As the old sayings goes -- or should -- "correlation is not causation, but it's a good place to start looking."
Pointing out that AE->B is useful because it shows other people where to look. It increases the amount of information. If we don't know that, we don't know where to look, but if we DO know that, we have a narrower area to investigate. That's why it's useful. Sure, the public will misinterpret this. But the public always misinterprets science, so that's not really much loss. If you're arguing that we shouldn't do research because it'll mislead the public, we're screwed. Look at this as basic research, similar to what particle physicists are doing: bashing things together to see what happens, and then letting someone else try and stitch that into a theoretical framework. That's why I was on about empirical science as opposed to theoretical science: it's valid to just do a zillion experiments, or even just observations, with no hypothesis and then publish the results, because someone else can use that data to help form a hypothesis. We didn't get the equation for gravity from theory, we got it from doing gobs of experiments on how fast things fell.
>some weird correlation between two factors that are obviously not directly related. What's a researcher these days, someone who gathers a whole bunch of data, looks for all the statistical correlation they can find and publish a paper as soon as they find "something"
But here's the thing: they did find "something", that is statistically significant. That means there *is* a direct relationship. That's the whole point of statistics. They're establishing that there's something going on, and then they go looking for what. The relationship may be direct (causation) or symptomatic of something else (correlation) but they are, indeed, showing that there *is* a relationship. There are a lot of books out there that discuss precisely this: what happens if you just go digging through data randomly and find statistically significant correlations, and then try and figure out why. "Freakonomics" discusses how there's a direct linear relationship between a mother's educational level and how she alters the spelling of her children's names. That's a *weird* correlation, but it is demonstrable. A more famous example -- from 1920, so it's not like weird correlation is a modern phenomenon -- is George Taylor's hemline index where he showed that the length of skirt hemlines was strongly correlated with the economy. And there are explanations for why that was/is happening. Sometimes weird things are interrelated. When people find them, it's a good place to go looking for underlying mechanisms. You can do good theoretical science by formulating a hypothesis and then testing it, but you can also do good empirical science by finding demonstrably interrelated systems and then figuring out how they are interrelated.
What you're talking about is called the Hygiene Hypothesis and there's a lot of evidence for it, although it's pretty hard to test.
But as for air quality -- the air quality in large cities was much worse in the '60's, but fewer people lived in cities. Rural air quality has been steadily declining, urban air quality has been increasing rapidly since about 1975, but urban air quality is still poor, especially as regards more or less inert particulates kicked up by traffic, and people keep moving from rural areas to cities throughout the world.
>but it stomps out generating new keys as much as you want.
Sort of. As the previous poster was alluding to, if the card numbers are generated sequentially and stored on the card, all you need to do is know your number, add about 100, put that number on your card, and wait for it to be activated so you can use it. You don't have to access the main server: you just wait for your number to show up. There was a neato scam running a while back where people would steal piles of seemingly useless blank gift cards, record the number off the card into a database, put them back in stores, wait a month, then try and use the number. If the card had been activated but not used (a gift card sitting in a present or a wallet somewhere) they bought what they could as fast as they could. I assume companies now sell entirely blank cards, that are programmed at time of sale, rather than pre-enumerated cards merely being scanned for activation.
I don't know about *your* VW Rabbit. *MY* VW Rabbit didn't have interior trim, at all. It was just bare metal on the inside. I think there was a panel covering the inside of the door, but there wasn't anything covering walls beside the passengers in the back seat other than bright yellow metal. The windows were manual, the two of them that rolled down. As I remember, it had a defroster and a heater that got to my feet, and a side rear-view mirror only on the driver's side, that was adjusted by opening the window and pushing the mirror where I wanted it to go. And of course manual doorlocks.
The point being: it's not primarily safety features that drive the weight gain of modern cars. It's comfort features that have come to be expected. I've read claims that you can quite accurately predict a car's weight if you know how many electric motors it includes. Even if you omit much of the hardware in the base model, the designers have still included the bracketry that would support those options, and often the wiring and other support equipment. (The wiring harness for a Ford Mustang weighs almost 30 kilograms; if stripped down to the bare essentials of legally required lighting, it loses 12 kilos, as an example I know off the top of my head.) Air conditioning, four-way in-door speakers (did the Rabbit even have a radio? I can't remember. I know my Jeep from the same time period didn't.) automatic windows, a decent climate control system, even little things like a parking brake that's shaped like my hand rather than just being a steel pipe with a button on the end, which is what I remember the Rabbit having, all add a lot of weight.
