back to get one built, and nothing's worth that...
When a company releases some software, it balances the money it hopes to rake in against the risks of additional costs, risks and liabilities, public exposure of intellectual property, and stuff like that. The buyer balances the benefits of buying a software package against the costs of the software and the time and effort they spend in getting familiar with it, the risks of non-support, and stuff like that. This balance of trust works both ways.
Most decent pieces of software should evolve and improve. Normally, if you are a serious user, you will want to have the latest version. The year's old source of GhostScript used to be free - but I was prepared to pay for this year's when I was doing something complex.
However, occasionally, the software does not evolve in the way you want. I know one image processing product that was lean and efficient, if a little dull. It was bought by another company, who bloated the code, stuck in all sorts of unwanted features, and slowed the thing to a crawl. The dedicated users are still using a version from about 5 years ago. Okay, shit happens, but they still had a working program. A 'sunset' clause would force these users to abandon their working product or fear litigation, but would not supply a workable alternative. Ten years or more ago, you used to get a lot of booby-trapped software, dongled code, and stuff like that, because it was sold by people who did not really understand software, what it cost, and why it was worth it.
We don't want to go back to those days. My gut feeling is that there ought to be a legal challenge to 'sunset' clauses. Whether you read them or not when you open the product packaging or click on the 'install' button, you ought to expect the one-machine use of the product in its supplied state, unless there are clearly explained exceptional reasons. Anything else would be a breach of the normal balance of trust between the user and the supplier. In the US, this sort of thing ought to attract federal anti-trust suits.
The solar chimney is a really neat idea for reasons that do not transfer to wind power.
All the moving bits are at the bottom (well - within 40M of the bottom). This means that you can get to service them without having to scale the chimney. You can swap out the generators for more efficient ones when they are developed without having to redesign the rest of the scheme.
There are windmill designs (the Savonius rotor) that have the generator at the bottom, and don't need pointing into the wind, but these are a design compromise between efficieny and servicability. The wind farms in Scotland have a
dynamo with a windmill on top of a big stick. I remember the 'Tomorrows World' presenter going up it, and going rather green: the really big ones are pretty scary places to work.
The chimney can also generate power when it is half-built. It won't be as efficient, but this may allow the building loan to be spread out. Once you have built the chimney, it may then make finiancial sense to expand the greenhouse area. A windmill is either there or it isn't.
Don't get me wrong - I like windmills, and a solar chimney in the Orkneys simply isn't on. However, the Orkneys windmill is paying because regular electricity was over 4 times the cost on the mainland. However, IMHO, the solar chimney is in a different league to windmills and tidal stations. I do hope it gets built.
Re:are artillery shells that delicate?
on
Battlefield Lasers
·
· Score: 1
People haven't really tried to make laser-proof
shells yet, but it wouldn't be hard.
Metal at normal temperatures is good at reflecting, and not good at absorbing. However, if you sink in enough power, then some of the metal is vapourized - this forms a layer of plasma just above the strike point, and that plasma absorbs the light just fine. If you are wanting to cut a hole in metal, then you would have to use a series of short pulses with gaps to allow the plasma to disperse. If you were machining something like plastic, or ceramic, then the thermal shock will tear out large chunks, but with metals you are usually boining of the stuff, atom by atom. If the object is spinning, then it is almost impossible to land the beam in the same place twice.
ICBMs are vulnerable in their boost phase, then there is lots of rocket propellant that goes off with a nice bang when hit. However, once the warhead is in the ballistic phase, then it is just a chunk of stuff flying through space: the ablating material might propel it a bit off course, but that's about it. For a short period during re-entry it may become delicate again as it finds its target, but then most of it's energy is in momentum, not propellant, so there is not much for your laser to make go bang.
The same is surely true of shells. If the fuse is hidden in the body of the shell, and there are no exposed and vulnerable non-metallic parts, then iut should be very difficult to attack.
