Given the plethora of proven connectivity options out there, I can't envision a scenario where I would chose this implementation over others. From TFA they talk about saving energy with the LED lighting system, but couldn't you by a cheaper LED lighting control system without their "value added" data transmission tech added to the cost?
Yes. However, so much of the cost of an LED system is in the LED's themselves, and so little is in the hardware that's running the driver, that adding extra functionality to the driver has marginal added cost to the overall package. Moreover, businesses and particularly government purchasing offices are *screaming* for managed light systems that they can remotely monitor and shut down per-unit. That means networking to the light, with control over whether it's on or off, is already included in such a design, so adding an ethernet port at the controller and having it send out packets is a pretty minor addition. (I work in LED driver design. I don't design them, I just test them, but the designers work down the hall, so I learn a lot about commercial lighting.)
It needs some work. First off, it can't handle temp tags at all -- and a surprising number of cars have temp tags. It's also pretty bad with mud or complicated license plate holders. He thinks it only gets about a 70% success rate. Secondly, since he needs quite a bit of info, even using a 2 megapixel cam he still needs like 300x500 pixels or thereabouts to get a good shot, which means it can't just be fixed and trying to grab the whole scene: it scans and looks for license-plate-like objects. Which means a lot of the time when he hits the asshole button it's looking the wrong direction and he doesn't get a screenshot, so he keeps hitting the button until it looks like it's pointed the right direction, giving him more stuff to sort through. But the idea's pretty cool, and I hope he gets the bugs worked out. It's related to something he's doing for work, so he's paid to work on it, *and* he has access to a lot of image-processing-specific talent and software, so he has a good tool set for getting it working.
I've a coworker who has built a cute setup for his car. He's a bright coder and we have a lot of neat hardware, so his first version was just to have a pan-and-tilt webcam scanning the parking lot and recognizing the car his boss drives, so he gets a heads-up email/sms when his boss gets to work. Since that worked pretty well, he's built (and is still playing with) what he calls assholecam. It sits on the dashboard of his car, scans nearby cars, and sticks their license plate numbers into a database. If someone is driving like an asshole he pushes a button and it saves a screenshot, and later on he goes through by hand and marks which car was being a jerk, which it stores in the database. As his database grows, he's beginning to have second, third, and fourth encounters with jerky drivers, which his little computer that does all the logging/database stuff alerts him of by beeping. Some day he'd like to add some stuff to make it less manual, but he's having trouble figuring out a good way to mark which car he's annoyed with real-time, or have it mark cars for him. Something that can dump material into a GPS could serve as a display, to show where jerk drivers are. But a good interface for input, that won't be a huge distraction to him, is a hard problem.
It's easy, with a reasonable number of people, to recreate 1870's technology. The problem once you get into 1930's tech is just the sheer number of man-hours it took to build the infrastructure required for decent development. I've read estimates that it took 50,000 man years of effort to turn Fleming's discovery of penicillin into mass-production of penicillin, and most of that wasn't basic research, where having foreknowledge saves you time, it was applications development: breeding hyperproductive strains, building bioreactors. Likewise, the tech required to build a factory capable of producing silicon with 99.9995% purity, needed for semiconductors to actually work, is hundreds of man-years of effort when you already have the equipment, and tens of man-years of effort go into building each of those pieces of equipment.
Sure we might have the knowlege, but you'd need a million people with that level of knowledge, to get to 1950's-era tech, and that's presupposing you conveniently manage to skip some horrible disease that wipes out half your population before you manage to develop antibiotics. Even getting *farming* productivity high enough so that half your people can spend their time developing equipment, rather than farming to keep from starving, is going to take several plant/harvest seasons.
BTW it's fairly easy to make muskets, and even engines from cast iron -- although smelting iron from ore is an extremely time-consuming process, and getting iron pure enough to be useful is even more difficult. Again, you run into dependency issues: good iron/steel is dependent upon a supply of high-purity oxygen, for refining the carbon out of the smelted iron, and high-purity oxygen is dependent upon having vacuum technology and fractional distillation technology, which is dependent upon having lathes and mills and a way to make vacuum seals, and so forth. I've made small engines from aluminum I've cast myself, and made gun parts and the like, and all the steps are quite easy, if you already have the tools and raw materials. I've even made a metal lathe from aluminum I've cast myself, which is getting very close to from-scratch, except for the whole refining-aluminum thing. (It would be possible in cast iron but it'd take a lot longer.) The issue is how very fractal it is, and how very much time each step takes, and when one of those steps would be 'feeding everyone' and another would be 'keeping everyone healthy' (and building, say, sewer systems so everyone doesn't die of cholera) you're already spending a decade just getting to the point where you can start building stuff beyond staying-alive-tomorrow stuff.
