You mean like the Fedora installer that pops up when you boot the live disk and says "Click here to install to disk"? It takes maybe five minutes to put Fedora 17 live onto a hard drive, saying "use entire disk" when it asks. Not as good as dd if=/dev/urandom of=/dev/sda' though (preferred because somebody who really really tried could probably resurrect the old data through a mere zeroing, but it would be a lot harder if overwritten with random bytes, and way harder if you executed this command five or six times in a row).
Yeah, I'm having a few problems with the idea. Temperature actually implies thermal equilibrium, which in turn requires interaction. However, those atoms/molecules are, shall we say cosmicly non-interacting, being so dispersed that they basically form a hard vacuum. There are then a number of problems with the picture. One, why do the molecules not simply fall back into the galaxy (or if you like, why were they pushed out of the galaxy in the first place)? Several billion years accumulation of solar wind and outflow from supernovae? Second, why don't they "cool" to equilibrium with the background blackbody radiation? That one at least is answerable -- they basically never collide (as you say). The third question is -- how can one detect the gas? The molecules are cold, non-interacting, and enormously diffuse. To the extent that they interact with photons, they would relatively quickly slow down to equilibrium, so they are almost by definition invisible. Furthermore, photons are photons, and being a gas detecting the origin of a photon approaching the Earth from any given direction (that is, the distance of the source) is quite impossible, ditto for absorption lines from distant stars. I mean, parallax won't work. Neither will red shift. Alterations of gravity (as in the way one infers "Dark Matter") might work, but won't give you the speed or "temperature" or ensure that the matter that is altering orbits is a gas of baryons as opposed to darkonium.
Indeed, the title is horribly misleading. It would be better to say that the galaxy may be surrounded by a very diffuse gas of particles at a very low temperature in a reference frame that has a very high velocity relative to the galactic core, and we infer this compared to all competing explanations by sacrificing a chicken with a black-handled knife on the keyboard of a computer.
Also, listening to music and tapping your feet has nothing to do with multitasking. These activities are sent to the background and can perfectly happen while you're focusing on a specific task. Having to lose focus all the time because of telephone calls is completely different and is a serious productivity killer.
Fair enough. However, the study was also apples and oranges -- that was my primary point. It measured the degradation in utility executing a single primary task without allowing for the utility of the interrupting task, or for the value of opportunity cost time. You may find it more pleasurable and less crazy-making to work on just one task to the exclusion of others, but your employer may find far greater utility in your employment by having you answer the phone on an interrupt-driven basis (that is, when it rings) to handle whatever problem associated with the call and work on longer term but less urgent project in between. The degradation in performance in that task may not matter to them as long as they don't have to hire another person to largely sit idle and "just" answer the phone, especially if they'd have to hire somebody with more or less your skillset to do the answering (e.g. providing deep support for some technical product).
If they'd measured the actual degradation in the value of the total of a person's work-related activities, that would have been very useful, especially if they'd done so in a way that could actually be rescaled and used to do quantitative optimization. But measuring the degradation in primary task execution subject to random irrelevant "noise", especially on an unrealistically short time granularity, is both less useful and verifying a proposition that nobody would have doubted anyway. Yes, if you slap somebody in the face or give them an electrical shock according to some randomized schedule with a mean time between events that is order of a minute or two, you will interfere with their productivity (and driven them quietly crazy), but who cares? If you force them to switch complex tasks every forty seconds, I'm certain that for most people it will degrade their productivity on both tasks compared to doing them without the switch (although people are quite variable and there may be people that can manage this for at least certain classes of task especially with practice) but how much their productivity degrades at least might be relevant, especially if you explore the task degradation quantitatively as a function of the switch interval and across a decent population (to have a decent chance of sampling outliers who do much better or much worse).
The latter's the rub. My wife (a physician) cannot work with information with music playing -- too distracting. I can and often do. My wife can juggle certain sets of "simultaneous" tasks -- tasks she switches to and from on demand -- while others drive her crazy. I do much better at general multitasking and usually have a much broader set of things I'm working on at any given time and can switch between them fairly easily at reasonable granularity. For example, I'm writing this now, but will be clearing email in a minute or two (and was clearing email before getting the message announcing this reply to my slashdot comment) and am in the middle of at least a dozen project getting slices of my time as opportunity and need dictate. Different people have very different tolerances and abilities in this regard, even for "random, annoying" interrupt-driven subtasks like letting the dog out or answering the phone or working someplace where people are hustling and bustling all around you. I'm not at all certain that even publishing an average degradation on memory or completion or error rate is useful -- the distribution itself might be, because it might be multimodal or have long tails or otherwise not be classically "normal" so that the mean has some relevant meaning.
The key is to know your OWN abilities and inclinations in this regard, and to
Except that is not what they tested, is it? They tested people on whether or not they could work efficiently if they are distracted, not whether or not multitasking improves or doesn't improve efficiency.
The silly thing is that we actually know quite a lot about task organization for efficient multitasking. It is a key component of task scheduling on any computer. Fine grained multitasking -- especially on a CPU that has a large latency component for switching tasks, is known as "thrashing", and is also known to degrade performance substantially, and whacking a computer with a steady stream of pointless interrupts so that it is always thrashing slows it down.
At the same time, executing tasks with the right granularity and with the right kinds of latency and parallelism can speed things up quite a lot compared to doing tasks one at a time. This, too, is true in life as much as it is in computers. Anybody who cooks knows that you cannot generally make a good meal in serial fashion. If you want to serve rice with a stir fry and end up with dessert in a timely manner, you have to be cooking the rice, chilling the dessert, and chopping up and frying the main course all "at the same time", with layered overlaps in the attention you pay to the different tasks. The tasks are all related and a skilled cook can juggle quite a few of them without cognitive or operational overload and finish a meal far faster than anyone would ever finish it cooking one thing at a time (to a soggy, cold, unproductive finish).
Most normal humans multitask all the time. I listen to music and work while wiggling my feet to maintain circulation. I hop from answering email to posting silly things like this reply to doing work on task A to doing work on task B to doing work on task C -- so much the more so when my tasks are all different, all use the computer (or a number of computers) and take different amounts of time (attended or unattended) to move on to the next stage of completion. Yes, I can be overloaded, I can thrash in my normal work if overloaded so little gets done, but that is entirely different from asking if I can work when somebody is randomly blasting uncorrelated and meaningless distractions into my workspace.
Actually, programming is more like writing a cookbook. Lots of people who cook can't write a cookbook. For example, illiterate ones. Similarly, it helps to be literate in a computer language in order to be able to program, which in turn requires an above average ability to deal with a peculiar kind of metaphor and an above average understanding of tools far more demanding than a source of heat, a knife, and some ingredients. The thing that prevents old women or old men or old monkeys or old dogs from moving from food to programming is a mix of intelligence, interest, and motivation.
It is, for example, fairly common belief that programmers make an income that is well above average. It's a common belief because it is true. Yet you don't see teen-age fry cooks piling in to programming to multiply their minimum wage income by close to an order of magnitude. Why not? Because programming is difficult, and you have to be both smart enough to do it and inclined to WANT to do it, regardless of the obvious rewards of working in an air conditioned environment sitting on your ass while making $60K/year or more with benefits compared to slinging greasy burgers at possibly armed and dangerous clients in a Burger King late at night for $200 a week on a good week -- ooo, and then there are those pesky social security deductions and a manager that laughs hysterically if you mention the word "benefits" right before he fires you.
I would have to disagree. Programming isn't algebra, it is metaphor. It is building machines out of words. In fact, I don't know if you meant it humorously, but you yourself say "define the WORD algorithm". I'd argue that most algorithms CAN be described in words. Often even English or Swahili words.
I'm not sure, but I think that's why they call the medium a "programming language". It isn't devoid of math or logic -- far from it -- but it isn't the same thing, either, and a person good at real mathematics can easily suck at programming or vice versa.