Safety equipment does add a lot of weight. But at the same time, many of the heavy -- and expensive -- bits like hardened alloy steel reinforcement in the doors and crumple zones designed to make sure the engine doesn't end up in my lap, are things I value more than gas mileage.
I'm not Gladwell, and I won't presume to argue his case for him. But, as I said, "He's not saying that only people born then become successful, or that every person born then becomes successful" just that a disproportionate number of people from a specific cohort, became bigwigs. He shows the same thing in Canadian hockey, in Asian school math scores, and a bunch of other areas. The book isn't about how Bill Gates became famous. The book is about how society sets up invisible gating systems in its attempt to select people based on merit, and then incorrectly concludes the people who succeed do so because of their merit. He spends a lot more time talking about the statistics that support his argument when he's talking about Canadian hockey: there really aren't enough tech billionares to get a trustworthy statistical analysis. However, once you've shown that you do have a trustworthy result, statistically speaking, you can go to other areas and show how they also look very similar.
Something the reviewer doesn't make clear, that Gladwell spends a lot of time talking about, is just how important BillG's birthdate was. There was apparently a narrow window of opportunity; if you were born before that, you were already entrenched in another field when computers became huge, and if you were born after that you couldn't ever manage to stay up: you weren't in the bubble. Gladwell submitted as evidence that the window of opportunity was about two years long, and in those two years were born Bill Gates, Steve Jobs, Bill Joy -- well, here, let me just quote a partial transcript of Malcolm Gladwell talking about the book: "January 1975 was the dawn of the personal computer age. The perfect age to be in 1975 is young enough to see the coming revolution but not so old as to have missed it. You want to be 20 or 21, born in 1954 or 1955. Bill Gates, the co-founder of Microsoft, was born on October 28, 1955. Paul Allen, the other co-founder of Microsoft, was born on January 21, 1953. Steve Ballmer, the present CEO of Microsoft, was born on March 24, 1956. Steve Jobs, the co-founder of Apple, was born on February 24, 1955. Eric Schmidt, the CEO of Google, was born on April 27, 1955. Bill Joy was born on November 8, 1954."
He's not saying that only people born then become successful, or that every person born then becomes successful, but that people born then have a much greater chance of being successful in a particular field that is just opening up.
>The fact that they're give it a specific time, down to the minute, would imply that all things are already decided.
Don't read stuff in there that's not there. It's like Patch Tuesday: it's not that the hacks come out on Monday. This happens a lot in decisions. There might be a Swedish law saying you have 40 days (or something) to come up with a verdict, and if it's not a law, there's some similar guideline set up that produces a deadline for the decision. I've been involved in a lawsuit and they had the time for the verdict already set up before the trial was finished. (Which is nice: a friend is involved in a child custody dispute and although the testimony ended two months ago, the judge still hasn't issued a decision and they have no idea when they'll know. Having a set deadline avoids this sort of situation.)
Lawyers aren't dumb. Everyone involved would have known this case couldn't be won two years ago. This whole thing was intended to provide evidence for the need for new laws.
If you want to be even cheaper, make your own. National Semiconductor LM3405 driver chip and half a dozen associated components (this link has what I consider to be the snazziest design software I've ever seen, that'll crank out a list of the precise parts you need and even send them to you) plus a couple Philips LumiLED LXHL-BW02's from Future Electronics (the cheapest source) and you have a lovely little light that'll run for days off a 9v battery. Since a 9V is a crappy way to run a light, price/performance-wise, you could use a boost driver like an LM2623 to run your LED's off 2 AA cells, and that's an easy design too. If it's not obvious from the foregoing, I design parts of these chips, so my referrals of their site isn't altruisic. They're good chips, anyway. Although so are Maxim's.
>What is needed is a really efficient current driver IC for LEDs.
They exist in large, large quantities. My company makes three dozen, currently, with another three dozen in development. So do half a dozen of our competitors. They also add about $1 to the cost of the lightbulb, and consumers mostly buy the bulbs that cost $1 less.
>it seems like our entire economic system has been tailored to next quarters results at the expense of building/investing for the long term.
It doesn't *seem* like that, it *is* that. Precisely that. Investors don't care what a company makes: all they care about is maximizing the return on their investment. That's the only reason they invested. A company cares about what they do, but they're required by the market to maximize their profit to retain investors. A company that thinks long-term gets outcompeted by short-term thinking, because it loses investors to the short-term companies. Everyone in this scenario is acting in their own self-interest: they're all behaving rationally and they're doing the best possible thing for their goals. The result is what we see: financial disaster, both for individual companies who burn themselves up trying to chase quarter-by-quarter profits, and for society as a whole when this behavior crashes parts of the financial sector. I've compared this in the past to a herd of deer running from a wolf towards a cliff. They all know that if they keep going they're all going to die, but they also all know that if any of them stops, that one is going to die sooner, so they keep running.
