ALL server farms radiate into space - they just use everybody else's atmosphere (or perhaps a river and then the atmosphere) to convect the heat to the less-infrared-opaque upper stratosphere.

Absorbing the energy Out There, processing the energy into computation, and radiating it Out There, can be expensive and inefficient Especially if it is done with big hermetic boxes and circulating coolant. We do boxes and coolant on Earth because chips are sensitive to contaminated air and dirty fingers. We crowd everything into small packages because of propagation delay, and because we learned to design software on uniprocessors, and use those same habits on parallel tasks.

A less-bad space processor is an array of gossamer-thin, widely spaced compute surfaces. The simplest "thinsat" ( see http://server-sky.com/ThinsatV... ) is a thin surface with solar cells on the sunward side, and a grid of small "naked"chips on the back, radiating heat into 2.7 Kelvin deep space. A shape like a "T" is better; the top bar of the T is the solar cell, and the larger area perpendicular downstroke of the T is the surface with the compute fabric. The naked chips will be bathed in radiation - but the newest silicon production processes bathe chips in 100eV (13.5 nm) extreme UV photons during manufacturing. In higher energy space radiation, there will be bit flips and an occasional fried transistor - but 10 nanometer transistors are a very small target for space radiation.

So, design with error detection, correction, and redundancy. Design code for loosely coupled autonomous processes. Spread the grid of small chips over a area much larger than the solar cell, so that it can radiate heat at lower temperatures. Rule of thumb, chip "wearout" lifetime doubles for a 10C drop in temperature, so why not aim for a very large area and 250C operating temperature?

If the sun-facing solar cell array reflects infrared and UV without absorption, and is 50% efficient with the 40%-of-1366W/m midband light that it does collect, the solar cell array waste heat is 275 W/m, and a 90% infrared-emissive solar cell back side would be 270K. The angular size of the Sun near Earth's orbit is 32 minutes of arc, about 0.01 radians, so the triangular dark shadow umbra behind the solar cell is 50 times longer than the solar cell is wide.

In theory, a two-sided triangular chip-covered "pennant" in back could have a Stephan-Boltzmann law radiating area (both sides) 50 times larger than the solar cell, hence 40% of the temperature in Kelvins of that solar cell - perhaps 110K - but it would be cheaper to make the heatsink blade thinner, hotter near the chips and colder between, with a truncated blade. A properly designed solar-powered computer segment needs no plumbing or coolant if the computer chips are small, numerous, and distributed.

Also, naked and replaceable. Chip technology advances, compute markets change, upgradability extends system lifetime. The mass of the chips will be small compared to the mass of the solar cells, substrates, and interchip connections. Extend lifetime by replacing old naked chips with new ones.

Another emerging technology is fiber-optic chip interconnect. Optical fiber fed by wavelength division multiplexing/demultiplexing etalon chiplets can move Tbps data over a 125nm fiber optic core, with practically zero path loss over sub-kilometer distances. My 300/300 Mbps fiber internet link is all optical from my house to the switching center 20 km away; nothing to de-power or burn out. A 10-meter-scale compute surface could have tens of thousands of short optical fibers embedded in a heat-sink "sandwich" - a few grams per square meter.

Chip size? Perhaps 1000x1000x20 micrometers, 50 micrograms. More than a decade ago, I licensed circuit technology to Hitachi ULSI for their "smart dust" RFID chips, 300x300x60 micrometers. Using the "fluidic self assembly" techniques developed by Alien Technologies (cool name!) of California, we could have made RFID product tags cost less than a paper price sticker, and saved storefront retail. Then came the Sendai earthquake; Hitachi shifted corporate focus to recovery, and Amazon soon engulfed retail.

Orbital location? 1400 km altitude (7778 km equatorial radius) Low Earth Orbit is SILLY; the orbit is in shadow 30% of the time, 2000 seconds every 1.9 hours. Besides the downtime, thousands of huge thermal stress cycles per year could cripple a low-mass satellite.

A better place for thinsat arrays is in front of Earth-Sun L1, 1.5 million km sunwards; 10 second round trip signalling time. Four space probes are there now (SOHO, Wind, ACE, and Aditya-L1) in "Lissajous orbits" around the L1 gravity cusp. Continuous sunlight, no thermal shock, very small station-keeping thrust - which can be created by "solar sailing". There is room in that region to collect almost 100 Earth's of sunlight without reducing Earth's illumination more that 1%.

