I will never live to see the day when a tech product borrows more than the Macintosh expropriated from Xerox. The greatest designer in the history of Apple is Douglas Engelbart.
I don't see the value in allowing Xerox to lock all these ideas up to see whether they could contrive a viable business model in ten to twenty years when Jobs managed to pull it off in three.
Apple didn't invent the touch screen. They didn't invent minimalism. They didn't invent polish. Good for Apple to show up first to market with the slickest and most polished device.
There's a fairly large component of mob psychology in the computer business. Trends catch on. Whatever trend caught on is viewed as being simpler than whatever trend didn't, but half the time that's just acculturation. Simplicity is when the industry moves as a whole. Cars all have the same steering wheel because it's a good idea for people to sit behind controls they recognize.
I don't believe in Jobs' IP model behind his presumed ourPhone. I didn't believe it when he was stealing from Xerox, and I don't believe it when Google is stealing from him.
Steve is beginning to remind me more and more of Howard Hughes. If he had Ellison's third testicle and the babes to match, the comparison would be a dead ringer.
If I find flaws in a proof, does that make me a math skeptic? A tolerable statement would have been:
Smoking gun skeptic Stephen McIntyre...
Stephen is not convinced that steam from the morning sunrise should be included in the assessment of smokingness. He's be less of a skeptic if more of the vapour was actually smoke.
So your idea of the best climate scientists is... people who aren't climate scientists?
Michael Mann defended weaknesses in his statistical methods on the basis that this paper survived peer review, despite the peer review failing to include a statistician with expertise on the statistical methods employed. How does that work in any other walk of life? What gives science the passing lane to miracle quorum?
Working in another field, Stephen McIntyre does have expertise on the application of statistical methods to inflated conclusions and he elucidated flaws in the approach to the tree ring analysis which notable statisticians have commended as very astute.
Mann responded by playing a game of "you can't have my data", so it was a long time before notable statisticians had anything to pronounce upon.
Mann is representative of the climate believers who feel it's more important to be right than to get to the right answer on the best possible foundation. In part this is a defense against well funded detractors who wish to distract the climate debate to go around in endless circles of mock debate. I understand the frustration.
The problem with Mann's approach to McIntyre is that McIntyre had actually filed a valid bug report. All Mann needed to do was fix the bug, publish a supplement to his paper with less convincing hockey sticks, then go back to the grindstone to find data or an analysis of the data the proved what we all suspect on a foundation of watertight analysis. What any scientist working in dull obscurity would accept as everyday life.
Mann behaved like a project manager who had a progress graph on his wall showing 80% complete after a developer comes to him and says "we've made a huge mistake in estimating the scope of one of the subsystems, so the remaining work is twice what we thought". A good manager updates his chart to show 60% complete, then works his ass off to follow through on the 40% that remains. A bad manager says, "but we had a board meeting and everyone signed off on 80%" Then the developer gets painted as a progress denier.
I am absolutely thrilled to see this analysis being repeated by a group of people I suspect would rather fall over dead than mutter some of the vague defenses employed by Michael Mann. I think Mann is a fairly decent guy who did a good piece of work on a very difficult subject, made a few extremely subtle mistakes, then reacted very badly when those mistakes were identified, primarily by saying things about science that no-one trained to speak about science would be caught dead uttering.
The thing about peer review is that it catches more problems than it misses most of the time. It's not rock solid in any particular instance. In the fullness of time, the process converges to good science. But the whole point of the climate debate is to incite a radical economic response far in advance of the fullness of time that makes science a faultless enterprise.
There's another group that wishes to claim that the radical economic response isn't actually that radical. People trained to study this are called economists, not scientists. I know, that's a horrifying reality. I've yet to meet a scientist with a fourth year credit in global intervention, yet there are no shortage of these guys telling the world what it needs to be doing.
Some of them are speaking with wisdom and common sense. Are they trained to take these positions? Absolutely not.
For example, there would be millions of man hours fewer required for people to 'make change' all day.
Someone needs to buy you a stopwatch. Any decent clerk can make change faster than the till prints the receipt.
Here's a much more direct way to reduce petty counting and improve interpersonal sanitation: eliminate recycling of pop cans.
But if we're thinking large why don't we just eliminate the pop cans altogether? Reductions in obesity, global warming, wasteful petty commerce, what's not to like?
Over all, I think they've handling this well. I just wish they would display the number of views remaining where it greats me every day with the phrase "Welcome, ohsupremeleader".
Actually, my NYT user name is from the Slartibartfast school of almost filthy. Refer to someone that way in the joint, they'd carry you out in a box. It combines the adolescent directness of fartface with the oozitude of sloppy seconds. I guess I was none too thrilled about the mandatory sign-up.
But honestly, they have pretty good content, and deserve to stay in business.
Not the departments I would choose necessarily but this is the type of thinking I am on board with.
Honestly, I laughed out loud at your use of the word "thinking". I detect no thinking here whatsoever.
Thinking would involve a detailed cost/benefit analysis written as a terms of reference for a viable transition plan. It wouldn't sound like a drunk at the Politburo playing daisy with the purse strings of power, where terms of reference are never more than a slam of a shoe.
If he's such a wondrous business man, why have Microsoft's shares flatlined?
To protect an existing profit level higher than nearly any technology company that has ever existed. Corporations, like technologies, have natural life cycles. Plus, Microsoft is discouraged by law from exploiting it's entrenched monopoly in many ways it might have liked to.
Ballmer sucks in any event, without ignoring the first week of economics 101.
Dyslexic much? Don't think you meant 1895, though it's more plausible than what you wrote. 1958 is the only vaguely plausible coordinate. The people who talked up these techno-fantasies then were the same people on the waiting list for LSD.