Conventional wisdom is that in cases like this, government regulation -- action by a body whose rational self-interest isn't driven by short-term considerations -- is called for. That's not a very popular viewpoint here, but I don't see any alternatives.
>The managers making the decisions didn't know what it all meant and the guys using the model didn't adequately explain the model's limitations.
Or the managers didn't understand their explanations -- or more likely yet, didn't *want* to understand their explanations. This doesn't look fundamentally different than the Challenger explosion: the technical staff knows there's a problem, keeps saying that there's a problem, but their upper management is invested in there not being a problem. It's really difficult to explain something to someone whose job depends on ideas that conflict with what you're explaining.
Or they were trying to bleach it and clean off all the accumulated crap, and found that it did that *and* changed the surface chemistry as well.
I used to restore old watches -- the mechanical ones. They're jammed with grease and wax that mechanics didn't put there and usually that's a large part of why they're not working. It's fairly routine to dunk something that looks like the Antikythera Mechanism into a cleaner just to get all the horribleness out. (A note to anyone considering doing this: avoid ammonia. Those are very delicate little bits of brass. Ammonia works spectacularly well. If you leave a watch movement in there for an hour rather than just a few minutes, you'll come back to find all the wax *and* all the metal completely gone except for the steel and a few of the large pivots.)
>Maybe coupled with some outboard piece of USB gear
AVR-USB: a firmware-only USB interface for atmel microcontrollers. Two resistors and two diodes, one ATTiny microcontroller, two power-ground caps, and you have your hardware interface on a board the size of a postage stamp. This reference board has 8 channels of power switching, that could drive reasonably large relays like, say, 8 Jameco 134949's (at $5 each), that'll each switch 20 amps at 220volts.
It could drive a three-axis CNC mill. Plug in a webcam and you have a security system. Add a relay board and you have a portable autonomous sentry setup. A usb-to-serial converter combined with fuse and owfs and a half-dozen Dallas Semiconductor one-wire devices and you can put dirt cheap thermometers all over your house. Add some Dallas I/O chips and some motors and you have zone heating for your house.
I can think of *lots* of interesting things to do with this.
Severance? I've heard of that but never seen it. Everyone shows up at work and there's a mid-level manager with a list and two armed security guards standing at the front door, and if your name's not on the list you go home and they send you your stuff. I guess that's your severance pay, that they don't keep your stuff.
Speaking as a pilot, I have to say it's landing that's more difficult. There are autopilot systems that can take off, fly, and land, but landing's the challenging one. But in any case, the reason you have a pilot is for when things go wrong. When things are going right, a monkey could take off, fly, and land a plane. When things go wrong, it takes knowledge to know what has gone wrong, and how to survive it. That's where the difficulty comes in.
I can't find the graph online but there's a neat graph that's included in any primary instruction manual, showing workload vs. fatigue for a typical flight. The highest peak on the workload graph is just before and during landing, which is also the lowest point on the fatigue curve, hence much of the reason that landings are stressful.
Now, disasters during takeoff are much more dangerous. They happen fast and you have lots of fuel and weight. But they're also much more rare, thankfully.
This is why, traditionally, men walked on the street side and women on the building side: because of the splatter-up from horse crap.
Once I was riding cross-country (by which I mean no road or trail) with my dad, through heavy bracken, and when we got out and into a meadow we both had cuts all over our legs. We came down a fast descent and found out it was a cattle pasture. A fresh cattle pasture. Nothing like using up all your water washing fresh cow plop out of open wounds on your legs.
The point I learned to bunnyhop was actually a bike race in Oregon. We were coming around a hard corner and I hit a banana slug with my rear wheel. Those things are very nearly frictionless. Rear end came around, guy behind me piled into me, and we both went off a sort of cliff and into a muddy creek. That banana slug had an okay revenge for its death.
I was thinking about poo on roads the other day, actually: around here, Canada Geese are the main antagonist. I realized that it's like playing Space Invaders: you have this narrow band of pavement coming towards you and you have to dodge all the static debris (goose droppings: nasty, nasty stuff, very slick) and sometimes moving stuff (the geese themselves) that's sometimes just moving randomly (avoiding you) and sometimes actively trying to hit into you (when they have babies they get hostile and come at you in a hurry -- until they realize you're going 50k and aren't going to stop.)