Here's a program:
"Take an integer variable named i, fill it sequentially with the integer values from one to one hundred in steps of one, and print out its value on the screen I'm looking at".
This is perfectly understandable English, and can be executed by a reasonably bright student to whom the words are directed where the screen is a whiteboard.
It isn't really any different when written "for (i=1;i <=100;i++) printf("%d\n",i);" or any of the myriad other ways of writing the same program in different programming languages. A machine built out of words/symbols that have a fairly carefully specified operational meaning. Not at all like proving that the angles in a plane triangle add up to pi.
And don't forget the Flynn Effect. It isn't even a constant 100, so 100 this decade doesn't mean what 100 meant last decade.
But still, no, it is not true that my brother (with Down's Syndrome) could have ever become a programmer, whereas I, with an IQ (FWIW) several standard deviations over the mean have gigabytes of source in my source directory, a rather large fraction of which I actually wrote, in several languages.
So the question is still a stupid question, as the answer is obviously no. Worse, it is basically trollbait BEYOND being a stupid question, as somebody of "normal" intelligence can probably write "a program" in a sufficiently simple environment without ever in their lifetime being capable of writing a 50,000 line program with 100 functional modules written on top of various APIs (some of which they created) running over the network on top of UDP socket layer code. Actually, a lot of fairly ABOVE average intelligence well-trained programmers might fail there, or do a poor job if they succeeded.
So the proper answer is "No, to be a good programmer you have to be smarter than the average human, and probably better educated too. Propensity to skip showers and live on Jolt Cola optional. Troll."
space-based solar power arrays parked in geosynchronous or near-geosynchronous orvbit.
Ah, I once thought as you do, but then a measure of common sense asserted itself. Consider the fact that the cost of getting to geosynchronous orbit is, per kilogram, larger than the energy output of a kilogram's worth of cells over a lifetime of "forever" (or damn near). Consider further that a gigawatt's worth of space array, beaming its energy back to the ground (at some cost in efficiency, transmission losses) is more or less a gigawatt-scale space weapon if it is aimed somewhere other than whatever patch of ground set aside as a receiver. What can go wrong? Consider that you can avoid this problem, sure, by using a very weak beam, but then you have to use a very large piece of ground as a receiver, one that increases in size with the geometry of latitude giving you a second trade-off between area of receiver and atmospheric loss at higher latitudes versus the difficulty of very long distance power transmission from the equator to the temperate zone. Consider that TOA insolation is only a factor of two or so larger than BOA insolation (so it's not like you get a lot more power by being out of the atmosphere) and land is cheap in the desert, and there is plenty of desert. Finally consider that land is REALLY cheap on your own rooftop, which very likely contains ALMOST enough area to completely supply your own house's energy needs and can "store" energy by simply dumping surplus back into the grid during the day at reverse cost to be drawn out again at night "for free", even without an ever-improving local storage option.
Consider that the cost of actually putting 5 kW of solar cells on your roof NOW is more than break even on a 20 year amortization or less (in many parts of the country) with the amortization schedule dropping with the cost of solar cells and other improvements in the technology. The cost of solar cells per delivered watt has been dropping exponentially with a halving time of around a decade for the last three or four decades. It is currently between $1 and $2 per watt, plus installation and hardware costs. At $1/watt -- already available to large commercial buyers -- the amortization time for a 5 kW rooftop installation is order of a decade: it will generate order of $1000 worth of electricity per year, enough to pay off a $7000-8000 loan and even make a profit over that time. I've spent more than that on high efficiency furnace/AC for my house -- several times over, sadly -- with an even longer amortization. And, of course, anything that is "profitable" on the scale of individual rooftops is far MORE profitable on an industrial scale with industrial economies of scale. $1/watt retail is $0.50/watt wholesale in volume, and even allowing for installation and operation and maintenance costs, POWER COMPANIES will be GIVING you units to put on your roof -- as long as they can sell you slightly discounted power from those units. Or building large arrays themselves, but then they have the pesky problem of buying kilometer-square chunks of land here and there.
So the real problem with putting solar cells in space is that if the price drops, as one can very reasonably expect, to under $1/watt full retail over the next decade, solar generation will proliferate like a weed all over the world not to save the whales or lower carbon footprint but because it is the cheapest or second cheapest way to make electricity. This will happen even if there ARE no breakthroughs in gigawatt-scale 24 hour plus storage, although I personally think that physicists and engineers will beat the storage problem too within the next decade -- the payoff for doing so is huge. Sure, we'll still need bridge power -- nuclear and probably coal or natural gas -- but the actual draw on those facilities will decrease to a fraction of what it is today. Whence, then, the incentive to put a massive Dr. Evil prequalified space maser up there at a cost of hundreds or thousands of dollars per watt, vulnerable
A very few smartphone manufacturers are making phones already that are shockproof and waterproof. My Casio, for example, is supposed to survive 30 minutes one meter underwater or being dropped a couple of meters onto concrete. All it takes is sealed caps on the ports (although I don't plan to TEST my Casio until I do it accidentally:-). My wife's iPhone, OTOH, was fried by a single lousy drop of water in its enormously wide and entirely nonstandard (well, except for Apple's own "standards") charging/playing port. Which voids the warranty (and of course trips a little colored switch so that they know you voided the warranty and all service plans) which then means you can drop a few hundred more getting a replacement even if you have insurance. So much for Apple's "superior" engineering...
Personally, I hate all of the charging ports. They suck. Micro and regular USB arguably suck the least -- at least they are open standards and one charger or USB cable works for all devices more or less. But they still die (unless sealed when not in use) with as little as a single drop of water. The cost of one phone is far, far greater than the cost of wasted power over the lifetime of ten phones, I would guestimate. But still, why can't anyone seem to engineer an unbreakable, waterproof charging port? How hard can it be?
Any nation whose entire system of weights and measures is fundamentally based on the weight and size of the barleycorn can't be all bad. One inch is three barleycorns, one grain is the weight of one barleycorn, and of course all of the volume measurements are related to the mouthful, the volume of barley-brewed beer or whiskey a man can reasonably consume in one gulp. Let's all drink to that!
16 atmospheres is a better way of putting the entire discussion. One does have to wonder if the 0.1 MPa = 1 atmosphere (more or less) as an eruption threshold is relative to atmospheric pressure or above it. I'd be a bit surprised to find any pressures below it. It's also the (surplus) pressure at the bottom of 160 meters of water. That's a lot to contain if an earthquake or godzilla creates a defect in the dome that leads to 1 atmosphere air.
But hey, if the historical record teaches us anything, it is that volcanoes happen, sometimes rather explosively.
I started out as a Unix sysadmin (professionally) in 1986/1987, initially with a Sun 386i workstation (but running a network whose core was a PDP 11 and a Sun 4/110 server, with various Sun 3's and an early SGI box), but grew with the network until it spanned some forty systems, mostly Suns but with a mix of SGIs (including a big SGI refrigerator compute server -- 220S?), Sparcstations, ELCs and SLCs, and the rebuilt Sun 4/310 which was eventually replaced with a Sparc Ultra server. Somewhere around 1993 or 1994 I first tried Linux (on a friend's 486 IIRC) and installed it myself at home on an early non-Intel P5 clone -- I got Slackware to install on my home 486 with 4 MB but it was "sad" and it did much better on the larger faster AMD. I built one of the first distributed parallel supercomputers on top of dual Pentium Pros -- it might have BEEN the first, I don't know for sure, we had to use the very first 2.0 kernels which had locking issues with the network (which I helped debug, for at least one network driver) and also had to screw around mightily with disk drivers as very few hardware manufacturers at the time helped AT ALL with linux drivers. A few years later I helped oversee the department's gradual transition from Sun to Linux in general -- SunOS 4.x was great, but Solaris was named Slowaris for a reason (broken kernel scheduler) and by the time the kernel spin lock issue was resolved Linux handily outperformed it on PPro or beyond hardware, EXCEPT in server context so we kept a quad Sparc server for a long time afterwards. Around that time I stopped being the only sysadmin for the Physics department and watched as we gradually transitioned away from Slackware and towards Red Hat (largely because of the scalability of Kickstart, which still rocks, BTW). Maybe a decade ago at this point we hired this guy named Seth Vidal as our primary sysadmin, and boy what a great decision that was! He experimented with a tool called "yup" -- the Yellowdog linux installer for Apple/Linux (based on kickstart), ported it to our department for general use, and forked it into "yum", which is now more or less the standard install tool for RH-derived kickstartable linuces because it is AWESOME. I made the mistake of writing the first YUM HOWTO and even though I have no time to update it and it is largely obsolete, it is still one of the most hit parts of my personal webspace. Fortunately Seth now works for Red Hat and Yum has long since transcended its documentation.