The 10 second delay is actually a bonus. I prefer that humans to have a wee bit of terrestrial reaction-time advantage compared to exawatt AI.

The internet is not some abstract infinite channel, where packets are created from nothing and move globally for free. It is a finite amount of expensive optical fiber and routing equipment with finite capacity. How is that to be used? Presumably to generate financial returns for those who paid for all that equipment. Where do the packets come from? Increasingly, from Hollywood, via services like Netflix.

Netflix video-on-demand is a horribly inefficient and intrusive way to use internet infrastructure. If you move gigabytes of high definition low latency video through a finite web, peer-to-peer, my web searches and information downloads are in competition. True net neutrality means you get as much bandwidth as I do, inbound and outbound, for the same price, which is a lot less bandwidth and lower Quality of Service than you are demanding now. At least Netflix understands that their video-on-demand model requires expensive physical equipment to implement; too bad their customer's brains have been turned into pudding by that video.

I used to offer unrestricted free wifi to my neighborhood. I liked the idea that poor schoolkids could work on their homework, travellers could consult online maps, visitors could check their email. Then the TV addicts discovered it, and filled my pipe with netflix packets. So I "violated net neutrality" and throttled the maximum bandwidth. The wifi can still be used for the beneficial uses I encourage, but video QOS sucks. As a bonus, I can worry less about a DMCA takedown. People can still move video, but they must wait a while.

Video addicts - please, please, please do boycott the internet. Feed your brain-rotting addiction with cable, DVDs, broadcast, satellite, and all the other low-cost ways that actually pay the producers and purveyors of your high-definition drivel. Your gigabyte road trains do not belong on the information highway until you pay for the extra resources to accomodate them.

  Freedom of speech means Libre, not Gratis. Pay for what you take.

I just learned about this when I went looking for a report, CR-2357, Peter Glaser et. al.'s "Feasibility study of a satellite solar power station" from 1974 . I've downloaded hundreds of NASA reports as part of my research, but not all of them. There are millions of pages of documents there; to review them to ITAR requirements would take everyone working at the agency years to do.

Sorry I'm late to the party, but I've been busy actually putting this government research to work, creating jobs and healing the environment. With this one heinous act, the politicians (R and D, it took both) just flushed a trillion dollars worth of scientific and technical research down the crapper, and possibly the survival of the U.S. You can bet that the (Name of Current Enemy Here) have already downloaded every one of those reports; all that the Feds have done is deny access to Americans. The really sad thing is that most Americans are such ignorant partisan illiterates that this modern "burning of the library of Alexandria" went unnoticed, perhaps was even applauded when the politicians reported their security triumph to the fools (R and D both) who elected them.

Any one of those reports, no matter how trivial, is more important to our future than the whole pack of politicians in DC. Let's ITAR house.gov, senate.gov, and whitehouse.gov instead, those maniacs are more dangerous than a nuclear arsenal.

CO(subscript 2). Slashdot editor kept the unicode subscript for previews, dropped them from the final. Fooey. Live and learn.

Funner fact: Where the effect happens is more important. The troposphere is close to IR opaque, with gas and black body temperatures closely coupled. The black body temperature of the earth, and hence the amount of IR radiation emitted into deep space, is the deep cold of the upper atmosphere. Clouds and sulfate particulates determine the amount of light reaching the surface (mostly ocean), where almost all is turned into heat. In the longer term, that heat is equal to the IR black body radiation, with whole system temperatures adjusting until they do.

Atmospheric temperature decreases about 6.5C per kilometer altitude, and density decreases by about a factor of two every 7 kilometers. Water vapor pressure drops rapidly with temperature, and the water freezes out around 0C forming clouds and precipitation. There is very little water vapor above the high cloud tops, and that region is ruled by CO and methane, which do not freeze out at atmospheric temperatures. At some even higher altitude, the remaining amount of CO is transparent to space. Double the CO, and the transparency altitude goes up around 7km, to a region 45C colder than 7km below it.

  If you could see the earth in the infrared, from space, it would appear colored by high altitude haze from the CO and methane, with lower altitude islands of cloud floating in an opaque sea of water vapor. The land and oceans are be invisible - and irrelevant to black body radiation. If you double the CO or methane, the long term effect is to increase the coloring, raise the water vapor "sea level" altitude somewhat. The temperature at this level stays close to constant (remember, that is the temperature where water freezes out), but at a higher altitude, there is more 6.5C/km air beneath. A very small change in atmospheric properties, raising the top of the troposphere a few percent, can easily result in an average 5C change down here on the surface.