It's not scalpels that are expensive. It's MRI's and PET scans, and more to the point, that when you're dying, and your doctor is, in essence, working on commission, you and your doctor both are very willing to try any and every high-tech, high-cost diagnostic and treatment, to put off dying in the hopes that you'll be that 1-in-100 miracle cure. Here's a good article about this problem written by an oncology surgeon in The New Yorker a couple months ago, where he talks about how 25% of Medicare's current budget is being spent by people who are in their last couple of months of life, and that money provides an average of less than two months' delay in death.
In the last days of WWII, high-ranking German staff who didn't have access to cyanide killed themselves with nicotine injections. It was claimed that soaking the nicotine out of a cigarette into a solvent, and injecting that, wouldn't kill you, but a cigar would.
Myself, I think about chugging a liter of nitroglycerine and jumping off a building -- but the heart's reaction to the nitroglycerine would be extremely unpleasant.
You ever watch either "Harold And Maude" or "Better Off Dead"? Two funny movies about people obsessed with suicide and death.
And wrt barbituates, I've several friends that have tried to kill themselves with overdosing on prescription drugs of various sorts, and they've mostly ended up very screwed up. Don't try overdosing on lithium, for instance, because you'll almost certainly live but your brain will be a disaster area for years. Likewise, having to explain obvious scars on your wrists/always wearing sweaters to cover them, is a big drag. Hence the heroin idea: nothing to explain if it doesn't work, and a minimum (under the circumstances) of upset to people if it does.
You know what's worse than a smart person who is lazy and doesn't show up on time? A dumb person that is lazy and doesn't show up on time. All of those traits he listed aren't qualities that solely belong to "smart people."
True, but also to some extent, irrelevant, because there's no question what to do then: you fire those people. The problem is people who contribute an unusually high amount of quality work, and also an unusually high amount of trouble.
If a job has a skillset for which there's more or less a gaussian distribution of skill among applicants, then you just decide whether a person is more weird than smart or more smart than weird, and make your decision that way. The problem is for jobs where there isn't a gaussian distribution: where most people can do a passable job, but a few people can do a spectacular job, maybe an order of magnitude better than most people. Then, it's not just a weird-vs-smart balance, because the people aren't actually replaceable.
The other problem is when a job is seen as being a rock-star-requirement job, and it isn't actually. And of course people have an incentive to portray jobs as being rock-star-type jobs. So then it gets tricky.
Our company got a 3Dsystems V-Flash, which I'm presuming is the $10K model you saw. It works pretty well, but some of the parts have significant distortion as the plastic cures. It also had a problem with the reel mechanism: it feeds out a layer of material on a plastic tape, cures some, then feeds back in, and if a flake of hardened material gets on that tape it'll shadow out and prevent curing of everything in the rest of the part above that, which can lead to swiss cheese parts. The feed material is also breathtakingly expensive. But it sure does clean up easily and nicely. We also got a Stratasys, which I believe is what's being sold by Hewlett Packard in the US, and it had *excellent* accuracy, within 10 mils, and has a very nice appearance. The printing material's pretty cheap, too. But you have to use weird solvents to get the part out of the backing/support material.
The next model that 3D systems offers, the ProJet, seems to have vastly better accuracy and stability during cure.
We build prototype cellphone cases, speaker cases, outlet switch boxes, light bulb reflector backing structures, to make sure everything fits mechanically and looks good, and then go to plastic injection molding companies to have the production runs done. We found that a single mistake on an injection molding die cost about as much as the V-flash. And meanwhile the two gamers at work are running off about a zillion RPG miniatures on the machine.
I know that if I planed to kill myself by a posion, I would want to know quite a bit about how it worked and what to expect.
Though, I am not sure thats the one I would choose.... nicotine maybe.... or nitrous oxide... glycol tastes sweet if I remember, its why dogs sometimes die from drinking antifreeze, so seems like a good choice to slip in food or drink... so... hard to say.
Why would you want to kill yourself with poison? Go for a heroin overdose.