I knew a guy who was so mad at Canada Goose poo in his yard he started luring the geese into his garage and killing them. (and eating them.) Not as easy with horses, though. People complain more.
>Just mention horses, public roads and an affinity for cycling (me, not the horses).
Learn to bunny-hop. I admit the rightness of your claim, but they're not going to change. And a good bunny-hop technique can save your neck when a car decides to lurch into your space because the driver's an idiot or asleep or having an aneurism and you need to cross a curb without breaking your collarbone. Ditto dealing with people who throw crap at you and you don't want to hit it, unexpected potholes, grates, and even crossing railroad tracks at an angle. It's a fairly easy skill to learn. The easy way relies on toeclips or clipless pedals. The cool way, you can do on bmx pedals but it's harder to learn. Look up Ryan Leech on YouTube: he has some instructional videos on this.
Sorry about the lag: vacation. We never got the rabbits to stop chewing on cables. So to prevent dead refrigerators or computers, we used: close supervision lifting all the power cords up over a foot off the floor running anything critical in flexible metal conduit. And of those, I'd say only the last one was truly effective. With two rabbits, you just can't do effective supervision, realistically speaking, and even though mine are really tiny rabbits, like a kilo a piece, they can still stand up a lot taller than you'd expect to get to stuff they want to eat.
I have/had a similar set of problems. 1. I've owned pet rats and know what they can eat. 2. I currently own house rabbits. 3. I have a recurring mouse problem.
So lemme sum up. A rabbit can eat through a 14 gauge stranded copper cord of the sort you'd use for your refrigerator. Guess how I figured that out? Since it's starting at one side it doesn't ever cross both the live and neutral, so it doesn't get electrocuted. It can eat every cord off the back of a computer in under three minutes. Guess how I figured that out? A rat doesn't have quite the toothy abilities of a rabbit but it's fairly close. They can certainly cut through thin copper. Neither the rabbits nor the rats -- nor my dog -- have been bothered by sprays intended to keep animals from digging/chewing on things. The super hot pepper-derived stuff stopped the rabbits but not the rats, and my dog loves the stuff. The sour/bitter stuff didn't slow any of them down even slightly. Plastic split conduit doesn't even slow them down. Even when soaked in bitter or hot do-not-chew stuff. Rats can chew through the side of a lead pipe and crawl through a hole the size of a US quarter. I don't have evidence that they can chew through copper pipe but I wouldn't be surprised. Reducing food doesn't work. Once they're established, you can't keep the place clean enough. I have no idea how wild mice manage to find nutrition but they do. We keep all our food in sealed containers and vacuum and roomba every other day, and neither the dog nor the sometimes cat deter the mice in the slightest. The mice do, however, drive the dog and the cat completely insane, so if you want to have your predator madly clawing at the wall where it can either hear or smell a mouse, go for it. Both dogs and clawed cats can dig through standard drywall, and then you have a repair to do. (and they remember it and keep trying. Pitbulls are very, very retentive dogs and she'll dig through 12mm thick plywood to get to where she remembers a mouse or rat or squirrel to have hidden, once, six months ago.)
hate to say it but d-con and other awful poisons are probably the best way to go, as far as eradication, and flexible conduit to protect the lines you can't easily replace.
As I said elsewhere, glue traps are probably more evil than poison, and oftentimes live traps are as well, because you don't check them often enough and the animal dies of dehydration. And if you're really lucky the animal will manage to drag the glue trap into a place you can't get to and if you're young and still have good ears you can hear its little high-pitched screams for a couple days before it does die.
My girlfriend's landlord put in glue traps. We started hearing this weird high-pitched noise, like a flyback transformer in a CRT going wonky. It was mice screaming because they'd been stuck to the glue trap for days and were starving to death.
If you want to get them loose without killing them don't pull them. A: you pull parts off and B: they're still covered in glue so they just stick to debris and fluff and leaves and whatever else. Vegetable oil will get them loose. Just, y'know, don't use a lot or you end up with drowned oily mice.
Slashdot Mirror
As the old sayings goes -- or should -- "correlation is not causation, but it's a good place to start looking."
Pointing out that AE->B is useful because it shows other people where to look. It increases the amount of information. If we don't know that, we don't know where to look, but if we DO know that, we have a narrower area to investigate. That's why it's useful.
Sure, the public will misinterpret this. But the public always misinterprets science, so that's not really much loss. If you're arguing that we shouldn't do research because it'll mislead the public, we're screwed.