In 1984 I was handed some bizarre Tokamak fusion propaganda by some kook in the SF airport which seemed like it came from a different planet. Many voices in the GOP now sound like they hail from the same planet. Useful as a premonition of mood. The fusion predictions, however, were not so good. By 1985 the only sane people who talked this way were people who had collections of Playboy, Popular Psychology, Popular Science, and Omni collecting dust under their beds. Sane, but a little behind.
The smart people could see the internet looming on the horizon. Had many talks about this. What we mostly failed to understand was the eventual economic model (which surprised nearly everyone) and that you wouldn't have to say "mother may I" to a utility company with every click. IBM was regarded as the epitome of evil control and there was much concern about escaping their dark Monopolis.
It wasn't until the dominance of wireless service that the world returned to the Mother-may-I format we might have predicted. Contrary to your assertion, none of us were lulled into regarding the Jetsons as a documentary.
The only future technologies I ever talked about was infomatics, genetics (anticipating the human genome project), radio astronomy, and a hint of nanotech. For the most part, I thought the Space Shuttle was a waste of money. The Hubble was a nice consolation prize, I will admit.
Moore's law is a lot like a paving machine. We would have given anything to know how much more asphalt was in the hopper. More than most of us suspected. How the brain struggles to comprehend 64GB (8*8 on the upcoming SandyBridge-E) when you're trapped inside 64kB without even a hard disk drive.
Strangely, fibre optics were closest to revealing their true potential.
The optical fiber amplifier was invented by H. J. Shaw and Michel Digonnet at Stanford University, California, in the early 1980s.
Erbium-doped was first demonstrated in 1789. Er, I mean 1987. I remember reading about it for the first time in Scientific American probably a year before then. I nearly jumped out of my seat at the immediate implications. Without the use of a futuristic jet pack.
Yeah, I don't read all that much SciFi and the chart quickly lead me to a whole bunch of softcore SF classics, all of which I've read.
Here's a question: How does it make a SF title any better to have been written in the last hundred million seconds out of 100,000 years? Isn't keeping up with the present the domain of the Twitterverse?
I'm a well-aged consumer of scotch, cheese, movies, and books. If I'm going to consume something fresh, it's probably a documentary that took ten years to finance and film.
Every public company is required by law to become the next Microsoft if the business opportunity presents itself in order to provide maximum return to shareholders.
But then, you can take it to a whole new level by submitting falsified video tapes to the DOJ.
The government produced its own videotape of the same process, revealing that Microsoft's videotape had conveniently removed a long and complex part of the procedure and that the Netscape icon was not placed on the desktop, requiring a user to search for it. Brad Chase, a Microsoft vice president, verified the government's tape and conceded that Microsoft's own tape was falsified.
So yes, Google is getting grubbier, but it has yet to descend to flinging feces around like chocolate bon bons.
Mod parent up. In ten years I've never said that. Old dog, new trick. Also, was the mistake thinking we needed to do this too quickly?
But you never have the necessary experience to design something right the first time, so that is not really a worthwhile discussion.
Absolutely, there are many questions here unworthy of extended discussion. One doesn't to consult Donald Knuth to master quicksort. There's a huge literature on overdesign, discounting through net-present-value, adoption risk, and the difficult promulgation of technical standards.
TCP/IP has had roughly the same amazing thirty year run as x86. I've followed the twists and turns with x86 far more than IP, so I have a pretty good idea what I'd write on a napkin to send back to the original x86 design team--supposing it was actually a going concern to make the design not suck and that suckitude was not actually the criteria by which IBM originally chose this part, leading it's relevance thirty years later. (On the former question, I think Intel wanted to make a good part; on the later question, I think IBM was seriously gun shy about the future arriving quick and capable. In Terminator mythology, the napkin to Intel would fail, and be followed by a cyanide capsule to quavering bean counters over at Big Blue.)
First suggestion for Intel: offset the segments by 8 bits instead of 4 bits. This gets you a 16MB address space which gets you to viable 32-bit multitasking without a foray through DOS extenders. Let me put this iron into the fire to demonstrate, in a few minutes, the pleasure it will bring to thousands of programmers if you don't follow this advice. I think there would have been very little practical cost to this--even if I don't tell them that segmented data protection would fail. That feeble 20-bit address space was penny wise and pound foolish any way you look at it.
Second suggestion: either set all the status bits, or none at all with extremely few exceptions. In the short term this simplifies compilers, in the long run it expedites data flow implementations (i.e. OOO scheduling).
Third: Variable instructions length is a mixed blessing. Good if the first byte determines instruction length. Have at most one byte in the instruction which encodes allowed combinations of what might otherwise become prefixes/overrides. Variable instruction length improves icache density; a 30% efficiency gain over 32K of icache more than pays for 500 extra transistors in the instruction decode logic. Think about slope and intercept on instruction encoding density. Ideally your first byte tells you if you have a override bitfield and which byte contains it.
Fourth: Have more index registers than just SI and DI. This one is tricky since any expansion of the register set slows task switching. Maybe it should wait and become "have more index registers than ESI and EDI". But I think not. 16 registers is plenty in a RMW architecture. You don't need 32 registers unless blinded by RISC purity.
Fifth: Plan to transition the x87 co-processor from stack-based to register addressed at the earliest possible opportunity.
Sixth: Add an instruction to bit reverse bytes. Add an instruction to popcnt bytes. Chicken and egg problem. If you build it, they will come. Always have an instruction to byte reverse your longest supported register.
All the other x86 complaints: ignore completely. It's the future trying to eat your lunch after depreciating your opportunity cost and adoption risk to zero.
If you recall clearly what the world looked like when ideas are first on the drawing board (x86 is my personal case study for thinking about this), it's a lot harder to send advice back through the time machine that isn't actively dangerous.
How many of my suggestions would have rewarded them right out of the gate? Not a one, unless they put a full court press on high quality C compiler, to exploit the suggested register
Unless the phone or other device can also take screenshots, or doesn't have that software installed... or... or... or...