1. It doesn't hurt while you're dying. In fact, it feels pretty good.
2. If you don't die you don't end up messed up for life: blind, scarred, brain-injured, whatever, as happens with most poisons.
3. And, most importantly, your loved ones all say "how sad: I didn't know that idiot was a junkie" rather than spend years feeling guilty and asking themselves "oh my god is there something I could have done to prevent this happening?"
Suicide is usually a really crappy choice, but if you're going to make it, might as well at least do it with the least harm to other people, and to yourself, should you not succeed.
Frankly, as a 40-something old fart I'm not that into MF either. But young geeks sure are, hence the inclusion on the list. Devon Aoki, however, or Julie Delpy -- well, they're on MY Christmas list, and I'm not sharing.
When I was doing backups from the 24 year old Amiga 2000 the other day, it was really a chore: I pulled out the SCSI drive and stuck it in the linux box, since linux can read the amiga file system without a problem. And the Amiga can read the 3 1/4" floppy discs from the old Mac. However I did have one tricky bit: the old mp3 player, a PJRC.com model I bought like 10 years ago, wouldn't fire up. I'd intended to copy a bunch of Underworld to the hard disc and put it behind the bathroom mirror so when the Party Switch is activated, and the disco ball starts spinning, "King Of Snake" comes busting out from the walls themselves. But it wouldn't start. I haven't run it in seven years, and I was kind of sad, so I started poking at it, and found that a hard drive that hasn't spun up in that long does much better if it gets thwhacked solidly on the side. I also had to reseat one of the socketed IC's. Then it spun right up and started cranking out old Beastie Boys tunes, until I rewrote the hard drive.
A much harder problem is the Hewlett Packard 4145A Semiconductor Parameter Analyzer, which uses a 5 1/2" disc, with a non-standard hardware interface and non-standard track spacing on the disc, and there I admit I'm pretty much hosed if the disc ever fails, since it reads boot information off the disc every time it starts, and checks for an intentionally damaged track as a copy-protection scheme.
I've posted about this many times before. I work for a semiconductor design and fab company. We have fabs overseas, of course. But we also have a prototype fab at our head office, in California, and right beside it we have a military fab. Any design that anyone wants fabricated with a guarantee of security, we'll run, with their engineers involved at every step of the way, with 100% verification, if they're willing to pay enough. And apparently -- I don't know this because I don't have access to this kind of information, but I'm told so by people who do -- most all the fab we do for military/DoD stuff is done in our secure fab, with different chip numbers. Even if it's a chip we fab overseas, they redo the chip design, fabricate it, package it, do on-die and in-package testing, applications engineering, and product engineering, entirely at that one site, with the customer as involved as the customer wishes to be. Since we also do work for many, many other companies, and I know our marketing and sales people are smart and aggressively interested in attracting business, I strongly suspect that we use our capability of making chips that are as secure as the customer wants, as a sales pitch. That means anyone who is buying chips overseas is doing so because price is much more important than security, and they get what they want. Anyone who cares about security has verification options, and I'm going to assume that anyone whose job depends on keeping stuff uncompromised is buying from sources where they can verify what they're getting.
With that said, I've built a foundry based on these books, and built a small mill as per the milling machine book, and I have some recommendations: build the foundry and do a bit of casting and see if you like it, and then switch to a Reil burner running off propane: it's cleaner, faster, and less hassle.
Likewise, Gingery's designs are okay, but there are *lots* of modified versions out there, with much deeper bed sections to stiffen up the lathe or mill beds, and there are advantages to using modern angular contact bearings over his plain bronze bearing designs. (There are good things about bronze bearings as well, so it's a tradeoff, but generally on smaller machines, higher RPM's are more useful than the stability under very heavy loads that plain bearings give you.)
I think part of the problem is the cost of PCB fabrication: buying someone else's board, that's capable of handling your specific use case, is going to add about $30 per axis, or about $60 for a multi-axis-on-one-board, to the cost of the electronics, and while that might not seem like much, to people who are used to doing their own, with homemade PCB's, that's 50-70% of the price of the entire motor controller, so it seems cost-effective to roll your own. Add to that, the expense of high-power drivers and the mismatch between drivers and motors. Generally, you find a deal somewhere on what you think are some really good stepper or servo motors, and then you go to find a matching driver and controller, and there's one, reasonably priced, that doesn't quite have enough power for your steppers, and another that costs an arm and a leg and has 60% more power than you need, and you mumble and curse and build your own that is well-matched to the hardware you have. I built a couple OSMC controllers, that if you looked at them, you'd probably think it was a completely custom circuit. Maybe this guy did the same. It's trading off time for money, but a set of geckodrive controllers to do the same thing I needed would run $500, and what I built cost $50 plus my time. (Which, since PCB fab and assembly is part of my 9-to-5 job, didn't take much time.)