These days I still use linux -- currently mostly Fedora 14 (because Gnome 3 sucks, sorry, but that's a fact) although I have bit the sorry bullet and put F16 on a few systems, waiting/hoping for the Gnomergens to come to their senses and dial back to 2 and try again. It's the Linux equivalent of the Blizzard Major Fail with Diablo 3 compared to 2 -- they didn't need to fix anything at all, just add more better, and instead changed all sorts of unbroken things and made a damn usable interface miserably useless. Although I tracked Microsoft from Dos 1.0 on the 64K motherboard IBM PC (yes, that is K, not M) up through Windows 3.2, these days if I use Windows at all it is under VirtualBox as a linux application. But I rarely have to do that, and if it weren't for games and a very few proprietary packages I still need occasional access to I wouldn't do it at all.
Exactly, but automagically and smoothly. You could turn off half of your SUV's motor cruising at constant speed on level ground -- more of it off going downhill -- and save another 1-2 MPG. I think it would be pretty easy to double the mileage of a typical SUV without compromising power or towing capacity.
Hi, this is something (that as an Excursion owner and physicist:-) I've often thought about. The solution, however, is not to build a hybrid electric-diesel engine. It is to build a gasoline-gasoline OR diesel-diesel hybrid. The technology for doing this has been around forever, but sadly, nobody has implemented it. Here's how it works.
If you actually own one of these vehicles, you know that there are three dominant sources of energy waste. I've owned both the 10 cylinder gas Excursion and the 8 cylinder gas Excursion, so I can directly compare my observations of fuel consumption using the built in trip computer. One of the largest ones (if not the largest one) is idle time. In the 10 cylinder version this was extreme -- sitting at a stop sign involved all ten cylinders pounding along, dropping average mileage visibly from any reset. Idle time alone seemed to be one of the largest "costs" of city driving -- it wasn't so much stopping and starting up again (although that was an important part too) as it was stop signs, which hit you BOTH by wasting your KE AND by burning gas keeping all those cylinders going.
The second controllable source of waste is engine size. One way electric-gas hybrids save is by allowing the electric engine to provide high torque during acceleration so that one can use a much smaller motor when cruising. Big cars with fixed numbers of cylinders, however, have little choice. They either have good fuel economy while cruising (and lousy torque at low speeds) or they have enough cylinders and power to get good torque at low speeds or for towing and lousy fuel economy. The 10 cylinder Excursion had great torque (for a four ton vehicle) but mediocre fuel economy at around 11 mpg, where the 8 cylinder has lousy torque -- it struggles just getting up a hill, especially towing my boat -- but can turn in 13-15 mpg on the highway.
Both of these problems are so very easily fixable by simply redesigning the gasoline engine so that it is a set of modular ganged pairs of cylinders (pairs to minimize operational vibration by symmetrizing their motion) that can be turned on and off at will in real time as the needs for torque vs idle vary. Take my Excursion. With 5 pairs of cylinders that are geared so that they smoothly go on and add their torque to the total as the accelerator is pressed (trivial for any sort of computerized electronic ignition system these days) one can idle at a stoplight on two, burning less gasoline than a non-hybrid four cylinder car! Indeed, one could probably dedicate one pair of cylinders JUST to idle and overdrive and make them deliberately smaller to burn about as much gas as a lawnmower during idle.
Then, when one accelerates away from the stop, one successively kicks on and in the pairs of cylinders. For a period of maybe 10 to 20 seconds, the car is a 10 cylinder car and you pull smoothly away to merge, get up to speed (including your boat or whatever) and sure, you burn gas during this stretch. But then one merges at speed, or reaches the normal speed of traffic. You no longer need 10 cylinders to provide the torque that overcomes just level-ground wind resistance and friction. My old Excursion burned on all 10 anyway, and got an easy 20% worse highway mileage than the 8 cylinder I have now, but I'd get great highway mileage on level ground if it ran on only 4 to 6, which is all it really needs to overcome wind, and kicked in the other 4 to 6 only when I hit a hill, a wind, or need to pass.
This idea is perfectly capable, as one can see, of stretching out the mileage of a heavy SUV without compromising torque by anywhere from 20% to 50%, without requiring a half-ton of batteries, a huge electrical motor, an electrical recovery system like that of the Prius, and so on. One could implement it "tomorrow" upon doing the absolutely straightforward engineering of the cylinder pairs, and if one made them modular one could fix another great evil of automobile manufacture, the fact th
Please note that I did not say CO2 wasn't a cause of warming, I simply claimed that many models are flawed.
Difficult to refute, phrased that way. I'm sure that they are. But not necessarily through the use of back-radiation in the model, although frankly I don't care for it either simply because it is so open to pointless argument.
That's fine, since I didn't reference Olsen at all. Did you even visit the link and view both sides of the discussion to which I *WAS* referring?
Sky Dragon is Olsen. He loves crank science. He finds it and puts it on his website. Nobody sane actually argues that the GHE is a cold system "warming" a warmer system, nor is that what climate models implement. It simply slows the cooling of a warm system heated by a system warmer still. Olsen -- in direct communications to me -- has asserted that the sun warms the Earth primarily in the IR band, I suppose to avoid that "warming by a system warmer still". That's insane. I confess to your accusation of not visiting, though. I've spent too much time on Sky Dragon recently and got fed up with the pompously presented crank physics. Which is easy to do. I'll even apologize for (perhaps mistakenly) assuming that you were trying to assert that the GHE doesn't exist, that the Earth is warmed by the Sun primarily via IR, and so on. Sky Dragon is not my idea of an authoritative source of information these days but you are right, I shouldn't assume that just because you reference something there that you are a crank too.
Outside of that, I don't much care what insane people think about the GHE, but I do try to point out what it really is in the fond hope that sanity will be restored to them if only I explain it clearly enough...;-)
As for back radiation, downwelling radiation, scattered radiation, blackbody radiation, emissivity and Kirchoff's law, extinction and optical path -- as far as the actual processes involved are concerned there isn't a huge difference between one kind of optical scattering and another, not at the molecular scale. I can refer you to your choice of graduate physics textbooks or a fairly good book of physical meteorology if you want to review or learn it someplace that simply presents it and the data that support it. I teach a lot of the basic physics at both the graduate and undergraduate level (and have written one of those textbooks that covers at least part of this), but of course you are right, I might not be competent and the field of meteorology and climate modeling is broad enough that even if I were, I could be mistaken about lots of things in one context or another.
As for Rayleigh scattering not being "back radiation" -- I was merely offering that as an example of how molecules can and do reflect and scatter photons in a way that varies with frequency. The photons that are (resonantly or otherwise) back scattered from atmospheric CO_2 near the surface -- whatever you want to call them -- reflect some fraction of the otherwise outgoing energy in those bands towards the source quite independent of the relative temperature of the source that is radiating. One can directly measure this spectrally resolved back scattered radiation in bottom of atmosphere spectrographs, looking up. Call it "back radiation" or not, call it whatever you like, but it is a term that acts as a differential gain in the heat balance of the ground, at least if you believe in Maxwell's equations and that the power incident on a surface is the flux of the Poynting vector through the surface.