Ion beams might be another matter - these can have momentum and a lot of penetration, but they don't tend to travel in straight lines in the atmosphere (hosepiping). Another poster has mentioned the thunder you might get if the beam heated the atmosphere. Now this has been proposed as a method for allowing an ion beam weapon to get a clear shot at the target.
Jus' another defence company putting out the high-tech begging bowl, I reckon...
I want to know what the insulator is. Silicon dioxide is a magnificent insulator. This wasn't that important in the early days, but the smaller things become, the more important it becomes.
You could have Silicon-on-Vacuum - a thin wafer
of silicon suspended by the edges, but that wouldn't take your heat away.
I still have a program somewhere for the Sinclair Spectrum, that generated speech. You could save data to a cassette recorder, so with the right
arrangement of ones and zeroes you could reproduce recognizeable speech. This might have run on the Z80.
The equivalent these days would be using a CD writer to make simple holograms. I vaguely remember someone at a UK university doing this, but I forgot the details.
I remember listening to HP calculator programs using a shortwave radio. I also remember some
PDPs (?) that had a loudspeaker.
An even older technique with ancient valve
computers was the 'stop loop'. You put
'A: GOTO A' at the end of your code, which
caused the electronics to whistle, to tell
you your data was ready. Not so good for
multiuser systems, though.
Aaarrr, those were the days, when you
programmed with a 60W soldering iron and
a pair of tinsnips. Kids today think clicking
a mouse is programming, don't know they've been born, blah blah etc....
The article is pretty vague about numbers. There was a better article about DNA computing in the New Scinetist a couple of years back.
A gram of material can contain 10^20-odd molecules. We are not really talking billions or trillions, but real monster numbers. Unfortunately the monster parallelism comes with severe I/O limits, and a low clock rate.
Suppose you wanted to crack an RSA cipher. You could use one type of molecule to represent prime numbers, and a second molecule to take one of the first type molecules, and try it on the cipher key. If you start off with a few cc's of prime numbers, you will probably have all of the 40-bit primes many times over, so many molecules will make the right conection.
Unfortunately, the molecules that make the right connection will be vastly outnumbered by the ones that don't, and the ones that went wrong, and the impurities, and everything else. To rescue the signal from the noise, you need another chemical stage. This should allow only the successful molecules to copy themselves. So you mix number solution 1 with RSA key solution 2, and stir it for a few minutes; then you add breeder solution 3, and wait for the most frequently encountered correct result to start crystallizing out.
This is a wonderfully parallel process for searching for a single solution to a simple problem. RSA hackers, and Goooogle might be able to use it, but you can't use it to do your 3-D renders. Awww.....
If we had to crack something like the Enigma codes today, then Bletchley Park would be developing DNA, instead of using relays and valves. The Bletchley Park Colossus was not a computer in today's sense - it was dedicated to solving a single problem - but the same people that developed it also worked on the earlier computers.
Other people have suggested making molecules with the electonic orbital equivalent of the electrical components we have in present circuits. But that was not what that article was about.
I remember a proposal for Quake-OS. You could
shoot files to delete them, and shoot executing tasks to halt them. To clean out a directory, you toss in a grenade.
To begin with, this was just a silly idea. However, Quake-OS did have some serious merits.
Instead of having that irritating prompt - Do you really want to delete? {Y/N] - you see a hand holding the grenade, and you can stop it if you
are quick enough. If there is someone working in a nearby directory, you can see the flashes and hear the bangs. Links would be portals. New files could be white, and old files could go yellow.
I am not sure the differences wold have been worth the effort, but it would have been fun...
Asynchronous design seems like GaAs. It has always been just about to sweep the world, but always the regular stuff managed to close the gap enough to make it not happen, which is, perhaps, as it should be.
There are many cunning things you can do with asynchronous design. Your ear has some really neat asynchronous components that allow you to judge the relative phase of signals up to 4 KHz. However, lately, people have theorized that even our brains use the phase between the alpha and theta signals as a sort-of clock to correlate disparate bits of information such as the position, shape, colour, and meaning of an object, all processed in different parts of the brain.