Santa just called. He wants his elf-master of the list back before this junket of free association into the infinite void permanently curls his toes into cranky hang nails. Santa's old experiment with CRM114 automation did not go well. Return the elf, now!
Mob psychology is the echo chamber of common sense. And that's the good outcome. Even worse is nature/nurture where there was never much common sense to begin with.
Yes, there are correlates on both sides despite one or more mixing rounds of bent functions.
How does one perform medical epidemiology on an encryption block your kid sister wrote? Let's say your kid sister is Judit Polgar and she's almost smart enough to get this right (having not actually majored in math or computer science), but then you truncate to two rounds, so it's more like scrambling eggs with a spatula rather than an egg beater. This is a good model for the human nature/nurture system.
Permute the keys bits holding the block bits constant. Permute the block bits holding the key bits constant. Throw this into a powerful statistics whizinart, then press "publish" to gasps and wows from the bleachers of humanity who are slow to grasp that their common sense on this matter is six feet under.
Yeah, you can probably partition into dozens of sub-regions of statistically significant linearity manipulating the input bits on either side. There's many discernible chunks of white and yolk in the spastic scramble.
Even if you take major features of culture (such as our universal 12 year educational system, which represents about a millisecond of our 80ka recent history) and correlate genes most amenable to this, there's a wide span of orbital radius as experienced by any particular member of the population.
If a trillion dollars worth of epidemiology tells you less about a person on a quick reading of their gene chart than you get from an astute five minute introductory conversation, what exactly has all this research accomplished?
What we will find eventually are a few genes or gene complexes which correlate strongly with the effectiveness/ineffectiveness of proposed interventions (such as reading assistance). The exact fractal coefficient on little peaks of signal exploitation is certainly a quantity of interest moving forward. I'm pretty sure it's bounded from above by log (panacea).
We'll all become a tiny bit better at playing to strength. And most of the rest of the signal will fall below the noise floor of messy human affairs.
Here's the surprise twist: astute assessment is thin on the ground (the sap to syrup problem) and nearly impossible to institutionalize. So unless our machines become astute ASAP (as some predict), we'll probably press forward with the institutional seal club of genomic aphorism.
My comment in that thread was that 90% of the workforce will be unemployed within half a century on present trends. We're heading into a post-commodity winner-take-all employment climate. You'd like to think present trends will change. But how?
I once read a lot about the publishing industry. There were a lot of people in publishing who really cared about books. They faced an increasingly faced a harsh reality, until only the harsh survived.
For every Rowling making a billion dollars (she says less) there are thousands of small time authors crowded out of the marketplace.
Does disintermediation contribute to a long tail wagging happy to see you? Or do we just end up concentrating more wealth among the winner-take-all block busters, while the back-catalog brims with also-rans and failed contenders that never cover the author's outlay?
I don't understand how this reviewer has never heard of Chaitin, but finds this book vastly too elementary. Oddly, I mentioned Chaitin in an earlier post this very day. Perhaps reviewer should tear a page out of the Roger Ebert school of criticism:
When you ask a friend if Hellboy is any good, you're not asking if it's any good compared to Mystic River, you're asking if it's any good compared to The Punisher. And my answer would be, on a scale of one to four, if Superman is four, then Hellboy is three and The Punisher is two. In the same way, if American Beauty gets four stars, then The United States of Leland clocks in at about two.
That's a lot of words to pour out without defining expectations or genre. And there are many sub-genres within science writing.
Rubbish, if you expand your horizons wider than oracles of ticker tape or Back to the Future parlour tricks.
There have been some pretty profound visionaries over the centuries. Jules Verne, da Vinci, Richard Feynman (There's Plenty of Room at the Bottom), Claude Shannon, Freeman Dyson (space chickens), Charles Babbage, Leibniz, William Gibson (cyberspace), Marshall McLuhan (global village), Archimedes if you could get him to talk. These are not men immortalized for aping Minority Report.
I shake my head at all these Margulis extropians, who think we're headed for post-sexual merger with mechanoid symbiotes, the under-skin super suit. Which would be cool if I had any clue what the 90% of world's population, the unemployed, will be doing with all that time.
The future is a moving target. Set your sights accordingly, and recognize transcendent wisdom bereft of gadgets.
I'm pretty sure all those salmon are swimming upstream to discover the cure for senescence ending the nasty business of spawning for ever and ever.
This is the least conservative line of investigation in the history of the human species. Stem cells? Totally passe. Surf upon mighty Morforgeddon, and despair!
But it will lower the abortion rate, so let's not be hasty to call it a bad thing. An aging base could potentially be a godsend once we repeal term limits.
Every true geek must read Metamath! or at least vaguely grok the concept.
It's the finishing touch on a brilliant little geek high-ball made from Heisenberg, Turing, Godel, Kolmogorov, Chaitin with a Mandelbrot cherry. Toss it back then bite the lemon! Warning: There's a lot of peel in A New Kind of Science. Chaitin is short and dense and accessible to a thinking 15 year old, for some value of "ignores assigned homework". Wolfram is pointing out that the mathematical Chaitinverse is a five minute walk from the financial district; a few steps past the edge of city limits you spot your first cactus.
In fact, as far as I can see it, the set of transcendental numbers being infinite, it seems that there must therefore be at least one transcendental number that encodes any given message within any given number of digits of its beginning in any given base.
There exists a message which this little strip of paper is too small to contain. Bursts into flames.
The significance of trillions of digits of Pi is akin to the Apollo program. An alien civilization who spies a little shard of metal emitted from the earth's atmosphere, winging very directly to the moon where it loiters for a few orbits, then winging back to earth (in a vacuum!) would probably be thinking "Petunias! Whale meat!" Life, in other words.
Catching a glimpse of a million digits of Pi from the billionth page on the Hollywood hills of a distant galaxy, they would think exactly the same thing--after Zaphod Spader conducts the kindergarten refresher class on algorithmic complexity theory.