In models that focus on ground temperatures, it is thus pretty natural to have it and easy to semi-empirically justify it. On top of atmosphere models looking down (which is what I personally prefer when trying to convince people that do not want to "believe" in the GHE because they have a hard time with differential gain terms in resonant absorptive bands or get lost in the thermodynamics of multi-channel cooling processes at the surface)
I've just been participating in a rather extended round of debate with Olsen (sky dragon slayer). Two comments again:
a) CO_2 warming models don't "rely" on back radiation. They are inferable from simple subtraction, using utterly empirical evidence. Find a website that shows top of atmosphere spectrographs, ideally ones taken at night over e.g. the arctic so that you can eliminate the confusion of reflected sunlight and focus only on radiation given off by the ground as it cools. Look at the hole in the CO_2 absorption/scattering band. The total power radiated from the area being photographed must, on average, match the total rate at which heat is delivered there by all sources. The CO_2 hole is clearly visible in daytime, equatorial, temperate, polar spectrographs. If one integrates the outgoing flux of the Poynting vector over the entire sphere above the TOA, this has to equal the rate power is delivered to the interior of that sphere. Since energy lost from the surface via blackbody radiation is blocked in the CO_2 band, the surface must warm up until total energy lost is (still) equal to the total energy input (averaged over both the surface and time and all wavelengths). End of story.
b) For what it's worth, one can actually take bottom of atmosphere spectrographs from the same place at the same time as one takes the TOA spectrographs. This has been done. The spectrographs compliment themselves (at least in the arctic where there is little water vapor or confounding signals from other stuff going on). One can see a nearly perfect match between downwelling radiation in the blocked bands so clearly visible up above.
IMO fairly professional opinion as a physicist, Olsen is not competent; his arguments are not sound, nor are they in any way quantitative (that I've been able to determine). For example, when he discusses how fission "must" be a missing source of energy and the real cause of climate change instead of CO_2 modulation, solar output modulation, albedo modulation, global circulation decadal oscillation modulation he fails to provide any sort of quantitative or plausible computation of the actual energy production one might expect in the interior, relying instead on a verbal/heuristic argument that reduces to "the interior is very hot, therefore fission must be important". I've tried fairly patiently to walk him through the spectrographic data -- which are for all practical purposes photographs of the greenhouse effect in action -- to no avail -- he doesn't seem to understand electromagnetic radiation theory very well (as in, as well as a bright undergrad physics major). Some things he states are blatently silly -- the assertion that a "space blanket" (reflective mylar sheet) works by blocking convection instead of trapping radiation (it's both, but the human body loses heat primarily through radiation, a simple fact that can once again be directly photographed). And he has somehow taken an experience where he was lost inside a cold cloud when he first learned to fly and turned it into an entire theory of cloud cooling.
For what it is worth, you can see back radiation and side radiation and all sorts of radiation from the sky with your eyes. Blue sky? Rayleigh scattering. Live in a city and want to do some stargazing? Sorry, too much backscattered radiation from the city lights, worst viewing on hazy nights (lots of greenhouse gas H_2O in the air), best on clear, cold, dry nights in the winter (not so much greenhouse H_2O in the air. Sure, you only see visible light "back radiation", but that's because of limitations in your eyes, not because it isn't there in other wavelengths of invisible light as well (such as infrared). In the CO_2 absorption band in the IR, the sky is so "hazy" it is completely opaque. Think of it as being a thick fog, visibility a few hundred feet. Not all radiation that is emitted from the surface in that band is reflected back -- some diffuses through to eventually escape above at a mu
Two comments -- just to be picky. One: Read Taleb's "The Black Swan". It is basically a systematic proof that in fact most experts aren't experts, with narrow exceptions in non-complex scientific fields such as physics. In complex systems, experts in fact are rarely experts, and almost invariably claim more knowledge than time and data prove that they have/had. It's quite understandable -- in the end you can understand why many experts aren't expert.
Two, chaotic systems are often made less so by increasing a driver. In fact, many of them have narrow parametric regions where they are chaotic, and if you move any parameter out of that region the system stabilizes.
As a single example, the most violent weather tends to occur when warm fronts and cold fronts are in close proximity, when/where high pressure systems and low pressure systems collide or interact. For any given heat input, temperature differentials on the surface of the Earth actually increase cooling efficiency because outgoing power is radiated proportional to the fourth power of the temperature but only the second power of the relevant surface length scale. The more uniform the temperature, the warmer the average temperature. It is therefore entirely possible for a warming climate to have more uniform temperatures and less violent weather. It is similarly quite possible for a globally cooling climate to be setting local temperature records (concentration of heat in a comparatively small area, from which it is relatively rapidly lost) while only cooling very slightly elsewhere, and to have more violent weather when cold fronts impinge on those heated areas.
I have code and descriptions if you want to numerically study a very simple actual chaotic system (or two, or three) so that you can see for yourself that you have to drive it at just the right frequencies, amplitudes, and dampings to observe a Feigenbaum tree (period doubling into chaos) and equally rapid emergence from the chaotic regime as you increase amplitude or frequency or damping. That doesn't make this a universal truth about chaotic systems, BTW, it just points out the danger of making sweeping statements about something you don't really know much about. One could go on -- is there a proof that adding more CO_2 creates greater instability? What, exactly, is greater instability (how do you define it)? I fully agree that adding more CO_2 (e.g. taking it to 600-700 ppm by 2100) is likely to raise global temperatures by some amount (the exact amount is a matter of considerable debate even among experts with a lower bound that is just over nothing).
It is by no means clear -- and to the best of my knowledge there is no statistically sound evidence to support the conclusion that -- the warming of the late twentieth century resulted in "greater instability" in the form of more violent weather, nor has any other kind of "instability" other than the motion of the mean global temperature itself been convincingly demonstrated. It has been drier, wetter, stormier, hotter, colder, both locally and globally, in the past without CO_2 forcing.
The really interesting thing is that many climate scientists are quite open about their lack of certain knowledge in climate science -- in a scientific forum where they might be called on it if they utter something really speculative as if they are sure. A George Mason survey of actual climate scientists found that roughly one in seven think that there will be little to no warming and no catastrophe by 2100. Over half think that there will be significant, but probably not catastrophic warming. In the end, I agree with you -- this honest lack of consensus among climate scientists probably rates some consideration.
For one thing, it makes the entire field more credible. When was the last time you were in a room full of scientists who agreed about everything, even important things for which there is far better experimental data and far more computable theory
Or, all together were probably less than Tambora: http://en.wikipedia.org/wiki/Mount_Tambora -- at 800 Mt, considerably less. Tambora holds many records: Largest explosion in recorded history, loudest sound in recorded history, largest single-event influence on the climate in recorded history (it basically eliminated "summer" for two years in a row in at least some temperate latitudes) and helped make the decade of 1810 the coldest decade on historical record (but not the coldest year or part of the coldest half-century or century).
But I don't think we can be certain of the effect of the nuclear tests. Many of the largest were low over water and kicked a lot of water into the stratosphere. We just don't have the data, and hence any conclusions are likely to be guesses.
Hmm, I guess it's time to go pop the cap off of a tasty homemade beer, one that would easily cost $2 a bottle in a store or $5 a bottle in a restaurant. Not that I disagree that most people think it is too difficult to make and would rather buy cheap swill in large quantities of cans.
It is a legitimate field of study. One where negative results should have been (and eventually were) published. And also one where confirmation bias produced "surprising" results indeed -- until you looked for the man behind the curtain.
Slashdot Mirror
You mean like the Fedora installer that pops up when you boot the live disk and says "Click here to install to disk"? It takes maybe five minutes to put Fedora 17 live onto a hard drive, saying "use entire disk" when it asks. Not as good as dd if=/dev/urandom of=/dev/sda' though (preferred because somebody who really really tried could probably resurrect the old data through a mere zeroing, but it would be a lot harder if overwritten with random bytes, and way harder if you executed this command five or six times in a row).