People have theorized that molecular-level computing will have to be asynchronous. Maybe some of it will, but I bet we will reinvent the clock too.
The circuits will have zero resistance only when they are in a stable state. Your static memory will not consume power provided you don't try and read it. Unfortunately, most interesting bits of computing will involve changing the electronic states, so there will still be power consumption, and trouble getting rid of heat.
Carbon will probably be the new silicon. It has a big 10eV band gap, and you can make it a resistor, a semiconductor, a conductor, or a superconductor by rearranging the bonds, without doping. If we can crack the self-assembly problems, then you may get a mole of bits in a few tens of grammes of material. Which may not be instantaneous calculations for no energy, but it is pretty good to be going on with.
Making a whole computer is also possible, but this may take a little longer.
Go back to the person who told you you were too old at 33, and set fire to them. This is important. There are so many people with responsibility for hiring people who think that way, particularly in the UK, and it may seem that setting light to one of them will hardly make a difference. But if every one of us set light to just one of these people a year, then we will begin to make a real difference.
Seriously though, my mother, started using a computer for the first time whe she was over 80, and she remembers the keyboard shortcuts that I never can. My brother-in-law got his present job because he was economical with the truth when asked about his age, and was told long after that he would never have been hired if his real age had been known. Being old does not of itself make you good at computers, but should not disqualify you either. If you are old, and you think you know computers, then you probably do. Someone whe rejects people on their age is missing a lot of good talent. It may be unfair, and politically incorrect as an earlier poster has pinted out, but above all, it is a terrible waste.
Do not pay too much attention to the exact requirements in the adverts, either. The wording is often done by recruiting agences, who like working with keywords, and badger people to specify which operating system, and which package. Sometimes it will be important - but, more often than not, an impressive collection of irrelevent talents that shows a willingness to learn is much better than an exact match.
BTW - no, I am not a sysadmin. But I have occasioanlly recruited people, so I have seen the system form the other side.
Slashdot Mirror
Okay, Britian has a long history of telling people what they ought to have built without actually putting very much together themselves. But it still strikes me as the right solution.
I also remember another device where a mono LCD used a colour CRT as a backlight. At the time (about 1985) this offered high black-and-white resolution, and the ability to display CMYK (inverse RGB, and black), which was quite interesting at the time. The CRT had a thick front plate, so the LCD was clearly 'floating' some way in front of the CRT image.
A holodeck, it ain't. Even quite modest volumes contain an awful lot of voxels. Think how many little cubes you get in a bag of sugar.
What people are trying to do is to focus a shockwave. It's sort-of like cracking a whip with rotational symmetry. You start with the sort of energy densities you can generate by conventional processes, such as lasers or pulsed power or plastic explosive, and you focus the energy and momentum into a small amount of matter. If all the stuff collides in the centre and the momentum all cancels itself out, you can end up with tremendous pressures and energy densities. For example, as a light-bulb filmament burns out, the current runs through a smaller and smaller area, until it gives out a brief burst of thermal X-rays before turning to plasma. The pop you hear comes a bit later, when the plasma has thinned from its near-solid densities and avalanching becomes possible.
The problem is getting a symmetric implosion. People tried this with microballoons - little glass spheres containing deuterium-tritium mixtures. Unfortunately, if one bit is a little hotter than the rest, or one bit of the imploding material is a bit softer or lighter than the rest, or even if the surface is slightly curved in the wrong way, then you will get one bit that reaches the centre before the rest, and spoils the fun (Rayleigh-Taylor instabilities, and such stuff).
This experimental work starts with a large bubble for sonouminescence - about 1mm. We suspect that the forces driving the implosion are fairly symmetric, and there are some fairly large forces for surface tension that ought to keep the surface spherical over the first stages of the implosion. It would be nice if all the energy went into a symmetric implosion, but things don't work like that. You will get adiabatic heating at the front of the shock wave, and when the temperature exceeds the triple point of the material you are imploding, then you can't expect surface tension to work. From then on in, you just have to cross your fingers a hope things are symmetric enough.