I'm not arguing with your overall message. However:
Despite the widespread and increasing availability of sexually explicit materials, according to national FBI Department of Justice statistics, the incidence of rape declined markedly from 1975 to 1995. This was particularly seen in the age categories 20–24 and 25–34, the people most likely to use the Internet.
Like I said, I agree with what you're saying, and particularly in the case of the reduction of child sex abuse in countries where people have a non-predatory outlet for their issues, I think that the way the US handles the situation is crazy, particularly because of its clearly religious motivation. But it's not clear that reductions in violent sexual crime are entirely or even mostly because of increased access to pornography, although I certainly think it's helped.
...are that it's proved to be a completely inappropriate way of measuring the age of a sample, particularly for older samples.
In fact for any sample over 2000 years old the errors are absolute.
So in fact, this is big, big news.
I'd be curious to see where it's been "proved" to be an inappropriate way of age measuring, since carbon-14 dating closely correlates with tree ring data out to 26,000 years back, using the INTCAL04 data group, which is internationally recognized as valid, and likewise it correlates well with deep ocean sediments, coral, cave rock formations, and other sources, all of which give similar age data to radiocarbon dating, which is currently using the INTCAL09 data for correction, that is internationally recognized as valid out to 50,000 years. So, if there's a problem with radiocarbon dating, the same problem is also affecting how fast sediments accumulate, coral grow, and stalactites form, and I've never heard of anyone suggesting anything that can affect all those, at the same time, and alter them all in a proportional manner. If you've any suggestions for something that could do that, I'd love to hear about it.
I ask, because the other day some of our corporate overlords visited my site and one said to the other, as they were walking by the door to my office, "and you'll notice this is the only site company-wide where everyone has a separate office, and we think that's why their productivity is so high" along with some other stuff. It's true: we do have very high productivity, and I think a chunk of that is being able to close the door when it's crunch time, play loud music, and not have interruptions.
I've been in the same crash as the GP. In fact, I've been in it, the last six times I've had crashes. Here's the scenario: red light, cars stopped. I stop. I have cars to the right of me. I have a concrete lane divider to the left of me. I have a car ahead of me. I look up into the rear-view mirror and realize that I have 1.2 seconds before the car behind me, whose driver is talking on a cellphone and hasn't seen that traffic is stopped. So, in that 1.2 seconds I have, where I'm unable to go right, left, or forwards, what am I supposed to do to avoid this crash?
In the big crash I had I have no idea what exactly happened, since I don't remember that entire month, but according to the eyewitnesses, traffic was stopped, and I'd slowed down to a near-stop with, again, cars on either side of me, when the semi ran into the back of my car at 65 mph. Again: what do you think I could have done to avoid injury, when I had a car directly ahead of me (and in the crash, my car was smashed into hers, so answering "leave more space ahead of your car so you can move forwards" counts as a completely useless answer, since the amount of space required to do that was larger than the amount of space I had available) and cars on both sides.
I'd be really interested in knowing how I can avoid these crashes, since I've had the same exact accident four times in the last four years and I'd really like to have your amazing powers of accident-avoidance to prevent having another one. Please keep in mind the criteria: cars on one side, immediately beside me. Concrete barrier on the other, close enough there's no room for me to actually go that direction. Car immediately ahead of me. Less than 2 seconds to recognize the person behind me is not going to stop and do something.
That's a somewhat arbitrary distinction -- quadratic means a number raised to the exponent of two, right? so the class of quadratic equations is contained within the class of exponential equations. I don't think the popular usage of 'exponential' is inaccurate, just not as accurate as it could be -- but then again, almost nothing is described as accurately as it could be.
I'd like to add some stuff from my experience of working with and partially managing three big lab spaces chock-full of electrical engineers.