Generally, the way that scientists design experiments is to vary one factor at a time while keeping the other factors constant, but, in many cases, the most effective way to test a biological system may be to tweak a large number of different factors at the same time and see what happens. ABE will let us do that.
The rest is all mystery meat about the actual method.
This is certainly the way of the future. IBM has pumping that notion right now, as well. I'm not inclined to think this is a huge step if they've only found signal under Gaussian noise.
The rule of thumb is for the top of your screen to be at eye height sitting erect. Any higher causes your lids to lift (so I once read) and can lead to dry eyes.
I personally had back problems associated with using the mouse with my right hand on the other side of the numeric keypad. Changed to my left hand, which is less used for other things. This moved the mouse closer to my body line and reduces stress on the upper shoulders. If I really need precision, it's easy enough to reach across with my right hand for a few minutes. My left hand became automatic for normal cursor work a long, long time ago. My right hand still feels a bit more precise.
I just did a comparison drag selecting words with right and left hands. The right hand feels more interactive, but it was uncertain of the mouse velocity in the select motion and actually took longer. My left hand much more accurately sweeps the word length. Consciously, I don't sense this, but objectively there it is.
The colour temperature of your monitors can affect stamina and mood. This varies a lot from one person to another. Recently scientists are saying that a high colour temperature boost productivity. And office temperature matters, too.
I once had a set up where I was experiencing eye strain. When I illuminated the desk in front of the displays with a warm yellow desk lamp, the eye strain went away. Uniformity of illumination also matters. Glare from behind the monitors is a killer.
A seat pan with tilt is a nice adjustment. I don't see the need for stupidly expensive chairs. I've had a couple. You really have to try different chairs and find one that works. If you're big like me (6'4" 240) you'll need to supplement the cushion on most medium grade chairs. I'm not sitting in a cheap chair, but without the extra cushion I soon feel the nuts and bolts. Pay special attention to the lip at the front of the chair where it cuts into the back of the leg.
Pay attention to pain and try to do something about it. It's very easy to negatively condition yourself to associate work with pain. Then you have a psychological barrier to fight with, too.
I got myself a full-sized Thumper almost ten years ago. Never had a problem with it. It's nothing close to a professional massage, but it provides welcome relief on those long, hard days. My chiropractor has the newer Maxi Pro and it's certainly a step up. I think it's a complete joke that it recommends no additional pressure beyond it's seven pound size. Yeah, that works fine for my quads. When it hits the tightness in my low back, it just sits there and bounces without some manual force. The tightness in my low back is epic sometimes. For a while I was doing a back exercise program at a clinic with special back equipment. On some machines I started with just a few plates like the wimp I am. On the back extension machine, it was the whole stack from day one. How many of these should I do with my cords of steel? I was running out of wind before feeling the burn. Everyone has their own thing. One of the best investments I've made.
The best upgrade to a Thumper is a patient and willing GF (with sharp elbows). These are best acquired with skill in the kitchen (with the contents of the refrigerator! cooking food served on plates!!) Invest in a good set of knives, heavy bottom pans, and a traditional iron wok. Fill you spice cupboard. Experiment. It's 90% water and heat. Trade your Aeron chair for Modernist Cuisine. I haven't gone quite that far myself, but my cooking skills are pretty good already.
Determine your caffeine tolerance level. More than one eight ounce cup every four hours is rarely optimal. If you get a headache when you miss your morning coffee, you're
Any narrow advantage held by the Power architecture was quickly disappearing by 2003 when the first Operton chips with AMD64/SSE2 hit the market (for users able to jump to 64-bits), and pretty much obliterated with the introduction of the Core Duo in January 2006.
x86 had an ugly childhood, but it turns out there wasn't anything desperately wrong with its performance potential. Jobs made such a big deal of x86 being somehow deeply inferior. The x86 is deeply inferior when you try to build efficient devices under a watt. The little tricks used in x86 to achieve high performance are expensive in power consumption and nearly impossible to fix without an instruction set overhaul. ARM was the true victor.
You really have the edict entirely backward. It was Jobs' edict that Apple would slug it out on a platform with a small market share, and for which IBM could not afford to develop to market-leading performance on an indefinite basis. Finally Jobs lifted the anti-Intel edict because he had no choice. Not just in performance, but also available production quantity.
When you get right down to it, I'm sure Jobs regarded PowerPC as a convenient walled garden. He didn't want his miracle machine to become too compatible. But it was probably also a huge burden to carry your own ISA for 10% market share. Look how they've done since.
The decade of PowerPC pretty much corresponds with the Apple doldrums. Imagine that.
I'm curious whether his reality distortion field penetrates from the after life, or whether Gandalf the White will arrive to perform an exorcism at long last.
When I was younger, I used "because" and "since" in my writing about twice as often. Never terribly pleased by the effect--that's just how it came out. They are fairly weak transitions, useful mostly if you want a weak transition which detracts less from a central element.
This excess tapered off as I became more deeply immersed in my subject matter with age and experience. In my own history, these words were sign posts of incomplete thinking.
Re:Corporations aren't evil. They're not anything.
on
IRS Auditing Google
·
· Score: 1
Corporations shouldn't be taxed, period.
That's one way to do it. I might even agree if we deleveraged corporate ownership.
If A and B own 51% and 49% respectively of C, and C owns 51% of D, then when there's a vote in company D, B gets.49*.51 percent of the vote. But that's not how it presently works.
Do you even see the connection? Wallets don't own and control other wallets.
While I appreciate your desire to be geeky, may I point out that "table salt" is 33% more efficient at conveying the intended information?
Tell that to my Himalayan salt or my Mediterranean sea salt, both of which imply unspecified exotic trace elements. I'm not quite so willing to energize on white bread custom as to equate the two. For brevity, I keep a shaker of NaCl.
for all I know, they've fixed whatever was so horribly broken in their bureaucracy that prevent them from ever including the choices I wanted, but I wouldn't know as I haven't felt any great desire to go back...