In these experiments, it is hard to know how symmetric the implosion is, though I would guess it is smoother than the laser microballoon experiments, for a lot less cost. It is hard to know what physics are going on as the bubble smacks together, but we do know it must be something pretty extraordinary if we get bluish light from it. Maybe the Livermore labs people will dust off their code, and give us a good simulation that predicts sonoluminescence.
You might even be able to explain away the lack of neutrons, because you have oxygen and carbon in the implosion.
Okay, it's probably another false alarm. But I am going to have a few pleasurable moments gloating over the possibilities before I tear up my lottery ticket...
You get it with piano strings too. Where two or more are tuned to the same note. The delta in tuning has an important on the sustain-decay profile of the notes.
Huygens figured out the general principle. If you have two things that are matched in frequency, and capable of influencing each other, then any influence, however tiny, will eventually drag them into some preferred phase relationship. If there is some difference in frequency, then this may destroy the coupling effect, if it is too small. You get it with piano strings too. Where two or more are tuned to the same note. The delta in tuning has an important on the sustain-decay profile of the notes. Arguing that entrainment must exist to some degree between two clocks is easy. Showing exactly what causes it is a lot harder. That is what the recent paper was about.
There's all sorts of problems with film, from the way Monday's print bath is different to the others, through scratches, and sparkle, and the fact that the colours change with age, and the truly monster costs of actually printing all the copies of a blockbuster film. Bit film is still a tough act to follow.
On the digital size of things, TI spent something like 2 billion dollars and 20 years in research in the DLP elements in many projection systems before they even began to roll our the early digital OHP projectors. However, the DLP is an odd gadget: it is the most ambitious piece of MEMS technology, but it is made (for historical reasons) from alumina rather than silicon. Now, there are lots of people who claim that a more 'traditional' MEMS solution using silicon and silicon dioxide might give better results for lest cost. Other people claim you can get use a TV-like display with phosphors that work like lasers. So what do you do , if you own a chain of cinemas? Well, what they have done is to put a digital projector or two in a few high-profile places (we have a few in London), and sit and wait to see how the technology goes. Those digital projectors cost hundreds of thousands of dollars, so they could get seriously burned if it turns out they backed the wrong technology, or the customers don't like it. I doubt if the cinema chains ever had any plans to put in the numbers of digital projectors suggested by the article.
The next problem is the data format. Film doesn't have pixels, so you can run a film with 2K pixels per line or 4K pixels per line on the same projector. You can stick on anamorphic lenses to get widescreen. If you haven't got an IMAX projector, you can probably get a print on 35mm. There are a bag of mature technogies. With digital, you have lots of similar decision to make, but none of the tecnologies are mature. You will probably have 2K pixels per line in the near future, with the promise of 3K or 4K sometime in the future. Digital projector A may have a better colour gamut than B, but projector B may be brighte and have less flicker, which makes the colours look better. If you are producing 'Monsters, Inc', then you can get your computer to make a version that fits the best projector resolution available, but this doesn't quite work with conventional films.
Lastly, there is the politics of fixing a data standard. Once, the data standard used to be decided by the company that had the best working solution. Nowdays, most data standards such as MPEG, JPEG, DICOM, TIFF, and ICC are an uneasy alliance of lots of companies, each wanting to get their patented process as part of the standard, and shake down the rest of the world for the licencing rights.
Eventually, digital to film will surely be like CD's to vinyl. But it's going to take a technical breakthrough or at least a decade of slow toil to shake out a few clear leaders from the contending technolgies, and get a digital standard that pleases most parties. Remember, the CD took a long time to actually get from the lab into our homes.
Canon made a point to point near IR laser link, called 'CanoBeam' in 1999. I have tried their website to see whether it works in fog, but I can't get an answer. Anyone?