Yeah, I'm having a few problems with the idea. Temperature actually implies thermal equilibrium, which in turn requires interaction. However, those atoms/molecules are, shall we say cosmicly non-interacting, being so dispersed that they basically form a hard vacuum. There are then a number of problems with the picture. One, why do the molecules not simply fall back into the galaxy (or if you like, why were they pushed out of the galaxy in the first place)? Several billion years accumulation of solar wind and outflow from supernovae? Second, why don't they "cool" to equilibrium with the background blackbody radiation? That one at least is answerable -- they basically never collide (as you say). The third question is -- how can one detect the gas? The molecules are cold, non-interacting, and enormously diffuse. To the extent that they interact with photons, they would relatively quickly slow down to equilibrium, so they are almost by definition invisible. Furthermore, photons are photons, and being a gas detecting the origin of a photon approaching the Earth from any given direction (that is, the distance of the source) is quite impossible, ditto for absorption lines from distant stars. I mean, parallax won't work. Neither will red shift. Alterations of gravity (as in the way one infers "Dark Matter") might work, but won't give you the speed or "temperature" or ensure that the matter that is altering orbits is a gas of baryons as opposed to darkonium.
Indeed, the title is horribly misleading. It would be better to say that the galaxy may be surrounded by a very diffuse gas of particles at a very low temperature in a reference frame that has a very high velocity relative to the galactic core, and we infer this compared to all competing explanations by sacrificing a chicken with a black-handled knife on the keyboard of a computer.
rgb
All? You mean there is more than one? And here I thought that it was just some individual's internet handle...
Also, listening to music and tapping your feet has nothing to do with multitasking. These activities are sent to the background and can perfectly happen while you're focusing on a specific task. Having to lose focus all the time because of telephone calls is completely different and is a serious productivity killer.
Fair enough. However, the study was also apples and oranges -- that was my primary point. It measured the degradation in utility executing a single primary task without allowing for the utility of the interrupting task, or for the value of opportunity cost time. You may find it more pleasurable and less crazy-making to work on just one task to the exclusion of others, but your employer may find far greater utility in your employment by having you answer the phone on an interrupt-driven basis (that is, when it rings) to handle whatever problem associated with the call and work on longer term but less urgent project in between. The degradation in performance in that task may not matter to them as long as they don't have to hire another person to largely sit idle and "just" answer the phone, especially if they'd have to hire somebody with more or less your skillset to do the answering (e.g. providing deep support for some technical product).
If they'd measured the actual degradation in the value of the total of a person's work-related activities, that would have been very useful, especially if they'd done so in a way that could actually be rescaled and used to do quantitative optimization. But measuring the degradation in primary task execution subject to random irrelevant "noise", especially on an unrealistically short time granularity, is both less useful and verifying a proposition that nobody would have doubted anyway. Yes, if you slap somebody in the face or give them an electrical shock according to some randomized schedule with a mean time between events that is order of a minute or two, you will interfere with their productivity (and driven them quietly crazy), but who cares? If you force them to switch complex tasks every forty seconds, I'm certain that for most people it will degrade their productivity on both tasks compared to doing them without the switch (although people are quite variable and there may be people that can manage this for at least certain classes of task especially with practice) but how much their productivity degrades at least might be relevant, especially if you explore the task degradation quantitatively as a function of the switch interval and across a decent population (to have a decent chance of sampling outliers who do much better or much worse).
The latter's the rub. My wife (a physician) cannot work with information with music playing -- too distracting. I can and often do. My wife can juggle certain sets of "simultaneous" tasks -- tasks she switches to and from on demand -- while others drive her crazy. I do much better at general multitasking and usually have a much broader set of things I'm working on at any given time and can switch between them fairly easily at reasonable granularity. For example, I'm writing this now, but will be clearing email in a minute or two (and was clearing email before getting the message announcing this reply to my slashdot comment) and am in the middle of at least a dozen project getting slices of my time as opportunity and need dictate. Different people have very different tolerances and abilities in this regard, even for "random, annoying" interrupt-driven subtasks like letting the dog out or answering the phone or working someplace where people are hustling and bustling all around you. I'm not at all certain that even publishing an average degradation on memory or completion or error rate is useful -- the distribution itself might be, because it might be multimodal or have long tails or otherwise not be classically "normal" so that the mean has some relevant meaning.
The key is to know your OWN abilities and inclinations in this regard, and to
Except that is not what they tested, is it? They tested people on whether or not they could work efficiently if they are distracted, not whether or not multitasking improves or doesn't improve efficiency.
The silly thing is that we actually know quite a lot about task organization for efficient multitasking. It is a key component of task scheduling on any computer. Fine grained multitasking -- especially on a CPU that has a large latency component for switching tasks, is known as "thrashing", and is also known to degrade performance substantially, and whacking a computer with a steady stream of pointless interrupts so that it is always thrashing slows it down.
At the same time, executing tasks with the right granularity and with the right kinds of latency and parallelism can speed things up quite a lot compared to doing tasks one at a time. This, too, is true in life as much as it is in computers. Anybody who cooks knows that you cannot generally make a good meal in serial fashion. If you want to serve rice with a stir fry and end up with dessert in a timely manner, you have to be cooking the rice, chilling the dessert, and chopping up and frying the main course all "at the same time", with layered overlaps in the attention you pay to the different tasks. The tasks are all related and a skilled cook can juggle quite a few of them without cognitive or operational overload and finish a meal far faster than anyone would ever finish it cooking one thing at a time (to a soggy, cold, unproductive finish).
Most normal humans multitask all the time. I listen to music and work while wiggling my feet to maintain circulation. I hop from answering email to posting silly things like this reply to doing work on task A to doing work on task B to doing work on task C -- so much the more so when my tasks are all different, all use the computer (or a number of computers) and take different amounts of time (attended or unattended) to move on to the next stage of completion. Yes, I can be overloaded, I can thrash in my normal work if overloaded so little gets done, but that is entirely different from asking if I can work when somebody is randomly blasting uncorrelated and meaningless distractions into my workspace.
rgb
Actually, programming is more like writing a cookbook. Lots of people who cook can't write a cookbook. For example, illiterate ones. Similarly, it helps to be literate in a computer language in order to be able to program, which in turn requires an above average ability to deal with a peculiar kind of metaphor and an above average understanding of tools far more demanding than a source of heat, a knife, and some ingredients. The thing that prevents old women or old men or old monkeys or old dogs from moving from food to programming is a mix of intelligence, interest, and motivation.
It is, for example, fairly common belief that programmers make an income that is well above average. It's a common belief because it is true. Yet you don't see teen-age fry cooks piling in to programming to multiply their minimum wage income by close to an order of magnitude. Why not? Because programming is difficult, and you have to be both smart enough to do it and inclined to WANT to do it, regardless of the obvious rewards of working in an air conditioned environment sitting on your ass while making $60K/year or more with benefits compared to slinging greasy burgers at possibly armed and dangerous clients in a Burger King late at night for $200 a week on a good week -- ooo, and then there are those pesky social security deductions and a manager that laughs hysterically if you mention the word "benefits" right before he fires you.
rgb
I would have to disagree. Programming isn't algebra, it is metaphor. It is building machines out of words. In fact, I don't know if you meant it humorously, but you yourself say "define the WORD algorithm". I'd argue that most algorithms CAN be described in words. Often even English or Swahili words.
I'm not sure, but I think that's why they call the medium a "programming language". It isn't devoid of math or logic -- far from it -- but it isn't the same thing, either, and a person good at real mathematics can easily suck at programming or vice versa.
Here's a program:
"Take an integer variable named i, fill it sequentially with the integer values from one to one hundred in steps of one, and print out its value on the screen I'm looking at".