Amen, brother.
3.0 woody on 19 July 2002 3.1 sarge on 6 June 2005
Back in the day, that was a long hiatus in internet dog years. And what I depended on especially at the time was an up-to-date LAMP stack. I read a lot on the Debian list at the time and I can't say I came away impressed with their sense of priority during that time.
Now Ubuntu has crumbled beneath my feet. I'm really tempted to jump to FreeBSD. But then I have a project coming up to play with OpenCL and I'm not sure FreeBSD will facilitate getting started. Just love it when my bug reports have a unique in the galaxy hardware configuration. People jump right in to help you out.
Slashdot Mirror
I will never live to see the day when a tech product borrows more than the Macintosh expropriated from Xerox. The greatest designer in the history of Apple is Douglas Engelbart.
I don't see the value in allowing Xerox to lock all these ideas up to see whether they could contrive a viable business model in ten to twenty years when Jobs managed to pull it off in three.
Apple didn't invent the touch screen. They didn't invent minimalism. They didn't invent polish. Good for Apple to show up first to market with the slickest and most polished device.
There's a fairly large component of mob psychology in the computer business. Trends catch on. Whatever trend caught on is viewed as being simpler than whatever trend didn't, but half the time that's just acculturation. Simplicity is when the industry moves as a whole. Cars all have the same steering wheel because it's a good idea for people to sit behind controls they recognize.
I don't believe in Jobs' IP model behind his presumed ourPhone. I didn't believe it when he was stealing from Xerox, and I don't believe it when Google is stealing from him.
Steve is beginning to remind me more and more of Howard Hughes. If he had Ellison's third testicle and the babes to match, the comparison would be a dead ringer.
If I find flaws in a proof, does that make me a math skeptic? A tolerable statement would have been:
Stephen is not convinced that steam from the morning sunrise should be included in the assessment of smokingness. He's be less of a skeptic if more of the vapour was actually smoke.
Michael Mann defended weaknesses in his statistical methods on the basis that this paper survived peer review, despite the peer review failing to include a statistician with expertise on the statistical methods employed. How does that work in any other walk of life? What gives science the passing lane to miracle quorum?
Working in another field, Stephen McIntyre does have expertise on the application of statistical methods to inflated conclusions and he elucidated flaws in the approach to the tree ring analysis which notable statisticians have commended as very astute.
Mann responded by playing a game of "you can't have my data", so it was a long time before notable statisticians had anything to pronounce upon.
Mann is representative of the climate believers who feel it's more important to be right than to get to the right answer on the best possible foundation. In part this is a defense against well funded detractors who wish to distract the climate debate to go around in endless circles of mock debate. I understand the frustration.
The problem with Mann's approach to McIntyre is that McIntyre had actually filed a valid bug report. All Mann needed to do was fix the bug, publish a supplement to his paper with less convincing hockey sticks, then go back to the grindstone to find data or an analysis of the data the proved what we all suspect on a foundation of watertight analysis. What any scientist working in dull obscurity would accept as everyday life.
Mann behaved like a project manager who had a progress graph on his wall showing 80% complete after a developer comes to him and says "we've made a huge mistake in estimating the scope of one of the subsystems, so the remaining work is twice what we thought". A good manager updates his chart to show 60% complete, then works his ass off to follow through on the 40% that remains. A bad manager says, "but we had a board meeting and everyone signed off on 80%" Then the developer gets painted as a progress denier.
I am absolutely thrilled to see this analysis being repeated by a group of people I suspect would rather fall over dead than mutter some of the vague defenses employed by Michael Mann. I think Mann is a fairly decent guy who did a good piece of work on a very difficult subject, made a few extremely subtle mistakes, then reacted very badly when those mistakes were identified, primarily by saying things about science that no-one trained to speak about science would be caught dead uttering.
The thing about peer review is that it catches more problems than it misses most of the time. It's not rock solid in any particular instance. In the fullness of time, the process converges to good science. But the whole point of the climate debate is to incite a radical economic response far in advance of the fullness of time that makes science a faultless enterprise.
There's another group that wishes to claim that the radical economic response isn't actually that radical. People trained to study this are called economists, not scientists. I know, that's a horrifying reality. I've yet to meet a scientist with a fourth year credit in global intervention, yet there are no shortage of these guys telling the world what it needs to be doing.
Some of them are speaking with wisdom and common sense. Are they trained to take these positions? Absolutely not.
Who reads your x-rays,
Yes, IIRC, the Google report indicated that SMART murmured not a peep in about half of all failures.
Someone needs to buy you a stopwatch. Any decent clerk can make change faster than the till prints the receipt.
Here's a much more direct way to reduce petty counting and improve interpersonal sanitation: eliminate recycling of pop cans.
But if we're thinking large why don't we just eliminate the pop cans altogether? Reductions in obesity, global warming, wasteful petty commerce, what's not to like?
Over all, I think they've handling this well. I just wish they would display the number of views remaining where it greats me every day with the phrase "Welcome, ohsupremeleader".
Actually, my NYT user name is from the Slartibartfast school of almost filthy. Refer to someone that way in the joint, they'd carry you out in a box. It combines the adolescent directness of fartface with the oozitude of sloppy seconds. I guess I was none too thrilled about the mandatory sign-up.
But honestly, they have pretty good content, and deserve to stay in business.
Honestly, I laughed out loud at your use of the word "thinking". I detect no thinking here whatsoever.
Thinking would involve a detailed cost/benefit analysis written as a terms of reference for a viable transition plan. It wouldn't sound like a drunk at the Politburo playing daisy with the purse strings of power, where terms of reference are never more than a slam of a shoe.
To protect an existing profit level higher than nearly any technology company that has ever existed. Corporations, like technologies, have natural life cycles. Plus, Microsoft is discouraged by law from exploiting it's entrenched monopoly in many ways it might have liked to.