Look at this as basic research, similar to what particle physicists are doing: bashing things together to see what happens, and then letting someone else try and stitch that into a theoretical framework. That's why I was on about empirical science as opposed to theoretical science: it's valid to just do a zillion experiments, or even just observations, with no hypothesis and then publish the results, because someone else can use that data to help form a hypothesis. We didn't get the equation for gravity from theory, we got it from doing gobs of experiments on how fast things fell.
>some weird correlation between two factors that are obviously not directly related. What's a researcher these days, someone who gathers a whole bunch of data, looks for all the statistical correlation they can find and publish a paper as soon as they find "something"
But here's the thing: they did find "something", that is statistically significant. That means there *is* a direct relationship. That's the whole point of statistics. They're establishing that there's something going on, and then they go looking for what. The relationship may be direct (causation) or symptomatic of something else (correlation) but they are, indeed, showing that there *is* a relationship.
There are a lot of books out there that discuss precisely this: what happens if you just go digging through data randomly and find statistically significant correlations, and then try and figure out why. "Freakonomics" discusses how there's a direct linear relationship between a mother's educational level and how she alters the spelling of her children's names. That's a *weird* correlation, but it is demonstrable. A more famous example -- from 1920, so it's not like weird correlation is a modern phenomenon -- is George Taylor's hemline index where he showed that the length of skirt hemlines was strongly correlated with the economy. And there are explanations for why that was/is happening.
Sometimes weird things are interrelated. When people find them, it's a good place to go looking for underlying mechanisms. You can do good theoretical science by formulating a hypothesis and then testing it, but you can also do good empirical science by finding demonstrably interrelated systems and then figuring out how they are interrelated.
What you're talking about is called the Hygiene Hypothesis and there's a lot of evidence for it, although it's pretty hard to test.
But as for air quality -- the air quality in large cities was much worse in the '60's, but fewer people lived in cities. Rural air quality has been steadily declining, urban air quality has been increasing rapidly since about 1975, but urban air quality is still poor, especially as regards more or less inert particulates kicked up by traffic, and people keep moving from rural areas to cities throughout the world.
>but it stomps out generating new keys as much as you want.
Sort of. As the previous poster was alluding to, if the card numbers are generated sequentially and stored on the card, all you need to do is know your number, add about 100, put that number on your card, and wait for it to be activated so you can use it. You don't have to access the main server: you just wait for your number to show up.
There was a neato scam running a while back where people would steal piles of seemingly useless blank gift cards, record the number off the card into a database, put them back in stores, wait a month, then try and use the number. If the card had been activated but not used (a gift card sitting in a present or a wallet somewhere) they bought what they could as fast as they could.
I assume companies now sell entirely blank cards, that are programmed at time of sale, rather than pre-enumerated cards merely being scanned for activation.
Look at McCain: he violated multiple copyrights with his ads, claimed that there should be a special exemption to the DMCA laws he voted for just for politicians, and once the campaign was over, everyone completely forgot about it except Jackson Browne, who is still pursuing a lawsuit against McCain, although it's generally considered wildly unlikely that'll go to trial (or that McCain will be fined $3000 for every case of infringement.)
I don't know about *your* VW Rabbit. *MY* VW Rabbit didn't have interior trim, at all. It was just bare metal on the inside. I think there was a panel covering the inside of the door, but there wasn't anything covering walls beside the passengers in the back seat other than bright yellow metal. The windows were manual, the two of them that rolled down. As I remember, it had a defroster and a heater that got to my feet, and a side rear-view mirror only on the driver's side, that was adjusted by opening the window and pushing the mirror where I wanted it to go. And of course manual doorlocks.
The point being: it's not primarily safety features that drive the weight gain of modern cars. It's comfort features that have come to be expected. I've read claims that you can quite accurately predict a car's weight if you know how many electric motors it includes. Even if you omit much of the hardware in the base model, the designers have still included the bracketry that would support those options, and often the wiring and other support equipment. (The wiring harness for a Ford Mustang weighs almost 30 kilograms; if stripped down to the bare essentials of legally required lighting, it loses 12 kilos, as an example I know off the top of my head.) Air conditioning, four-way in-door speakers (did the Rabbit even have a radio? I can't remember. I know my Jeep from the same time period didn't.) automatic windows, a decent climate control system, even little things like a parking brake that's shaped like my hand rather than just being a steel pipe with a button on the end, which is what I remember the Rabbit having, all add a lot of weight.