Not so. If we go on making smaller and smaller conventional semiconductor circuits, then it gets harder to make a simple conducting wire. Tunnelling effects make your wire appear thicker then it actually is. If your conducting layer is 2-3 atoms of Al thick, then your signal may get reflected from the steps betwen regions of 2 atoms and regions of 3 atoms. If you have a III-V semiconductor, the edge of your track may be charged depending on whether the edge lies on an III or a V. And there are all sorts of quantum funnies, I haven't mentioned. Strangely enough, making something like a transistor may get easier. So, when designing at the nanoscale, everything is topsy-turvy: you have to concentrate on the wires, and forget about the devices.
At some point, we may have to abandon the idea of a clock and a bus because we cannot get them to work reliably at the nanoscale. Instead, we may have a sludge of processing components that have settled out of a solution, with a certain fraction of duff components, and working ones. Okay, there are several good years left in Si, but we must not waste them: it is going to be a massive leap to change over from designed, synchronous circuits in crystalline Si to a semirandom, aynchronous circuits in organic molecules.
Problem number one: how do you program a massively parallel asynchronous array of processors. HP's elegant answer to this one was the Teramac project (see one of the other follow-ups).
Problem number two: how do you connect these processors to real devices? Well, an ordinary chip has tiny components, but real-sized legs. Between the millimetre scale and the micron scale, there are a whole set of technologies to bridge the gap in scale, making the connections and amplifying the signals. If (when?) we go to nanotech circuits, then we are going to need a whole new tier of these connecting technologies to bridge the gap between the micro-scale and the nano-scale. Your nanoprocessor sludge will probably sit on a conventional semiconductor chip. Again, people at HP (and elsewhere) are trying to address these sorts of things. So, your future nanoprocessor thingy will probably look a lot like a conventional circuit from the outside.
I love it when a plan comes together....
These little gadgets are so small that it is possible to make them out of a single, faultless piece of material. Okay, if you had a dislocation or an inclusion in your bike chain, then it would fail pretty quickly, if it worked at all; but if you get a good one, then it will seem almost immortal when compared to macroscopic objects. So, you make a few spares, and throw away the duds.
We are used to seeing silicon and silicon dioxide as crystalline. However, if you take out the small features that allow a crack to propagate through a crystal, then these materials can seem very tough and flexible. Think of glass fibres and glass. The Sandia site used to have a downloadable video of a minature moving mirror getting trodden on by a flea: it bends but does not crumple, and springs back unharmed.
There are other changes as you get to submicron sizes. Surface tension and other chemical effects seem huge. Water drops seem to have a tough skin on them at this scale, and drops will sit on a surface rather than wet it. This is just as well: a water drop could glue the chain together if it could wet. As things are, these gadgets seem to survive in the open atmosphere just fine.
If you think that is weird, the nanoscale stuff is much weirder. Interesting times, or what?
Instead of making a mask, you produce a pattern of raised points on a conducting plate. These points will emit electrons when a field is applied. You can then use reducing electron optics to produduce a much smaller image of your source on your semiconductor target. The whole chip could be imaged at once, so this would be no slower than conventional sources.
You could use a similar trick with ions. These have a bigger mass/charge ratio, so they take more volts to accelerate, but they are less put off by stray electrical and magnetic fields. The better ion beam milling microscopes use sub-nm beams these days.
These days with intelligent controls, we can make all sorts of unstable things fly. But we don't see flying wings, or wings on backwards. Why? Well, I expect there is a lot more expertise with designing conventional aircraft, so, unless the Dept. of Defense is picking up the tab, they stick to small variations on what they know.
I would like to see hollow wing passsenger aircraft. You would have loads of space to exercise on those long flights. However, we might need the Cold War back to get one built, and nothing's worth that...
Most decent pieces of software should evolve and improve. Normally, if you are a serious user, you will want to have the latest version. The year's old source of GhostScript used to be free - but I was prepared to pay for this year's when I was doing something complex.