I started out as a climate change sceptic, horrified by the pseudo-scientific and pseudo-technical nonsense spewed by the media. As a responsible technologist, I studied primary sources and ran the numbers myself, and changed my mind. Atmospheric scientists know a lot, have a lot more to learn, and cannot produce a definite prediction of exactly when our fiendishly complex, multi-billion cubic kilometer atmosphere will be broken beyond repair. The fate of civilization is dependent on the slow integration of effects that depend on other integrals of integrals; when the first pebbles of the landslide reach us, the unstoppable wall of rock will follow soon after, far too late to stabilize the slope.

Nature could soak up much of the excess CO, turning it into climax forest, plankton sea floor sediment, and perennial-plant root-mediated carbonate rock, but we are destroying the absorbers with annual-crop agriculture, especially atrocities such as "biofuel". We could reduce our methane use; instead we invest in erratic sources of "energy" such as wind turbines and grid-scale solar, which must be backed kilowatt-for-kilowatt with fast turn-on natural gas turbines instead of slow-response base-load hydro and thermal generation.

The natgas pipelines are overcapacity, leaking methane, bursting, even catching fire and exploding, compounding the problem of leaky production fields and leaky aging cities. If this is a problem in Europe and the US, imagine how bad it could get if impoverished India and China tried to copy our example.

No responsible technologist should blindly repeat numbers without checking them out, whether those numbers are ideologically comforting or not. Better to shut the hell up than to drown out the few responsible people who, whatever conclusions they come to, at least try to build those conclusions from direct observation, primary data, and replicable calculation.

Self-education and calculation is not only responsible, it is lucrative. Instead of the highly touted, vastly expensive "alternative energy" proposals touted by idealogues, there are many possibilities that are inexpensive, highly profitable, and a lot of fun to work on. I'm working long hours on a few, as are a few colleagues, and after stage 3 ("..."), we might even get paid for it. Help is welcome, competition is welcome, mindless slogans are not.

Apple doesn't shut down their data centers when the wind stops or when it is cloudy, and their biogas generators works only as fast as the microbes make methane, a process that is difficult to quickly throttle up and down. Perhaps they have vast underground tank farms storing methane for winter. Or perhaps this is all a fashion show to impress gullible green-wannabee customers.

So Apple is drawing the bulk of their power, most of the time, from the same grid the rest of us draw power from. They are building a new data center here in the Pacific Northwest, making the same claims about wind power. Wind power availability is random, on average less than 15% of nameplate capacity, and often entirely stopped for weeks at a time. On the rare occasions when wind farms produce much more than their tiny average, the Bonneville Power Administration is forced to buy it, regardless of the state of the power grid. BPA is being forced to spill water over the dams on the Columbia, which super-nitrogenates the water and kills salmon, thus violating their former "prime directive" - preserve fish runs. While maintaining navigation, controlling floods, providing irrigation, and about a dozen other important goals written into their charter. No wonder their director recently quit.

Solar - similar deal. The worst case solar insolation in winter is tiny, even in "sunny" New Mexico. No sunlight at night, of course, thin sunlight under week-long cloudy skies. Look at a December or January solar insolation map from NREL, and do the math yourself.

Where does the makeup energy come from? Given the short-term intermittency and quick-start requirements, mostly from natural gas turbines, increasingly fed through the national pipeline network from frack gas fields in the east and midwest. Gas magnate T. Boone Pickens /loves/ windmills.

Many of us here are technologists, capable of finding the numbers and doing the math. We do not have to rely on the pronouncements of corporate PR departments, government agencies, or conspiracist crackpot websites to figure out what is going on. We can go to primary sources, we can build our own spreadsheets, and we can replace magic thinking with nuts-and-bolts analysis to learn what works and what doesn't. The risk to the global climate is real (key phrases: IR column opacity versus altitude, adiabatic lapse rate), but the popular notions of how to deal with it are often worse than doing nothing.

And that is scary as hell.

Buy Apple products if you think the unregulated pollution from the power plants feeding Apple Chinese factories is better than the unregulated pollution due to their competitor's Chinese factories. If you don't, consider using free software on recycled/rebuilt/hot-rodded computers, like I am doing right now.

Every dollar a company charges above marginal cost is a big red flag to potential competitors - "money to be made here!" Australia has many crackerjack programmers. Perhaps as a write this, some small Australian startup is selling a product or service that is superior to Adobe products in some specialized market niche. From there, the upstart could grow too big to dislodge, while Adobe's attention is focused elsewhere. From that secure position, the upstart could grow to dominate the world market. It's happened plenty of times before.