Safety equipment does add a lot of weight. But at the same time, many of the heavy -- and expensive -- bits like hardened alloy steel reinforcement in the doors and crumple zones designed to make sure the engine doesn't end up in my lap, are things I value more than gas mileage.
I'm not Gladwell, and I won't presume to argue his case for him. But, as I said,
"He's not saying that only people born then become successful, or that every person born then becomes successful"
just that a disproportionate number of people from a specific cohort, became bigwigs. He shows the same thing in Canadian hockey, in Asian school math scores, and a bunch of other areas.
The book isn't about how Bill Gates became famous. The book is about how society sets up invisible gating systems in its attempt to select people based on merit, and then incorrectly concludes the people who succeed do so because of their merit. He spends a lot more time talking about the statistics that support his argument when he's talking about Canadian hockey: there really aren't enough tech billionares to get a trustworthy statistical analysis. However, once you've shown that you do have a trustworthy result, statistically speaking, you can go to other areas and show how they also look very similar.
Something the reviewer doesn't make clear, that Gladwell spends a lot of time talking about, is just how important BillG's birthdate was. There was apparently a narrow window of opportunity; if you were born before that, you were already entrenched in another field when computers became huge, and if you were born after that you couldn't ever manage to stay up: you weren't in the bubble. Gladwell submitted as evidence that the window of opportunity was about two years long, and in those two years were born Bill Gates, Steve Jobs, Bill Joy -- well, here, let me just quote a partial transcript of Malcolm Gladwell talking about the book:
"January 1975 was the dawn of the personal computer age. The perfect age to be in 1975 is young enough to see the coming revolution but not so old as to have missed it. You want to be 20 or 21, born in 1954 or 1955.
Bill Gates, the co-founder of Microsoft, was born on October 28, 1955.
Paul Allen, the other co-founder of Microsoft, was born on January 21, 1953.
Steve Ballmer, the present CEO of Microsoft, was born on March 24, 1956.
Steve Jobs, the co-founder of Apple, was born on February 24, 1955.
Eric Schmidt, the CEO of Google, was born on April 27, 1955.
Bill Joy was born on November 8, 1954."
He's not saying that only people born then become successful, or that every person born then becomes successful, but that people born then have a much greater chance of being successful in a particular field that is just opening up.
>The fact that they're give it a specific time, down to the minute, would imply that all things are already decided.
Don't read stuff in there that's not there. It's like Patch Tuesday: it's not that the hacks come out on Monday. This happens a lot in decisions. There might be a Swedish law saying you have 40 days (or something) to come up with a verdict, and if it's not a law, there's some similar guideline set up that produces a deadline for the decision. I've been involved in a lawsuit and they had the time for the verdict already set up before the trial was finished. (Which is nice: a friend is involved in a child custody dispute and although the testimony ended two months ago, the judge still hasn't issued a decision and they have no idea when they'll know. Having a set deadline avoids this sort of situation.)
Lawyers aren't dumb. Everyone involved would have known this case couldn't be won two years ago. This whole thing was intended to provide evidence for the need for new laws.
If you want to be even cheaper, make your own. National Semiconductor LM3405 driver chip and half a dozen associated components (this link has what I consider to be the snazziest design software I've ever seen, that'll crank out a list of the precise parts you need and even send them to you) plus a couple Philips LumiLED LXHL-BW02's from Future
Electronics (the cheapest source) and you have a lovely little light that'll run for days off a 9v battery. Since a 9V is a crappy way to run a light, price/performance-wise, you could use a boost driver like an LM2623 to run your LED's off 2 AA cells, and that's an easy design too.
If it's not obvious from the foregoing, I design parts of these chips, so my referrals of their site isn't altruisic. They're good chips, anyway. Although so are Maxim's.
>What is needed is a really efficient current driver IC for LEDs.
They exist in large, large quantities. My company makes three dozen, currently, with another three dozen in development. So do half a dozen of our competitors.
They also add about $1 to the cost of the lightbulb, and consumers mostly buy the bulbs that cost $1 less.
>it seems like our entire economic system has been tailored to next quarters results at the expense of building/investing for the long term.
It doesn't *seem* like that, it *is* that. Precisely that. Investors don't care what a company makes: all they care about is maximizing the return on their investment. That's the only reason they invested. A company cares about what they do, but they're required by the market to maximize their profit to retain investors. A company that thinks long-term gets outcompeted by short-term thinking, because it loses investors to the short-term companies.