However, occasionally, the software does not evolve in the way you want. I know one image processing product that was lean and efficient, if a little dull. It was bought by another company, who bloated the code, stuck in all sorts of unwanted features, and slowed the thing to a crawl. The dedicated users are still using a version from about 5 years ago. Okay, shit happens, but they still had a working program. A 'sunset' clause would force these users to abandon their working product or fear litigation, but would not supply a workable alternative. Ten years or more ago, you used to get a lot of booby-trapped software, dongled code, and stuff like that, because it was sold by people who did not really understand software, what it cost, and why it was worth it.
We don't want to go back to those days. My gut feeling is that there ought to be a legal challenge to 'sunset' clauses. Whether you read them or not when you open the product packaging or click on the 'install' button, you ought to expect the one-machine use of the product in its supplied state, unless there are clearly explained exceptional reasons. Anything else would be a breach of the normal balance of trust between the user and the supplier. In the US, this sort of thing ought to attract federal anti-trust suits.
All the moving bits are at the bottom (well - within 40M of the bottom). This means that you can get to service them without having to scale the chimney. You can swap out the generators for more efficient ones when they are developed without having to redesign the rest of the scheme.
There are windmill designs (the Savonius rotor) that have the generator at the bottom, and don't need pointing into the wind, but these are a design compromise between efficieny and servicability. The wind farms in Scotland have a dynamo with a windmill on top of a big stick. I remember the 'Tomorrows World' presenter going up it, and going rather green: the really big ones are pretty scary places to work.
The chimney can also generate power when it is half-built. It won't be as efficient, but this may allow the building loan to be spread out. Once you have built the chimney, it may then make finiancial sense to expand the greenhouse area. A windmill is either there or it isn't.
Don't get me wrong - I like windmills, and a solar chimney in the Orkneys simply isn't on. However, the Orkneys windmill is paying because regular electricity was over 4 times the cost on the mainland. However, IMHO, the solar chimney is in a different league to windmills and tidal stations. I do hope it gets built.
Metal at normal temperatures is good at reflecting, and not good at absorbing. However, if you sink in enough power, then some of the metal is vapourized - this forms a layer of plasma just above the strike point, and that plasma absorbs the light just fine. If you are wanting to cut a hole in metal, then you would have to use a series of short pulses with gaps to allow the plasma to disperse. If you were machining something like plastic, or ceramic, then the thermal shock will tear out large chunks, but with metals you are usually boining of the stuff, atom by atom. If the object is spinning, then it is almost impossible to land the beam in the same place twice.
ICBMs are vulnerable in their boost phase, then there is lots of rocket propellant that goes off with a nice bang when hit. However, once the warhead is in the ballistic phase, then it is just a chunk of stuff flying through space: the ablating material might propel it a bit off course, but that's about it. For a short period during re-entry it may become delicate again as it finds its target, but then most of it's energy is in momentum, not propellant, so there is not much for your laser to make go bang.
The same is surely true of shells. If the fuse is hidden in the body of the shell, and there are no exposed and vulnerable non-metallic parts, then iut should be very difficult to attack.
Ion beams might be another matter - these can have momentum and a lot of penetration, but they don't tend to travel in straight lines in the atmosphere (hosepiping). Another poster has mentioned the thunder you might get if the beam heated the atmosphere. Now this has been proposed as a method for allowing an ion beam weapon to get a clear shot at the target.
Jus' another defence company putting out the high-tech begging bowl, I reckon...
You could have Silicon-on-Vacuum - a thin wafer of silicon suspended by the edges, but that wouldn't take your heat away.
The equivalent these days would be using a CD writer to make simple holograms. I vaguely remember someone at a UK university doing this, but I forgot the details.
An even older technique with ancient valve computers was the 'stop loop'. You put 'A: GOTO A' at the end of your code, which caused the electronics to whistle, to tell you your data was ready. Not so good for multiuser systems, though.
Aaarrr, those were the days, when you programmed with a 60W soldering iron and a pair of tinsnips. Kids today think clicking a mouse is programming, don't know they've been born, blah blah etc....