The base platform for this hypothetical startup could copy code from Gimp and Inkscape, which are not identical to Adobe products and hence (to the mentally inflexible single-trick end-user) "sucks a bag of aids dicks". Doesn't matter. You can't copy code from Adobe products, so only Adobe (a tiny fraction of the world's programmers) gets to build new products with it. The rest of the world gets to build on open source tools, forking and remixing and sometimes stumbling across something that redefines the game.

The real victims of this will be inflexible Adobe customers, who are trapped deeper and deeper into one way of doing things with every project, and who must add an "Adobe tax" onto every project to pay for their expensive tools. When game change becomes unavoidable, they will struggle to find time to become adept with new tools, while still maintaining competence with (and paying for) the Adobe tools to support legacy projects. Most will fail, becoming obsolete like all the competent-but-struggling musicians who depended on the old record company label system for occasional gigs, and are now unemployable in the age of online independent artists and globe-spanning internet collaboration.

Because of the profit-rich environment that Adobe and its distributors have fostered in Australia, those new tools and new collaborations may originate there, fostered by the spirit of an intelligent and fiercely independent nation. Adobe, with their pricing, have made their own demise inevitable, taking far too many of their loyal customers with them.

I run a modded Thinkpad T60, and have stockpiled two more, and extra screens and keyboards. A lifetime supply, if necessary, though batteries will be a problem someday. I want 4x3. I want trackpoint. I want Linux compatibility. And I break warranties - look inside, repair inside, change inside.

The T60 is my main work tool. I calculate. I design. I program. I write papers and books. I can display two 8.5x11 pages on a 4x3 screen and fill it nicely. I can drive a 4x3 computer projector (still the most common) without clipping. 16x10 and 16x9 may present movies and games better, but I don't make movies or games.

Donald Knuth borrowed the T60 I'm typing on now. He's not going to borrow your windows 8 media vending machine.

I modded the T60 to a 15 inch 2048x1536 screen, from an NEC prototype run that M$ did not support. I built a kludge to rewrite the EDID eprom so the Thinkpad BIOS accepts it.

I bought my first Thinkpad (a 560) when my Dell laptop spent TWO MONTHS waiting for a power supply repair. That first Thinkpad developed a similar problem, and IBM fixed it 36 hours door-to-door. In fact, the morning after I sent it, I got a call from a service tech at IBM, who said "We see you are running Redhat 5. We have a new BIOS, which we've tested with Redhat 5. Before we ship your machine out in a few minutes, will you permit us to upgrade the BIOS?" Oh. My. God. SOLD.

That kind of service (no longer offered by Lenovo) is worth a huge premium. Laptops pack a lot of technology and heat into a small space, an unavoidable reliability compromise. Excellent manufacture and excellent service turn a consumable into a capital investment. Sadly, Lenovo is now following the race to the bottom. Thank goodness it is not leading it. Yet.

I am 59 years old. I've fooled with computers since I was 15, and built two from scratch. I'm old enough to NEED a trackpoint (hands too jittery to use a touch-touch-touch-touchpad). My visual acuity is dropping - handheld screen text is too small, and I must have a matte screen.

And there are millions like me. Most boomers are self-absorbed gimme-pigs, but some of us still design the hardware that the world depends on. The tools we need are disappearing. The tools that genX and genY engineer/entrepreneurs will need when they pass 40 won't be there. I weep for them.

Perhaps someday, some manufacturer will wise up, make a run of 15 inch 4x3 display glass, put a decent computer and keyboard underneath it, and own the engineer/business/senior market segment. They won't sell 100 million units, but they will get ten times the profit. It would be great if this was a branch of Lenovo, but my loyalty is to the usefulness and quality of the product, not to the nameplate.

Change? I'm all for change, if it is improvement, not atrophy.

For the last four years, I (and a tiny part-time team of volunteers) have been working on a different way to think about satellites:

http://server-sky.com/

These will be 3 gram, dinner-plate-diameter 50 micron thick satellites, based on some recent advances in semiconductor technology and Ivan Bekey's "Advanced Space System Concepts and Technologies: 2010-2030+". We are doing most of this as open technology, and I make presentations to groups that might help. Monday at NIST Gaithersburg, for example.