1. It's a well-known generalization that the more people you have sharing a space, the less care each person takes of equipment and of keeping the area tidy. Don't buy anything that's going to make you cringe when it gets dropped or has coffee spilled into it. An organizational scheme I learned from renovating a bathroom whilst living with a nice woman who had a startlingly large array of small bottles full of various weird colloidal materials was: if you want to collect a huge load of crap all over everything provide lots of open flat area, and if you don't, make your surfaces small enough to only contain what you're working on/with, and make everything else slanted. Specifically: figure out a way to mount your oscilloscopes and bench multimeters slanted upwards so people don't stack stuff on top of them, *particularly* containers full of liquid.
2. You can get a small, reasonably capable drillpress from the likes of Harbor Freight or Sears. The smallish model -- 1/2" drillbit, tilting table, working distance from the column of about 20 cm, seems to be exactly the same from any source, so get the cheapest one. Check out drills from a locked desk, because they *will* break and disappear. Double goes for taps. Chain the drillpress key to the drillpress because it's completely useless once that's gone and buying a key to match a no-name chuck is not actually very much fun.
3. While AVR and particularly Arduino are *fantastic* for their existing codebase, ST's microcontroller development kits, particularly the MSP430 LaunchPad, are truly amazing for what they deliver at the price they ask. Not as much code available, though, as for the PIC or AVR.
4. Buy test&measurement stuff off ebay. If you're doing this for people who are just dropping in, aren't paying for it, and you're not being paid to keep them happy, you can buy half-broken stuff for chump change. I bought a 125 MHz oscilloscope for $70 the other day because one of the two channels didn't work. Still, $70 is a good price for even one channel -- and I happen to know how to fix this scope, so it was a *great* price.
5. If you have the space, provide one workspace with an ESD mat and a grounding armband, and provide one workspace with an insulated surface, clearly marked "low voltage" and "high voltage". Trying to mix them means electronics are certainly going to get fried and it's quite possible people are, too.
I'm a big fan of the pegboard-up-one-wall plan. Another trick I learned, this one from a very bright ex-girlfriend: put all the tools up on the pegboard, and get some bright pink spraypaint, and spraypaint everything. That way you have a shadow in the shape of each tool so you know what went there, and all your tools are BRIGHT PINK which reduces their vapor pressure, if you know what I mean.
It's also nice if you can provide enough space for a bench vise and have a hacksaw handy.
We're facing at least two major problems with human risk assessment and risk mitigation.
1. The security theater surrounding searches and detectors isn't primarily designed to make *us* feel safer. It's designed to make the people who are responsible for aviation safety feel safer, because if and when something *does* go wrong, they're the ones who get in enormous amounts of trouble. As such, the minimal loss of civil liberties that they personally experience because of the increasingly intrusive security procedures they come up with, is dwarfed by the reduction in their risk of being on the front page of the New York Times in a story that says "ignored all the obvious warning signs" and "didn't enact simple safety procedures" and other hindsight-is-20/20 statements from scads of scared and upset people. There is a tremendous incentive, and no downside, to the people coming up with MOAR SECURITY NOW! ideas: if they propose processes that are immoral, unethical, and unconstitutional, so what? it's not their problem, and an airplane blowing up IS their problem, so they keep pushing all the draconian stuff they can.
2. People, in general, are really bad at return-on-investment calculations for very low-probability situations. I was reading an economics book last night that talked about this -- "Sex, Drugs, and Economics" by Diane Coyle, that I think everyone should read. One of the things mentioned in there was that if you ask people about the amount they'd pay to reduce the risk of something really nasty happening from 1:1000 to 1:10,000, and then from 1:100,000 to 1:1,000,000, they'll pay very close to the same amount. That's particularly the case if the 'really nasty' thing is described graphically, like a vivid description of dying from cancer from carcinogen-contaminated groundwater.
So, we can yell all day about security theater and the Constitution, but if we're actually going to try to slow down the reduction in our civil liberties, we're going to have to come up with ways of addressing both these issues.
Slashdot Mirror
Given the plethora of proven connectivity options out there, I can't envision a scenario where I would chose this implementation over others. From TFA they talk about saving energy with the LED lighting system, but couldn't you by a cheaper LED lighting control system without their "value added" data transmission tech added to the cost?