A gram of material can contain 10^20-odd molecules. We are not really talking billions or trillions, but real monster numbers. Unfortunately the monster parallelism comes with severe I/O limits, and a low clock rate.
Suppose you wanted to crack an RSA cipher. You could use one type of molecule to represent prime numbers, and a second molecule to take one of the first type molecules, and try it on the cipher key. If you start off with a few cc's of prime numbers, you will probably have all of the 40-bit primes many times over, so many molecules will make the right conection.
Unfortunately, the molecules that make the right connection will be vastly outnumbered by the ones that don't, and the ones that went wrong, and the impurities, and everything else. To rescue the signal from the noise, you need another chemical stage. This should allow only the successful molecules to copy themselves. So you mix number solution 1 with RSA key solution 2, and stir it for a few minutes; then you add breeder solution 3, and wait for the most frequently encountered correct result to start crystallizing out.
This is a wonderfully parallel process for searching for a single solution to a simple problem. RSA hackers, and Goooogle might be able to use it, but you can't use it to do your 3-D renders. Awww.....
If we had to crack something like the Enigma codes today, then Bletchley Park would be developing DNA, instead of using relays and valves. The Bletchley Park Colossus was not a computer in today's sense - it was dedicated to solving a single problem - but the same people that developed it also worked on the earlier computers.
Other people have suggested making molecules with the electonic orbital equivalent of the electrical components we have in present circuits. But that was not what that article was about.
To begin with, this was just a silly idea. However, Quake-OS did have some serious merits. Instead of having that irritating prompt - Do you really want to delete? {Y/N] - you see a hand holding the grenade, and you can stop it if you are quick enough. If there is someone working in a nearby directory, you can see the flashes and hear the bangs. Links would be portals. New files could be white, and old files could go yellow.
I am not sure the differences wold have been worth the effort, but it would have been fun...
Asynchronous design seems like GaAs. It has always been just about to sweep the world, but always the regular stuff managed to close the gap enough to make it not happen, which is, perhaps, as it should be.
There are many cunning things you can do with asynchronous design. Your ear has some really neat asynchronous components that allow you to judge the relative phase of signals up to 4 KHz. However, lately, people have theorized that even our brains use the phase between the alpha and theta signals as a sort-of clock to correlate disparate bits of information such as the position, shape, colour, and meaning of an object, all processed in different parts of the brain.
People have theorized that molecular-level computing will have to be asynchronous. Maybe some of it will, but I bet we will reinvent the clock too.
The circuits will have zero resistance only when they are in a stable state. Your static memory will not consume power provided you don't try and read it. Unfortunately, most interesting bits of computing will involve changing the electronic states, so there will still be power consumption, and trouble getting rid of heat.
Carbon will probably be the new silicon. It has a big 10eV band gap, and you can make it a resistor, a semiconductor, a conductor, or a superconductor by rearranging the bonds, without doping. If we can crack the self-assembly problems, then you may get a mole of bits in a few tens of grammes of material. Which may not be instantaneous calculations for no energy, but it is pretty good to be going on with.
Making a whole computer is also possible, but this may take a little longer.
Seriously though, my mother, started using a computer for the first time whe she was over 80, and she remembers the keyboard shortcuts that I never can. My brother-in-law got his present job because he was economical with the truth when asked about his age, and was told long after that he would never have been hired if his real age had been known. Being old does not of itself make you good at computers, but should not disqualify you either. If you are old, and you think you know computers, then you probably do. Someone whe rejects people on their age is missing a lot of good talent. It may be unfair, and politically incorrect as an earlier poster has pinted out, but above all, it is a terrible waste.
Do not pay too much attention to the exact requirements in the adverts, either. The wording is often done by recruiting agences, who like working with keywords, and badger people to specify which operating system, and which package. Sometimes it will be important - but, more often than not, an impressive collection of irrelevent talents that shows a willingness to learn is much better than an exact match.
BTW - no, I am not a sysadmin. But I have occasioanlly recruited people, so I have seen the system form the other side.