Ballmer sucks in any event, without ignoring the first week of economics 101.
Dyslexic much? Don't think you meant 1895, though it's more plausible than what you wrote. 1958 is the only vaguely plausible coordinate. The people who talked up these techno-fantasies then were the same people on the waiting list for LSD.
In 1984 I was handed some bizarre Tokamak fusion propaganda by some kook in the SF airport which seemed like it came from a different planet. Many voices in the GOP now sound like they hail from the same planet. Useful as a premonition of mood. The fusion predictions, however, were not so good. By 1985 the only sane people who talked this way were people who had collections of Playboy, Popular Psychology, Popular Science, and Omni collecting dust under their beds. Sane, but a little behind.
The smart people could see the internet looming on the horizon. Had many talks about this. What we mostly failed to understand was the eventual economic model (which surprised nearly everyone) and that you wouldn't have to say "mother may I" to a utility company with every click. IBM was regarded as the epitome of evil control and there was much concern about escaping their dark Monopolis.
It wasn't until the dominance of wireless service that the world returned to the Mother-may-I format we might have predicted. Contrary to your assertion, none of us were lulled into regarding the Jetsons as a documentary.
The only future technologies I ever talked about was infomatics, genetics (anticipating the human genome project), radio astronomy, and a hint of nanotech. For the most part, I thought the Space Shuttle was a waste of money. The Hubble was a nice consolation prize, I will admit.
Moore's law is a lot like a paving machine. We would have given anything to know how much more asphalt was in the hopper. More than most of us suspected. How the brain struggles to comprehend 64GB (8*8 on the upcoming SandyBridge-E) when you're trapped inside 64kB without even a hard disk drive.
Strangely, fibre optics were closest to revealing their true potential.
Erbium-doped was first demonstrated in 1789. Er, I mean 1987. I remember reading about it for the first time in Scientific American probably a year before then. I nearly jumped out of my seat at the immediate implications. Without the use of a futuristic jet pack.
Yeah, I don't read all that much SciFi and the chart quickly lead me to a whole bunch of softcore SF classics, all of which I've read.
Here's a question: How does it make a SF title any better to have been written in the last hundred million seconds out of 100,000 years? Isn't keeping up with the present the domain of the Twitterverse?
I'm a well-aged consumer of scotch, cheese, movies, and books. If I'm going to consume something fresh, it's probably a documentary that took ten years to finance and film.
Every public company is required by law to become the next Microsoft if the business opportunity presents itself in order to provide maximum return to shareholders.
But then, you can take it to a whole new level by submitting falsified video tapes to the DOJ.
So yes, Google is getting grubbier, but it has yet to descend to flinging feces around like chocolate bon bons.
Mod parent up. In ten years I've never said that. Old dog, new trick. Also, was the mistake thinking we needed to do this too quickly?
Absolutely, there are many questions here unworthy of extended discussion. One doesn't to consult Donald Knuth to master quicksort. There's a huge literature on overdesign, discounting through net-present-value, adoption risk, and the difficult promulgation of technical standards.
TCP/IP has had roughly the same amazing thirty year run as x86. I've followed the twists and turns with x86 far more than IP, so I have a pretty good idea what I'd write on a napkin to send back to the original x86 design team--supposing it was actually a going concern to make the design not suck and that suckitude was not actually the criteria by which IBM originally chose this part, leading it's relevance thirty years later. (On the former question, I think Intel wanted to make a good part; on the later question, I think IBM was seriously gun shy about the future arriving quick and capable. In Terminator mythology, the napkin to Intel would fail, and be followed by a cyanide capsule to quavering bean counters over at Big Blue.)
First suggestion for Intel: offset the segments by 8 bits instead of 4 bits. This gets you a 16MB address space which gets you to viable 32-bit multitasking without a foray through DOS extenders. Let me put this iron into the fire to demonstrate, in a few minutes, the pleasure it will bring to thousands of programmers if you don't follow this advice. I think there would have been very little practical cost to this--even if I don't tell them that segmented data protection would fail. That feeble 20-bit address space was penny wise and pound foolish any way you look at it.
Second suggestion: either set all the status bits, or none at all with extremely few exceptions. In the short term this simplifies compilers, in the long run it expedites data flow implementations (i.e. OOO scheduling).
Third: Variable instructions length is a mixed blessing. Good if the first byte determines instruction length. Have at most one byte in the instruction which encodes allowed combinations of what might otherwise become prefixes/overrides. Variable instruction length improves icache density; a 30% efficiency gain over 32K of icache more than pays for 500 extra transistors in the instruction decode logic. Think about slope and intercept on instruction encoding density. Ideally your first byte tells you if you have a override bitfield and which byte contains it.
Fourth: Have more index registers than just SI and DI. This one is tricky since any expansion of the register set slows task switching. Maybe it should wait and become "have more index registers than ESI and EDI". But I think not. 16 registers is plenty in a RMW architecture. You don't need 32 registers unless blinded by RISC purity.
Fifth: Plan to transition the x87 co-processor from stack-based to register addressed at the earliest possible opportunity.
Sixth: Add an instruction to bit reverse bytes. Add an instruction to popcnt bytes. Chicken and egg problem. If you build it, they will come. Always have an instruction to byte reverse your longest supported register.
All the other x86 complaints: ignore completely. It's the future trying to eat your lunch after depreciating your opportunity cost and adoption risk to zero.
If you recall clearly what the world looked like when ideas are first on the drawing board (x86 is my personal case study for thinking about this), it's a lot harder to send advice back through the time machine that isn't actively dangerous.
How many of my suggestions would have rewarded them right out of the gate? Not a one, unless they put a full court press on high quality C compiler, to exploit the suggested register
Santa just called. He wants his elf-master of the list back before this junket of free association into the infinite void permanently curls his toes into cranky hang nails. Santa's old experiment with CRM114 automation did not go well. Return the elf, now!