This is perfectly understandable English, and can be executed by a reasonably bright student to whom the words are directed where the screen is a whiteboard.
It isn't really any different when written "for (i=1;i <=100;i++) printf("%d\n",i);" or any of the myriad other ways of writing the same program in different programming languages. A machine built out of words/symbols that have a fairly carefully specified operational meaning. Not at all like proving that the angles in a plane triangle add up to pi.
rgb
Um, did you mean something IN Basic?
Just checking...
Although it is sadly all too likely to be true... even today.
rgb
And don't forget the Flynn Effect. It isn't even a constant 100, so 100 this decade doesn't mean what 100 meant last decade.
But still, no, it is not true that my brother (with Down's Syndrome) could have ever become a programmer, whereas I, with an IQ (FWIW) several standard deviations over the mean have gigabytes of source in my source directory, a rather large fraction of which I actually wrote, in several languages.
So the question is still a stupid question, as the answer is obviously no. Worse, it is basically trollbait BEYOND being a stupid question, as somebody of "normal" intelligence can probably write "a program" in a sufficiently simple environment without ever in their lifetime being capable of writing a 50,000 line program with 100 functional modules written on top of various APIs (some of which they created) running over the network on top of UDP socket layer code. Actually, a lot of fairly ABOVE average intelligence well-trained programmers might fail there, or do a poor job if they succeeded.
So the proper answer is "No, to be a good programmer you have to be smarter than the average human, and probably better educated too. Propensity to skip showers and live on Jolt Cola optional. Troll."
rgb
space-based solar power arrays parked in geosynchronous or near-geosynchronous orvbit.
Ah, I once thought as you do, but then a measure of common sense asserted itself. Consider the fact that the cost of getting to geosynchronous orbit is, per kilogram, larger than the energy output of a kilogram's worth of cells over a lifetime of "forever" (or damn near). Consider further that a gigawatt's worth of space array, beaming its energy back to the ground (at some cost in efficiency, transmission losses) is more or less a gigawatt-scale space weapon if it is aimed somewhere other than whatever patch of ground set aside as a receiver. What can go wrong? Consider that you can avoid this problem, sure, by using a very weak beam, but then you have to use a very large piece of ground as a receiver, one that increases in size with the geometry of latitude giving you a second trade-off between area of receiver and atmospheric loss at higher latitudes versus the difficulty of very long distance power transmission from the equator to the temperate zone. Consider that TOA insolation is only a factor of two or so larger than BOA insolation (so it's not like you get a lot more power by being out of the atmosphere) and land is cheap in the desert, and there is plenty of desert. Finally consider that land is REALLY cheap on your own rooftop, which very likely contains ALMOST enough area to completely supply your own house's energy needs and can "store" energy by simply dumping surplus back into the grid during the day at reverse cost to be drawn out again at night "for free", even without an ever-improving local storage option.
Consider that the cost of actually putting 5 kW of solar cells on your roof NOW is more than break even on a 20 year amortization or less (in many parts of the country) with the amortization schedule dropping with the cost of solar cells and other improvements in the technology. The cost of solar cells per delivered watt has been dropping exponentially with a halving time of around a decade for the last three or four decades. It is currently between $1 and $2 per watt, plus installation and hardware costs. At $1/watt -- already available to large commercial buyers -- the amortization time for a 5 kW rooftop installation is order of a decade: it will generate order of $1000 worth of electricity per year, enough to pay off a $7000-8000 loan and even make a profit over that time. I've spent more than that on high efficiency furnace/AC for my house -- several times over, sadly -- with an even longer amortization. And, of course, anything that is "profitable" on the scale of individual rooftops is far MORE profitable on an industrial scale with industrial economies of scale. $1/watt retail is $0.50/watt wholesale in volume, and even allowing for installation and operation and maintenance costs, POWER COMPANIES will be GIVING you units to put on your roof -- as long as they can sell you slightly discounted power from those units. Or building large arrays themselves, but then they have the pesky problem of buying kilometer-square chunks of land here and there.
So the real problem with putting solar cells in space is that if the price drops, as one can very reasonably expect, to under $1/watt full retail over the next decade, solar generation will proliferate like a weed all over the world not to save the whales or lower carbon footprint but because it is the cheapest or second cheapest way to make electricity. This will happen even if there ARE no breakthroughs in gigawatt-scale 24 hour plus storage, although I personally think that physicists and engineers will beat the storage problem too within the next decade -- the payoff for doing so is huge. Sure, we'll still need bridge power -- nuclear and probably coal or natural gas -- but the actual draw on those facilities will decrease to a fraction of what it is today. Whence, then, the incentive to put a massive Dr. Evil prequalified space maser up there at a cost of hundreds or thousands of dollars per watt, vulnerable
A very few smartphone manufacturers are making phones already that are shockproof and waterproof. My Casio, for example, is supposed to survive 30 minutes one meter underwater or being dropped a couple of meters onto concrete. All it takes is sealed caps on the ports (although I don't plan to TEST my Casio until I do it accidentally:-). My wife's iPhone, OTOH, was fried by a single lousy drop of water in its enormously wide and entirely nonstandard (well, except for Apple's own "standards") charging/playing port. Which voids the warranty (and of course trips a little colored switch so that they know you voided the warranty and all service plans) which then means you can drop a few hundred more getting a replacement even if you have insurance. So much for Apple's "superior" engineering...
Personally, I hate all of the charging ports. They suck. Micro and regular USB arguably suck the least -- at least they are open standards and one charger or USB cable works for all devices more or less. But they still die (unless sealed when not in use) with as little as a single drop of water. The cost of one phone is far, far greater than the cost of wasted power over the lifetime of ten phones, I would guestimate. But still, why can't anyone seem to engineer an unbreakable, waterproof charging port? How hard can it be?
rgb
Any nation whose entire system of weights and measures is fundamentally based on the weight and size of the barleycorn can't be all bad. One inch is three barleycorns, one grain is the weight of one barleycorn, and of course all of the volume measurements are related to the mouthful, the volume of barley-brewed beer or whiskey a man can reasonably consume in one gulp. Let's all drink to that!
rgb
16 atmospheres is a better way of putting the entire discussion. One does have to wonder if the 0.1 MPa = 1 atmosphere (more or less) as an eruption threshold is relative to atmospheric pressure or above it. I'd be a bit surprised to find any pressures below it. It's also the (surplus) pressure at the bottom of 160 meters of water. That's a lot to contain if an earthquake or godzilla creates a defect in the dome that leads to 1 atmosphere air.
But hey, if the historical record teaches us anything, it is that volcanoes happen, sometimes rather explosively.
rgb
...because Gnome 3 sucks. Seriously. What were they thinking?
rgb
I started out as a Unix sysadmin (professionally) in 1986/1987, initially with a Sun 386i workstation (but running a network whose core was a PDP 11 and a Sun 4/110 server, with various Sun 3's and an early SGI box), but grew with the network until it spanned some forty systems, mostly Suns but with a mix of SGIs (including a big SGI refrigerator compute server -- 220S?), Sparcstations, ELCs and SLCs, and the rebuilt Sun 4/310 which was eventually replaced with a Sparc Ultra server. Somewhere around 1993 or 1994 I first tried Linux (on a friend's 486 IIRC) and installed it myself at home on an early non-Intel P5 clone -- I got Slackware to install on my home 486 with 4 MB but it was "sad" and it did much better on the larger faster AMD. I built one of the first distributed parallel supercomputers on top of dual Pentium Pros -- it might have BEEN the first, I don't know for sure, we had to use the very first 2.0 kernels which had locking issues with the network (which I helped debug, for at least one network driver) and also had to screw around mightily with disk drivers as very few hardware manufacturers at the time helped AT ALL with linux drivers. A few years later I helped oversee the department's gradual transition from Sun to Linux in general -- SunOS 4.x was great, but Solaris was named Slowaris for a reason (broken kernel scheduler) and by the time the kernel spin lock issue was resolved Linux handily outperformed it on PPro or beyond hardware, EXCEPT in server context so we kept a quad Sparc server for a long time afterwards. Around that time I stopped being the only sysadmin for the Physics department and watched as we gradually transitioned away from Slackware and towards Red Hat (largely because of the scalability of Kickstart, which still rocks, BTW). Maybe a decade ago at this point we hired this guy named Seth Vidal as our primary sysadmin, and boy what a great decision that was! He experimented with a tool called "yup" -- the Yellowdog linux installer for Apple/Linux (based on kickstart), ported it to our department for general use, and forked it into "yum", which is now more or less the standard install tool for RH-derived kickstartable linuces because it is AWESOME. I made the mistake of writing the first YUM HOWTO and even though I have no time to update it and it is largely obsolete, it is still one of the most hit parts of my personal webspace. Fortunately Seth now works for Red Hat and Yum has long since transcended its documentation.