Everyone in this scenario is acting in their own self-interest: they're all behaving rationally and they're doing the best possible thing for their goals. The result is what we see: financial disaster, both for individual companies who burn themselves up trying to chase quarter-by-quarter profits, and for society as a whole when this behavior crashes parts of the financial sector.
I've compared this in the past to a herd of deer running from a wolf towards a cliff. They all know that if they keep going they're all going to die, but they also all know that if any of them stops, that one is going to die sooner, so they keep running.
Conventional wisdom is that in cases like this, government regulation -- action by a body whose rational self-interest isn't driven by short-term considerations -- is called for. That's not a very popular viewpoint here, but I don't see any alternatives.
>The managers making the decisions didn't know what it all meant and the guys using the model didn't adequately explain the model's limitations.
Or the managers didn't understand their explanations -- or more likely yet, didn't *want* to understand their explanations.
This doesn't look fundamentally different than the Challenger explosion: the technical staff knows there's a problem, keeps saying that there's a problem, but their upper management is invested in there not being a problem. It's really difficult to explain something to someone whose job depends on ideas that conflict with what you're explaining.
Or they were trying to bleach it and clean off all the accumulated crap, and found that it did that *and* changed the surface chemistry as well.
I used to restore old watches -- the mechanical ones. They're jammed with grease and wax that mechanics didn't put there and usually that's a large part of why they're not working. It's fairly routine to dunk something that looks like the Antikythera Mechanism into a cleaner just to get all the horribleness out.
(A note to anyone considering doing this: avoid ammonia. Those are very delicate little bits of brass. Ammonia works spectacularly well. If you leave a watch movement in there for an hour rather than just a few minutes, you'll come back to find all the wax *and* all the metal completely gone except for the steel and a few of the large pivots.)
>Ask him how he picks up women.
He says, "marry me and I'll convince total strangers to send you $4000 worth of flowers."
>Maybe coupled with some outboard piece of USB gear
AVR-USB: a firmware-only USB interface for atmel microcontrollers. Two resistors and two diodes, one ATTiny microcontroller, two power-ground caps, and you have your hardware interface on a board the size of a postage stamp. This reference board has 8 channels of power switching, that could drive reasonably large relays like, say, 8 Jameco 134949's (at $5 each), that'll each switch 20 amps at 220volts.
It could drive a three-axis CNC mill. Plug in a webcam and you have a security system. Add a relay board and you have a portable autonomous sentry setup. A usb-to-serial converter combined with fuse and owfs and a half-dozen Dallas Semiconductor one-wire devices and you can put dirt cheap thermometers all over your house. Add some Dallas I/O chips and some motors and you have zone heating for your house.
I can think of *lots* of interesting things to do with this.
How many Marxists does it take to change a lightbulb?
None. The lightbulb contains the seeds of its own revolution.
>Severance turns into a payed vacation
Severance? I've heard of that but never seen it. Everyone shows up at work and there's a mid-level manager with a list and two armed security guards standing at the front door, and if your name's not on the list you go home and they send you your stuff. I guess that's your severance pay, that they don't keep your stuff.
Needless to say, I prefer paycuts to layoffs.
Speaking as a pilot, I have to say it's landing that's more difficult. There are autopilot systems that can take off, fly, and land, but landing's the challenging one.
But in any case, the reason you have a pilot is for when things go wrong. When things are going right, a monkey could take off, fly, and land a plane. When things go wrong, it takes knowledge to know what has gone wrong, and how to survive it. That's where the difficulty comes in.
I can't find the graph online but there's a neat graph that's included in any primary instruction manual, showing workload vs. fatigue for a typical flight. The highest peak on the workload graph is just before and during landing, which is also the lowest point on the fatigue curve, hence much of the reason that landings are stressful.
Now, disasters during takeoff are much more dangerous. They happen fast and you have lots of fuel and weight. But they're also much more rare, thankfully.
This is why, traditionally, men walked on the street side and women on the building side: because of the splatter-up from horse crap.
Once I was riding cross-country (by which I mean no road or trail) with my dad, through heavy bracken, and when we got out and into a meadow we both had cuts all over our legs. We came down a fast descent and found out it was a cattle pasture. A fresh cattle pasture. Nothing like using up all your water washing fresh cow plop out of open wounds on your legs.
The point I learned to bunnyhop was actually a bike race in Oregon. We were coming around a hard corner and I hit a banana slug with my rear wheel. Those things are very nearly frictionless. Rear end came around, guy behind me piled into me, and we both went off a sort of cliff and into a muddy creek. That banana slug had an okay revenge for its death.