Mob psychology is the echo chamber of common sense. And that's the good outcome. Even worse is nature/nurture where there was never much common sense to begin with.
Yes, there are correlates on both sides despite one or more mixing rounds of bent functions.
How does one perform medical epidemiology on an encryption block your kid sister wrote? Let's say your kid sister is Judit Polgar and she's almost smart enough to get this right (having not actually majored in math or computer science), but then you truncate to two rounds, so it's more like scrambling eggs with a spatula rather than an egg beater. This is a good model for the human nature/nurture system.
Permute the keys bits holding the block bits constant. Permute the block bits holding the key bits constant. Throw this into a powerful statistics whizinart, then press "publish" to gasps and wows from the bleachers of humanity who are slow to grasp that their common sense on this matter is six feet under.
Yeah, you can probably partition into dozens of sub-regions of statistically significant linearity manipulating the input bits on either side. There's many discernible chunks of white and yolk in the spastic scramble.
Even if you take major features of culture (such as our universal 12 year educational system, which represents about a millisecond of our 80ka recent history) and correlate genes most amenable to this, there's a wide span of orbital radius as experienced by any particular member of the population.
If a trillion dollars worth of epidemiology tells you less about a person on a quick reading of their gene chart than you get from an astute five minute introductory conversation, what exactly has all this research accomplished?
What we will find eventually are a few genes or gene complexes which correlate strongly with the effectiveness/ineffectiveness of proposed interventions (such as reading assistance). The exact fractal coefficient on little peaks of signal exploitation is certainly a quantity of interest moving forward. I'm pretty sure it's bounded from above by log (panacea).
We'll all become a tiny bit better at playing to strength. And most of the rest of the signal will fall below the noise floor of messy human affairs.
Here's the surprise twist: astute assessment is thin on the ground (the sap to syrup problem) and nearly impossible to institutionalize. So unless our machines become astute ASAP (as some predict), we'll probably press forward with the institutional seal club of genomic aphorism.
My comment in that thread was that 90% of the workforce will be unemployed within half a century on present trends. We're heading into a post-commodity winner-take-all employment climate. You'd like to think present trends will change. But how?
I once read a lot about the publishing industry. There were a lot of people in publishing who really cared about books. They faced an increasingly faced a harsh reality, until only the harsh survived.
For every Rowling making a billion dollars (she says less) there are thousands of small time authors crowded out of the marketplace.
Does disintermediation contribute to a long tail wagging happy to see you? Or do we just end up concentrating more wealth among the winner-take-all block busters, while the back-catalog brims with also-rans and failed contenders that never cover the author's outlay?
Gleick has done some highly regarded work. I waded through some material on his web site many years ago, and felt a strong respect.
From my old notes, here's an audio interview about a previous book. A Miracle Made Lyrical: Jim Gleick's Isaac Newton
Also high praise for Chaos from I Missed the Complexity Revolution
I don't understand how this reviewer has never heard of Chaitin, but finds this book vastly too elementary. Oddly, I mentioned Chaitin in an earlier post this very day. Perhaps reviewer should tear a page out of the Roger Ebert school of criticism:
That's a lot of words to pour out without defining expectations or genre. And there are many sub-genres within science writing.
Rubbish, if you expand your horizons wider than oracles of ticker tape or Back to the Future parlour tricks.
There have been some pretty profound visionaries over the centuries. Jules Verne, da Vinci, Richard Feynman (There's Plenty of Room at the Bottom), Claude Shannon, Freeman Dyson (space chickens), Charles Babbage, Leibniz, William Gibson (cyberspace), Marshall McLuhan (global village), Archimedes if you could get him to talk. These are not men immortalized for aping Minority Report.
I shake my head at all these Margulis extropians, who think we're headed for post-sexual merger with mechanoid symbiotes, the under-skin super suit. Which would be cool if I had any clue what the 90% of world's population, the unemployed, will be doing with all that time.
The future is a moving target. Set your sights accordingly, and recognize transcendent wisdom bereft of gadgets.
I'm pretty sure all those salmon are swimming upstream to discover the cure for senescence ending the nasty business of spawning for ever and ever.
This is the least conservative line of investigation in the history of the human species. Stem cells? Totally passe. Surf upon mighty Morforgeddon, and despair!
Hi. Will you marry me? For a week?
But it will lower the abortion rate, so let's not be hasty to call it a bad thing. An aging base could potentially be a godsend once we repeal term limits.
Every true geek must read Metamath! or at least vaguely grok the concept.
It's the finishing touch on a brilliant little geek high-ball made from Heisenberg, Turing, Godel, Kolmogorov, Chaitin with a Mandelbrot cherry. Toss it back then bite the lemon! Warning: There's a lot of peel in A New Kind of Science. Chaitin is short and dense and accessible to a thinking 15 year old, for some value of "ignores assigned homework". Wolfram is pointing out that the mathematical Chaitinverse is a five minute walk from the financial district; a few steps past the edge of city limits you spot your first cactus.
There exists a message which this little strip of paper is too small to contain. Bursts into flames.
The significance of trillions of digits of Pi is akin to the Apollo program. An alien civilization who spies a little shard of metal emitted from the earth's atmosphere, winging very directly to the moon where it loiters for a few orbits, then winging back to earth (in a vacuum!) would probably be thinking "Petunias! Whale meat!" Life, in other words.
Catching a glimpse of a million digits of Pi from the billionth page on the Hollywood hills of a distant galaxy, they would think exactly the same thing--after Zaphod Spader conducts the kindergarten refresher class on algorithmic complexity theory.
The only paragraph I found at all useful:
The rest is all mystery meat about the actual method.
This is certainly the way of the future. IBM has pumping that notion right now, as well. I'm not inclined to think this is a huge step if they've only found signal under Gaussian noise.