These days I still use linux -- currently mostly Fedora 14 (because Gnome 3 sucks, sorry, but that's a fact) although I have bit the sorry bullet and put F16 on a few systems, waiting/hoping for the Gnomergens to come to their senses and dial back to 2 and try again. It's the Linux equivalent of the Blizzard Major Fail with Diablo 3 compared to 2 -- they didn't need to fix anything at all, just add more better, and instead changed all sorts of unbroken things and made a damn usable interface miserably useless. Although I tracked Microsoft from Dos 1.0 on the 64K motherboard IBM PC (yes, that is K, not M) up through Windows 3.2, these days if I use Windows at all it is under VirtualBox as a linux application. But I rarely have to do that, and if it weren't for games and a very few proprietary packages I still need occasional access to I wouldn't do it at all.
rgb
All we need is a John Carter now to lead a desperate mission to keep the atmosphere machine running... a half-trillion dollars or so later.
Damn, for some mod points...
Exactly, but automagically and smoothly. You could turn off half of your SUV's motor cruising at constant speed on level ground -- more of it off going downhill -- and save another 1-2 MPG. I think it would be pretty easy to double the mileage of a typical SUV without compromising power or towing capacity.
rgb
Hi, this is something (that as an Excursion owner and physicist:-) I've often thought about. The solution, however, is not to build a hybrid electric-diesel engine. It is to build a gasoline-gasoline OR diesel-diesel hybrid. The technology for doing this has been around forever, but sadly, nobody has implemented it. Here's how it works.
If you actually own one of these vehicles, you know that there are three dominant sources of energy waste. I've owned both the 10 cylinder gas Excursion and the 8 cylinder gas Excursion, so I can directly compare my observations of fuel consumption using the built in trip computer. One of the largest ones (if not the largest one) is idle time. In the 10 cylinder version this was extreme -- sitting at a stop sign involved all ten cylinders pounding along, dropping average mileage visibly from any reset. Idle time alone seemed to be one of the largest "costs" of city driving -- it wasn't so much stopping and starting up again (although that was an important part too) as it was stop signs, which hit you BOTH by wasting your KE AND by burning gas keeping all those cylinders going.
The second controllable source of waste is engine size. One way electric-gas hybrids save is by allowing the electric engine to provide high torque during acceleration so that one can use a much smaller motor when cruising. Big cars with fixed numbers of cylinders, however, have little choice. They either have good fuel economy while cruising (and lousy torque at low speeds) or they have enough cylinders and power to get good torque at low speeds or for towing and lousy fuel economy. The 10 cylinder Excursion had great torque (for a four ton vehicle) but mediocre fuel economy at around 11 mpg, where the 8 cylinder has lousy torque -- it struggles just getting up a hill, especially towing my boat -- but can turn in 13-15 mpg on the highway.
Both of these problems are so very easily fixable by simply redesigning the gasoline engine so that it is a set of modular ganged pairs of cylinders (pairs to minimize operational vibration by symmetrizing their motion) that can be turned on and off at will in real time as the needs for torque vs idle vary. Take my Excursion. With 5 pairs of cylinders that are geared so that they smoothly go on and add their torque to the total as the accelerator is pressed (trivial for any sort of computerized electronic ignition system these days) one can idle at a stoplight on two, burning less gasoline than a non-hybrid four cylinder car! Indeed, one could probably dedicate one pair of cylinders JUST to idle and overdrive and make them deliberately smaller to burn about as much gas as a lawnmower during idle.
Then, when one accelerates away from the stop, one successively kicks on and in the pairs of cylinders. For a period of maybe 10 to 20 seconds, the car is a 10 cylinder car and you pull smoothly away to merge, get up to speed (including your boat or whatever) and sure, you burn gas during this stretch. But then one merges at speed, or reaches the normal speed of traffic. You no longer need 10 cylinders to provide the torque that overcomes just level-ground wind resistance and friction. My old Excursion burned on all 10 anyway, and got an easy 20% worse highway mileage than the 8 cylinder I have now, but I'd get great highway mileage on level ground if it ran on only 4 to 6, which is all it really needs to overcome wind, and kicked in the other 4 to 6 only when I hit a hill, a wind, or need to pass.
This idea is perfectly capable, as one can see, of stretching out the mileage of a heavy SUV without compromising torque by anywhere from 20% to 50%, without requiring a half-ton of batteries, a huge electrical motor, an electrical recovery system like that of the Prius, and so on. One could implement it "tomorrow" upon doing the absolutely straightforward engineering of the cylinder pairs, and if one made them modular one could fix another great evil of automobile manufacture, the fact th
Please note that I did not say CO2 wasn't a cause of warming, I simply claimed that many models are flawed.
Difficult to refute, phrased that way. I'm sure that they are. But not necessarily through the use of back-radiation in the model, although frankly I don't care for it either simply because it is so open to pointless argument.
That's fine, since I didn't reference Olsen at all. Did you even visit the link and view both sides of the discussion to which I *WAS* referring?
Sky Dragon is Olsen. He loves crank science. He finds it and puts it on his website. Nobody sane actually argues that the GHE is a cold system "warming" a warmer system, nor is that what climate models implement. It simply slows the cooling of a warm system heated by a system warmer still. Olsen -- in direct communications to me -- has asserted that the sun warms the Earth primarily in the IR band, I suppose to avoid that "warming by a system warmer still". That's insane. I confess to your accusation of not visiting, though. I've spent too much time on Sky Dragon recently and got fed up with the pompously presented crank physics. Which is easy to do. I'll even apologize for (perhaps mistakenly) assuming that you were trying to assert that the GHE doesn't exist, that the Earth is warmed by the Sun primarily via IR, and so on. Sky Dragon is not my idea of an authoritative source of information these days but you are right, I shouldn't assume that just because you reference something there that you are a crank too.
Outside of that, I don't much care what insane people think about the GHE, but I do try to point out what it really is in the fond hope that sanity will be restored to them if only I explain it clearly enough...;-)
As for back radiation, downwelling radiation, scattered radiation, blackbody radiation, emissivity and Kirchoff's law, extinction and optical path -- as far as the actual processes involved are concerned there isn't a huge difference between one kind of optical scattering and another, not at the molecular scale. I can refer you to your choice of graduate physics textbooks or a fairly good book of physical meteorology if you want to review or learn it someplace that simply presents it and the data that support it. I teach a lot of the basic physics at both the graduate and undergraduate level (and have written one of those textbooks that covers at least part of this), but of course you are right, I might not be competent and the field of meteorology and climate modeling is broad enough that even if I were, I could be mistaken about lots of things in one context or another.