Satellites are surfaces that combine stimuli and solar power to make microwave transmissions to earth. A one micron thick layer of graded junction indium phosphide makes a 20% efficient solar cell and is very rad-hard. Ditto for Penryn-process silicon. That recent work, plus a number of other fortuitous recent discoveries, plus the kind of manufacturing techniques we use to make LCD displays, plus solar-sail-like maneuvering means we can built extremely thin satellites. They need to be heavier than 100g/m^2 to stabilize their orbits against light pressure, but they still can beat current satellite watt-to-kilogram ratios by more than a factor of 100.

The "satellites" will be arrays of thousands of these thinsats - a 99 kg array will contain 33 thousand of them, and collect about 160 kilowatts. That fits the sub-100kg definition of "micro satellite", though I hate the abuse of the scaling unit.

This is all very speculative, of course, and the thinsats that get launched will be very different from our early designs.

See the server sky website for more. Pretty chaotic right now. I use it as a public notebook, slowly improving the content as we learn new things, attempting to establish all this as public domain and attract informed discussion. While most readers will be highly skeptical until experiments get launched, non-analytical skepticism merely keeps competitors befuddled, giving us a head start. Folks with imagination, who can do research and the math and physics, have the opinions we care about.

For "conventional" cubesats, there are many universities working with the US Naval Academy. They have a "2U" cubesat design with a slightly-smaller-than-1U-sized plug-in compartment, which provides power and communications to a plug-in experiment provided by a university research partner. So, the university partner provides the experimental plugin, and a Really Good Story to convince the USNA that the experiment is worthwhile. USNA launches the cubesat, and the middies manage it, to get experience managing satellite assets without risking big military birds. At least, that is what I remember from attending the Amsat symposium a couple of years ago to give a paper on something Completely Different.

My name is on 12 patents. Mostly electronics, though one can be construed as a software patent. I would rather not have my name on most of them, and many result from the same situation you are in.

With my first employer, this was in my original contract years before I gave any thought to it. When I did, I said "no patents, or I quit". I did valuable work, so the quid pro quo extended another 5 years (the sneaky patent attorneys did get a European patent on some of my work, and never told me or my coinventors, but ...). When I finally left, they took out four US patents on the stuff in my notebooks, and threatened legal action if I didn't go along.

Then I helped a startup. They took out three US patents on my work for them. This time, to refuse would have torpedoed the company. In time, the foolish V.C.s torpedoed them anyway (high growth, high profit, market dominance, but the product was hard to explain at IPO - so the VCs made them work on easy-to-explain but low value products).

Then I consulted. I started putting "no patents" into my contracts. With one startup, that resulted in a four month delay in contract approval, resulting in a delay in further work. Which meant the startup was missing essential skills, the prototype failed, and so did the startup.

I learned to make my ideas look like someone else's idea at the clients. There were still some patents filed, but at least my name wasn't on them.

Now I have my own company. With four patents. And three on the way. I hope to release four of them into some kind of "public patents commons" arrangement, if I can find an organization that can leverage them properly into more public patents

I would rather that all patents go away, even if the global abolishment means the three I expect to continue to draw income from go away also. Sadly, in the world I live in, the alternative structures just aren't well developed, so if I don't patent an idea, a competitor will, locking me out of using my own idea. Or a client will treat the work and customization I provide them with as less valuable, because they notice "IP" more than they notice "good product". That is a cultural problem which I cannot fix myself. I hope to help with the fix, though, and I am always looking for alternatives.

I have managed to get a dozen or so potentially patentable ideas released directly into the public domain, mostly through publication in professional journals, and professional standards work. This is the best route available, given the existing situation. So:

Publish. Teach. Participate in standards bodies. Share all the ideas publically that your job permits. Write open source code. Use open source code, and attempt to make it the way things are done where you work. At the end of the day, those who want to contribute ideas to the public should just do it. Frequently. If that is not your job within the organization, morph your job until it is. If you cannot morph your job that way, find a new one elsewhere.

In the golden age of U.S. science, many great ideas were publically shared by scientists and engineers working for private companies. Many Nobel prizes resulted. Companies shared this work because it helped them hire the best, and showed their customers that they did first rate science, leading to first rate products. That age seems to be ending, but I think that is because the researchers value their salaries more than their scientific reputation. Now open source software seems to be heading into that golden age, and it will only stay that way if software innovators value their reputations more than the size of their paycheck. I respect the hell out of these people, and help out where I can. But it is a tough choice, it means forgoing some lucrative jobs, and if you have a family to feed, there is no easy answer. You have to decide whether you want to eat well, or sleep well.