I was thinking about poo on roads the other day, actually: around here, Canada Geese are the main antagonist. I realized that it's like playing Space Invaders: you have this narrow band of pavement coming towards you and you have to dodge all the static debris (goose droppings: nasty, nasty stuff, very slick) and sometimes moving stuff (the geese themselves) that's sometimes just moving randomly (avoiding you) and sometimes actively trying to hit into you (when they have babies they get hostile and come at you in a hurry -- until they realize you're going 50k and aren't going to stop.)
I knew a guy who was so mad at Canada Goose poo in his yard he started luring the geese into his garage and killing them. (and eating them.) Not as easy with horses, though. People complain more.
>Just mention horses, public roads and an affinity for cycling (me, not the horses).
Learn to bunny-hop. I admit the rightness of your claim, but they're not going to change. And a good bunny-hop technique can save your neck when a car decides to lurch into your space because the driver's an idiot or asleep or having an aneurism and you need to cross a curb without breaking your collarbone. Ditto dealing with people who throw crap at you and you don't want to hit it, unexpected potholes, grates, and even crossing railroad tracks at an angle.
It's a fairly easy skill to learn. The easy way relies on toeclips or clipless pedals. The cool way, you can do on bmx pedals but it's harder to learn. Look up Ryan Leech on YouTube: he has some instructional videos on this.
Sorry about the lag: vacation.
We never got the rabbits to stop chewing on cables.
So to prevent dead refrigerators or computers, we used:
close supervision
lifting all the power cords up over a foot off the floor
running anything critical in flexible metal conduit.
And of those, I'd say only the last one was truly effective. With two rabbits, you just can't do effective supervision, realistically speaking, and even though mine are really tiny rabbits, like a kilo a piece, they can still stand up a lot taller than you'd expect to get to stuff they want to eat.
I have/had a similar set of problems.
1. I've owned pet rats and know what they can eat.
2. I currently own house rabbits.
3. I have a recurring mouse problem.
So lemme sum up. A rabbit can eat through a 14 gauge stranded copper cord of the sort you'd use for your refrigerator. Guess how I figured that out? Since it's starting at one side it doesn't ever cross both the live and neutral, so it doesn't get electrocuted. It can eat every cord off the back of a computer in under three minutes. Guess how I figured that out? A rat doesn't have quite the toothy abilities of a rabbit but it's fairly close. They can certainly cut through thin copper.
Neither the rabbits nor the rats -- nor my dog -- have been bothered by sprays intended to keep animals from digging/chewing on things. The super hot pepper-derived stuff stopped the rabbits but not the rats, and my dog loves the stuff. The sour/bitter stuff didn't slow any of them down even slightly.
Plastic split conduit doesn't even slow them down. Even when soaked in bitter or hot do-not-chew stuff.
Rats can chew through the side of a lead pipe and crawl through a hole the size of a US quarter. I don't have evidence that they can chew through copper pipe but I wouldn't be surprised.
Reducing food doesn't work. Once they're established, you can't keep the place clean enough. I have no idea how wild mice manage to find nutrition but they do. We keep all our food in sealed containers and vacuum and roomba every other day, and neither the dog nor the sometimes cat deter the mice in the slightest. The mice do, however, drive the dog and the cat completely insane, so if you want to have your predator madly clawing at the wall where it can either hear or smell a mouse, go for it. Both dogs and clawed cats can dig through standard drywall, and then you have a repair to do. (and they remember it and keep trying. Pitbulls are very, very retentive dogs and she'll dig through 12mm thick plywood to get to where she remembers a mouse or rat or squirrel to have hidden, once, six months ago.)
hate to say it but d-con and other awful poisons are probably the best way to go, as far as eradication, and flexible conduit to protect the lines you can't easily replace.
As I said elsewhere, glue traps are probably more evil than poison, and oftentimes live traps are as well, because you don't check them often enough and the animal dies of dehydration. And if you're really lucky the animal will manage to drag the glue trap into a place you can't get to and if you're young and still have good ears you can hear its little high-pitched screams for a couple days before it does die.
My girlfriend's landlord put in glue traps.
We started hearing this weird high-pitched noise, like a flyback transformer in a CRT going wonky.
It was mice screaming because they'd been stuck to the glue trap for days and were starving to death.
If you want to get them loose without killing them don't pull them. A: you pull parts off and B: they're still covered in glue so they just stick to debris and fluff and leaves and whatever else. Vegetable oil will get them loose. Just, y'know, don't use a lot or you end up with drowned oily mice.