As for Rayleigh scattering not being "back radiation" -- I was merely offering that as an example of how molecules can and do reflect and scatter photons in a way that varies with frequency. The photons that are (resonantly or otherwise) back scattered from atmospheric CO_2 near the surface -- whatever you want to call them -- reflect some fraction of the otherwise outgoing energy in those bands towards the source quite independent of the relative temperature of the source that is radiating. One can directly measure this spectrally resolved back scattered radiation in bottom of atmosphere spectrographs, looking up. Call it "back radiation" or not, call it whatever you like, but it is a term that acts as a differential gain in the heat balance of the ground, at least if you believe in Maxwell's equations and that the power incident on a surface is the flux of the Poynting vector through the surface.
In models that focus on ground temperatures, it is thus pretty natural to have it and easy to semi-empirically justify it. On top of atmosphere models looking down (which is what I personally prefer when trying to convince people that do not want to "believe" in the GHE because they have a hard time with differential gain terms in resonant absorptive bands or get lost in the thermodynamics of multi-channel cooling processes at the surface)
I've just been participating in a rather extended round of debate with Olsen (sky dragon slayer). Two comments again:
a) CO_2 warming models don't "rely" on back radiation. They are inferable from simple subtraction, using utterly empirical evidence. Find a website that shows top of atmosphere spectrographs, ideally ones taken at night over e.g. the arctic so that you can eliminate the confusion of reflected sunlight and focus only on radiation given off by the ground as it cools. Look at the hole in the CO_2 absorption/scattering band. The total power radiated from the area being photographed must, on average, match the total rate at which heat is delivered there by all sources. The CO_2 hole is clearly visible in daytime, equatorial, temperate, polar spectrographs. If one integrates the outgoing flux of the Poynting vector over the entire sphere above the TOA, this has to equal the rate power is delivered to the interior of that sphere. Since energy lost from the surface via blackbody radiation is blocked in the CO_2 band, the surface must warm up until total energy lost is (still) equal to the total energy input (averaged over both the surface and time and all wavelengths). End of story.
b) For what it's worth, one can actually take bottom of atmosphere spectrographs from the same place at the same time as one takes the TOA spectrographs. This has been done. The spectrographs compliment themselves (at least in the arctic where there is little water vapor or confounding signals from other stuff going on). One can see a nearly perfect match between downwelling radiation in the blocked bands so clearly visible up above.
IMO fairly professional opinion as a physicist, Olsen is not competent; his arguments are not sound, nor are they in any way quantitative (that I've been able to determine). For example, when he discusses how fission "must" be a missing source of energy and the real cause of climate change instead of CO_2 modulation, solar output modulation, albedo modulation, global circulation decadal oscillation modulation he fails to provide any sort of quantitative or plausible computation of the actual energy production one might expect in the interior, relying instead on a verbal/heuristic argument that reduces to "the interior is very hot, therefore fission must be important". I've tried fairly patiently to walk him through the spectrographic data -- which are for all practical purposes photographs of the greenhouse effect in action -- to no avail -- he doesn't seem to understand electromagnetic radiation theory very well (as in, as well as a bright undergrad physics major). Some things he states are blatently silly -- the assertion that a "space blanket" (reflective mylar sheet) works by blocking convection instead of trapping radiation (it's both, but the human body loses heat primarily through radiation, a simple fact that can once again be directly photographed). And he has somehow taken an experience where he was lost inside a cold cloud when he first learned to fly and turned it into an entire theory of cloud cooling.
For what it is worth, you can see back radiation and side radiation and all sorts of radiation from the sky with your eyes. Blue sky? Rayleigh scattering. Live in a city and want to do some stargazing? Sorry, too much backscattered radiation from the city lights, worst viewing on hazy nights (lots of greenhouse gas H_2O in the air), best on clear, cold, dry nights in the winter (not so much greenhouse H_2O in the air. Sure, you only see visible light "back radiation", but that's because of limitations in your eyes, not because it isn't there in other wavelengths of invisible light as well (such as infrared). In the CO_2 absorption band in the IR, the sky is so "hazy" it is completely opaque. Think of it as being a thick fog, visibility a few hundred feet. Not all radiation that is emitted from the surface in that band is reflected back -- some diffuses through to eventually escape above at a mu
Two comments -- just to be picky. One: Read Taleb's "The Black Swan". It is basically a systematic proof that in fact most experts aren't experts, with narrow exceptions in non-complex scientific fields such as physics. In complex systems, experts in fact are rarely experts, and almost invariably claim more knowledge than time and data prove that they have/had. It's quite understandable -- in the end you can understand why many experts aren't expert.
Two, chaotic systems are often made less so by increasing a driver. In fact, many of them have narrow parametric regions where they are chaotic, and if you move any parameter out of that region the system stabilizes.
As a single example, the most violent weather tends to occur when warm fronts and cold fronts are in close proximity, when/where high pressure systems and low pressure systems collide or interact. For any given heat input, temperature differentials on the surface of the Earth actually increase cooling efficiency because outgoing power is radiated proportional to the fourth power of the temperature but only the second power of the relevant surface length scale. The more uniform the temperature, the warmer the average temperature. It is therefore entirely possible for a warming climate to have more uniform temperatures and less violent weather. It is similarly quite possible for a globally cooling climate to be setting local temperature records (concentration of heat in a comparatively small area, from which it is relatively rapidly lost) while only cooling very slightly elsewhere, and to have more violent weather when cold fronts impinge on those heated areas.
I have code and descriptions if you want to numerically study a very simple actual chaotic system (or two, or three) so that you can see for yourself that you have to drive it at just the right frequencies, amplitudes, and dampings to observe a Feigenbaum tree (period doubling into chaos) and equally rapid emergence from the chaotic regime as you increase amplitude or frequency or damping. That doesn't make this a universal truth about chaotic systems, BTW, it just points out the danger of making sweeping statements about something you don't really know much about. One could go on -- is there a proof that adding more CO_2 creates greater instability? What, exactly, is greater instability (how do you define it)? I fully agree that adding more CO_2 (e.g. taking it to 600-700 ppm by 2100) is likely to raise global temperatures by some amount (the exact amount is a matter of considerable debate even among experts with a lower bound that is just over nothing).
It is by no means clear -- and to the best of my knowledge there is no statistically sound evidence to support the conclusion that -- the warming of the late twentieth century resulted in "greater instability" in the form of more violent weather, nor has any other kind of "instability" other than the motion of the mean global temperature itself been convincingly demonstrated. It has been drier, wetter, stormier, hotter, colder, both locally and globally, in the past without CO_2 forcing.
The really interesting thing is that many climate scientists are quite open about their lack of certain knowledge in climate science -- in a scientific forum where they might be called on it if they utter something really speculative as if they are sure. A George Mason survey of actual climate scientists found that roughly one in seven think that there will be little to no warming and no catastrophe by 2100. Over half think that there will be significant, but probably not catastrophic warming. In the end, I agree with you -- this honest lack of consensus among climate scientists probably rates some consideration.
For one thing, it makes the entire field more credible. When was the last time you were in a room full of scientists who agreed about everything, even important things for which there is far better experimental data and far more computable theory
Or, all together were probably less than Tambora: http://en.wikipedia.org/wiki/Mount_Tambora -- at 800 Mt, considerably less. Tambora holds many records: Largest explosion in recorded history, loudest sound in recorded history, largest single-event influence on the climate in recorded history (it basically eliminated "summer" for two years in a row in at least some temperate latitudes) and helped make the decade of 1810 the coldest decade on historical record (but not the coldest year or part of the coldest half-century or century).
But I don't think we can be certain of the effect of the nuclear tests. Many of the largest were low over water and kicked a lot of water into the stratosphere. We just don't have the data, and hence any conclusions are likely to be guesses.
Hmm, I guess it's time to go pop the cap off of a tasty homemade beer, one that would easily cost $2 a bottle in a store or $5 a bottle in a restaurant. Not that I disagree that most people think it is too difficult to make and would rather buy cheap swill in large quantities of cans.
It is a legitimate field of study. One where negative results should have been (and eventually were) published. And also one where confirmation bias produced "surprising" results indeed -- until you looked for the man behind the curtain.