Yes. However, so much of the cost of an LED system is in the LED's themselves, and so little is in the hardware that's running the driver, that adding extra functionality to the driver has marginal added cost to the overall package. Moreover, businesses and particularly government purchasing offices are *screaming* for managed light systems that they can remotely monitor and shut down per-unit. That means networking to the light, with control over whether it's on or off, is already included in such a design, so adding an ethernet port at the controller and having it send out packets is a pretty minor addition. (I work in LED driver design. I don't design them, I just test them, but the designers work down the hall, so I learn a lot about commercial lighting.)
It needs some work. First off, it can't handle temp tags at all -- and a surprising number of cars have temp tags. It's also pretty bad with mud or complicated license plate holders. He thinks it only gets about a 70% success rate. Secondly, since he needs quite a bit of info, even using a 2 megapixel cam he still needs like 300x500 pixels or thereabouts to get a good shot, which means it can't just be fixed and trying to grab the whole scene: it scans and looks for license-plate-like objects. Which means a lot of the time when he hits the asshole button it's looking the wrong direction and he doesn't get a screenshot, so he keeps hitting the button until it looks like it's pointed the right direction, giving him more stuff to sort through. But the idea's pretty cool, and I hope he gets the bugs worked out. It's related to something he's doing for work, so he's paid to work on it, *and* he has access to a lot of image-processing-specific talent and software, so he has a good tool set for getting it working.
I've a coworker who has built a cute setup for his car. He's a bright coder and we have a lot of neat hardware, so his first version was just to have a pan-and-tilt webcam scanning the parking lot and recognizing the car his boss drives, so he gets a heads-up email/sms when his boss gets to work. Since that worked pretty well, he's built (and is still playing with) what he calls assholecam. It sits on the dashboard of his car, scans nearby cars, and sticks their license plate numbers into a database. If someone is driving like an asshole he pushes a button and it saves a screenshot, and later on he goes through by hand and marks which car was being a jerk, which it stores in the database. As his database grows, he's beginning to have second, third, and fourth encounters with jerky drivers, which his little computer that does all the logging/database stuff alerts him of by beeping. Some day he'd like to add some stuff to make it less manual, but he's having trouble figuring out a good way to mark which car he's annoyed with real-time, or have it mark cars for him. Something that can dump material into a GPS could serve as a display, to show where jerk drivers are. But a good interface for input, that won't be a huge distraction to him, is a hard problem.
Sure we might have the knowlege, but you'd need a million people with that level of knowledge, to get to 1950's-era tech, and that's presupposing you conveniently manage to skip some horrible disease that wipes out half your population before you manage to develop antibiotics. Even getting *farming* productivity high enough so that half your people can spend their time developing equipment, rather than farming to keep from starving, is going to take several plant/harvest seasons.
BTW it's fairly easy to make muskets, and even engines from cast iron -- although smelting iron from ore is an extremely time-consuming process, and getting iron pure enough to be useful is even more difficult. Again, you run into dependency issues: good iron/steel is dependent upon a supply of high-purity oxygen, for refining the carbon out of the smelted iron, and high-purity oxygen is dependent upon having vacuum technology and fractional distillation technology, which is dependent upon having lathes and mills and a way to make vacuum seals, and so forth. I've made small engines from aluminum I've cast myself, and made gun parts and the like, and all the steps are quite easy, if you already have the tools and raw materials. I've even made a metal lathe from aluminum I've cast myself, which is getting very close to from-scratch, except for the whole refining-aluminum thing. (It would be possible in cast iron but it'd take a lot longer.) The issue is how very fractal it is, and how very much time each step takes, and when one of those steps would be 'feeding everyone' and another would be 'keeping everyone healthy' (and building, say, sewer systems so everyone doesn't die of cholera) you're already spending a decade just getting to the point where you can start building stuff beyond staying-alive-tomorrow stuff.
It's not scalpels that are expensive. It's MRI's and PET scans, and more to the point, that when you're dying, and your doctor is, in essence, working on commission, you and your doctor both are very willing to try any and every high-tech, high-cost diagnostic and treatment, to put off dying in the hopes that you'll be that 1-in-100 miracle cure. Here's a good article about this problem written by an oncology surgeon in The New Yorker a couple months ago, where he talks about how 25% of Medicare's current budget is being spent by people who are in their last couple of months of life, and that money provides an average of less than two months' delay in death.
Myself, I think about chugging a liter of nitroglycerine and jumping off a building -- but the heart's reaction to the nitroglycerine would be extremely unpleasant.