The rule of thumb is for the top of your screen to be at eye height sitting erect. Any higher causes your lids to lift (so I once read) and can lead to dry eyes.
I personally had back problems associated with using the mouse with my right hand on the other side of the numeric keypad. Changed to my left hand, which is less used for other things. This moved the mouse closer to my body line and reduces stress on the upper shoulders. If I really need precision, it's easy enough to reach across with my right hand for a few minutes. My left hand became automatic for normal cursor work a long, long time ago. My right hand still feels a bit more precise.
I just did a comparison drag selecting words with right and left hands. The right hand feels more interactive, but it was uncertain of the mouse velocity in the select motion and actually took longer. My left hand much more accurately sweeps the word length. Consciously, I don't sense this, but objectively there it is.
The colour temperature of your monitors can affect stamina and mood. This varies a lot from one person to another. Recently scientists are saying that a high colour temperature boost productivity. And office temperature matters, too.
Scientifically Proven Tips For a More Productive Office
I once had a set up where I was experiencing eye strain. When I illuminated the desk in front of the displays with a warm yellow desk lamp, the eye strain went away. Uniformity of illumination also matters. Glare from behind the monitors is a killer.
A seat pan with tilt is a nice adjustment. I don't see the need for stupidly expensive chairs. I've had a couple. You really have to try different chairs and find one that works. If you're big like me (6'4" 240) you'll need to supplement the cushion on most medium grade chairs. I'm not sitting in a cheap chair, but without the extra cushion I soon feel the nuts and bolts. Pay special attention to the lip at the front of the chair where it cuts into the back of the leg.
Pay attention to pain and try to do something about it. It's very easy to negatively condition yourself to associate work with pain. Then you have a psychological barrier to fight with, too.
I got myself a full-sized Thumper almost ten years ago. Never had a problem with it. It's nothing close to a professional massage, but it provides welcome relief on those long, hard days. My chiropractor has the newer Maxi Pro and it's certainly a step up. I think it's a complete joke that it recommends no additional pressure beyond it's seven pound size. Yeah, that works fine for my quads. When it hits the tightness in my low back, it just sits there and bounces without some manual force. The tightness in my low back is epic sometimes. For a while I was doing a back exercise program at a clinic with special back equipment. On some machines I started with just a few plates like the wimp I am. On the back extension machine, it was the whole stack from day one. How many of these should I do with my cords of steel? I was running out of wind before feeling the burn. Everyone has their own thing. One of the best investments I've made.
The best upgrade to a Thumper is a patient and willing GF (with sharp elbows). These are best acquired with skill in the kitchen (with the contents of the refrigerator! cooking food served on plates!!) Invest in a good set of knives, heavy bottom pans, and a traditional iron wok. Fill you spice cupboard. Experiment. It's 90% water and heat. Trade your Aeron chair for Modernist Cuisine. I haven't gone quite that far myself, but my cooking skills are pretty good already.
Determine your caffeine tolerance level. More than one eight ounce cup every four hours is rarely optimal. If you get a headache when you miss your morning coffee, you're
Any narrow advantage held by the Power architecture was quickly disappearing by 2003 when the first Operton chips with AMD64/SSE2 hit the market (for users able to jump to 64-bits), and pretty much obliterated with the introduction of the Core Duo in January 2006.
x86 had an ugly childhood, but it turns out there wasn't anything desperately wrong with its performance potential. Jobs made such a big deal of x86 being somehow deeply inferior. The x86 is deeply inferior when you try to build efficient devices under a watt. The little tricks used in x86 to achieve high performance are expensive in power consumption and nearly impossible to fix without an instruction set overhaul. ARM was the true victor.
You really have the edict entirely backward. It was Jobs' edict that Apple would slug it out on a platform with a small market share, and for which IBM could not afford to develop to market-leading performance on an indefinite basis. Finally Jobs lifted the anti-Intel edict because he had no choice. Not just in performance, but also available production quantity.
When you get right down to it, I'm sure Jobs regarded PowerPC as a convenient walled garden. He didn't want his miracle machine to become too compatible. But it was probably also a huge burden to carry your own ISA for 10% market share. Look how they've done since.
The money made by Microsoft, Apple and Google, 1985 until today
The decade of PowerPC pretty much corresponds with the Apple doldrums. Imagine that.
I'm curious whether his reality distortion field penetrates from the after life, or whether Gandalf the White will arrive to perform an exorcism at long last.
When I was younger, I used "because" and "since" in my writing about twice as often. Never terribly pleased by the effect--that's just how it came out. They are fairly weak transitions, useful mostly if you want a weak transition which detracts less from a central element.
This excess tapered off as I became more deeply immersed in my subject matter with age and experience. In my own history, these words were sign posts of incomplete thinking.
That's one way to do it. I might even agree if we deleveraged corporate ownership.
If A and B own 51% and 49% respectively of C, and C owns 51% of D, then when there's a vote in company D, B gets .49*.51 percent of the vote. But that's not how it presently works.
Do you even see the connection? Wallets don't own and control other wallets.
Tell that to my Himalayan salt or my Mediterranean sea salt, both of which imply unspecified exotic trace elements. I'm not quite so willing to energize on white bread custom as to equate the two. For brevity, I keep a shaker of NaCl.
Amen, brother.
3.0 woody on 19 July 2002
3.1 sarge on 6 June 2005
Back in the day, that was a long hiatus in internet dog years. And what I depended on especially at the time was an up-to-date LAMP stack. I read a lot on the Debian list at the time and I can't say I came away impressed with their sense of priority during that time.
Now Ubuntu has crumbled beneath my feet. I'm really tempted to jump to FreeBSD. But then I have a project coming up to play with OpenCL and I'm not sure FreeBSD will facilitate getting started. Just love it when my bug reports have a unique in the galaxy hardware configuration. People jump right in to help you out.