You ever watch either "Harold And Maude" or "Better Off Dead"? Two funny movies about people obsessed with suicide and death.
And wrt barbituates, I've several friends that have tried to kill themselves with overdosing on prescription drugs of various sorts, and they've mostly ended up very screwed up. Don't try overdosing on lithium, for instance, because you'll almost certainly live but your brain will be a disaster area for years. Likewise, having to explain obvious scars on your wrists/always wearing sweaters to cover them, is a big drag. Hence the heroin idea: nothing to explain if it doesn't work, and a minimum (under the circumstances) of upset to people if it does.
You know what's worse than a smart person who is lazy and doesn't show up on time? A dumb person that is lazy and doesn't show up on time. All of those traits he listed aren't qualities that solely belong to "smart people."
True, but also to some extent, irrelevant, because there's no question what to do then: you fire those people. The problem is people who contribute an unusually high amount of quality work, and also an unusually high amount of trouble.
If a job has a skillset for which there's more or less a gaussian distribution of skill among applicants, then you just decide whether a person is more weird than smart or more smart than weird, and make your decision that way. The problem is for jobs where there isn't a gaussian distribution: where most people can do a passable job, but a few people can do a spectacular job, maybe an order of magnitude better than most people. Then, it's not just a weird-vs-smart balance, because the people aren't actually replaceable.
The other problem is when a job is seen as being a rock-star-requirement job, and it isn't actually. And of course people have an incentive to portray jobs as being rock-star-type jobs. So then it gets tricky.
The next model that 3D systems offers, the ProJet, seems to have vastly better accuracy and stability during cure.
We build prototype cellphone cases, speaker cases, outlet switch boxes, light bulb reflector backing structures, to make sure everything fits mechanically and looks good, and then go to plastic injection molding companies to have the production runs done. We found that a single mistake on an injection molding die cost about as much as the V-flash. And meanwhile the two gamers at work are running off about a zillion RPG miniatures on the machine.
I know that if I planed to kill myself by a posion, I would want to know quite a bit about how it worked and what to expect.
Though, I am not sure thats the one I would choose.... nicotine maybe.... or nitrous oxide... glycol tastes sweet if I remember, its why dogs sometimes die from drinking antifreeze, so seems like a good choice to slip in food or drink... so... hard to say.
Why would you want to kill yourself with poison? Go for a heroin overdose.
1. It doesn't hurt while you're dying. In fact, it feels pretty good.
2. If you don't die you don't end up messed up for life: blind, scarred, brain-injured, whatever, as happens with most poisons.
3. And, most importantly, your loved ones all say "how sad: I didn't know that idiot was a junkie" rather than spend years feeling guilty and asking themselves "oh my god is there something I could have done to prevent this happening?"
Suicide is usually a really crappy choice, but if you're going to make it, might as well at least do it with the least harm to other people, and to yourself, should you not succeed.
Frankly, as a 40-something old fart I'm not that into MF either. But young geeks sure are, hence the inclusion on the list. Devon Aoki, however, or Julie Delpy -- well, they're on MY Christmas list, and I'm not sharing.
I of course meant 5 1/4" and 3 1/2" discs, but it's been so long since I've used such things I reversed them.
A much harder problem is the Hewlett Packard 4145A Semiconductor Parameter Analyzer, which uses a 5 1/2" disc, with a non-standard hardware interface and non-standard track spacing on the disc, and there I admit I'm pretty much hosed if the disc ever fails, since it reads boot information off the disc every time it starts, and checks for an intentionally damaged track as a copy-protection scheme.
1. Space shuttle.
2. F22.
3. Megan Fox.
4. Nuclear-powered aircraft carrier.
5. Cray X1E
6. Having Woz as your on-staff technical advisor.
7. A copy of the NIST F1 atomic clock.
8. A gigawatt laser.
9. All the digits of pi
10. Your own website that's as popular as /.
I've posted about this many times before. I work for a semiconductor design and fab company. We have fabs overseas, of course. But we also have a prototype fab at our head office, in California, and right beside it we have a military fab. Any design that anyone wants fabricated with a guarantee of security, we'll run, with their engineers involved at every step of the way, with 100% verification, if they're willing to pay enough. And apparently -- I don't know this because I don't have access to this kind of information, but I'm told so by people who do -- most all the fab we do for military/DoD stuff is done in our secure fab, with different chip numbers. Even if it's a chip we fab overseas, they redo the chip design, fabricate it, package it, do on-die and in-package testing, applications engineering, and product engineering, entirely at that one site, with the customer as involved as the customer wishes to be. Since we also do work for many, many other companies, and I know our marketing and sales people are smart and aggressively interested in attracting business, I strongly suspect that we use our capability of making chips that are as secure as the customer wants, as a sales pitch. That means anyone who is buying chips overseas is doing so because price is much more important than security, and they get what they want. Anyone who cares about security has verification options, and I'm going to assume that anyone whose job depends on keeping stuff uncompromised is buying from sources where they can verify what they're getting.
It's okay, tomorrow you won't have said it. Or yesterday, I'm not sure.
Dave Gingery Build A Metalworking Shop From Scrap Index
The most relevant book being the charcoal foundry book.
With that said, I've built a foundry based on these books, and built a small mill as per the milling machine book, and I have some recommendations: build the foundry and do a bit of casting and see if you like it, and then switch to a Reil burner running off propane: it's cleaner, faster, and less hassle.
Likewise, Gingery's designs are okay, but there are *lots* of modified versions out there, with much deeper bed sections to stiffen up the lathe or mill beds, and there are advantages to using modern angular contact bearings over his plain bronze bearing designs. (There are good things about bronze bearings as well, so it's a tradeoff, but generally on smaller machines, higher RPM's are more useful than the stability under very heavy loads that plain bearings give you.)
I think part of the problem is the cost of PCB fabrication: buying someone else's board, that's capable of handling your specific use case, is going to add about $30 per axis, or about $60 for a multi-axis-on-one-board, to the cost of the electronics, and while that might not seem like much, to people who are used to doing their own, with homemade PCB's, that's 50-70% of the price of the entire motor controller, so it seems cost-effective to roll your own. Add to that, the expense of high-power drivers and the mismatch between drivers and motors. Generally, you find a deal somewhere on what you think are some really good stepper or servo motors, and then you go to find a matching driver and controller, and there's one, reasonably priced, that doesn't quite have enough power for your steppers, and another that costs an arm and a leg and has 60% more power than you need, and you mumble and curse and build your own that is well-matched to the hardware you have. I built a couple OSMC controllers, that if you looked at them, you'd probably think it was a completely custom circuit. Maybe this guy did the same. It's trading off time for money, but a set of geckodrive controllers to do the same thing I needed would run $500, and what I built cost $50 plus my time. (Which, since PCB fab and assembly is part of my 9-to-5 job, didn't take much time.)
Despite the widespread and increasing availability of sexually explicit materials, according to national FBI Department of Justice statistics, the incidence of rape declined markedly from 1975 to 1995. This was particularly seen in the age categories 20–24 and 25–34, the people most likely to use the Internet.
Leavitt and Dubner, in Freakonomics, reiterate the previous claims that the widespread availability of abortion has directly resulted in a massive reduction in the rate of crime of young adults, and also the three strikes law which greatly increases incarceration rates for repeat offenders, both of which have likely been major drivers of the reduction in violent crime we've seen over the last 20 or so years.
Like I said, I agree with what you're saying, and particularly in the case of the reduction of child sex abuse in countries where people have a non-predatory outlet for their issues, I think that the way the US handles the situation is crazy, particularly because of its clearly religious motivation. But it's not clear that reductions in violent sexual crime are entirely or even mostly because of increased access to pornography, although I certainly think it's helped.
...are that it's proved to be a completely inappropriate way of measuring the age of a sample, particularly for older samples.
In fact for any sample over 2000 years old the errors are absolute.
So in fact, this is big, big news.
I'd be curious to see where it's been "proved" to be an inappropriate way of age measuring, since carbon-14 dating closely correlates with tree ring data out to 26,000 years back, using the INTCAL04 data group, which is internationally recognized as valid, and likewise it correlates well with deep ocean sediments, coral, cave rock formations, and other sources, all of which give similar age data to radiocarbon dating, which is currently using the INTCAL09 data for correction, that is internationally recognized as valid out to 50,000 years. So, if there's a problem with radiocarbon dating, the same problem is also affecting how fast sediments accumulate, coral grow, and stalactites form, and I've never heard of anyone suggesting anything that can affect all those, at the same time, and alter them all in a proportional manner. If you've any suggestions for something that could do that, I'd love to hear about it.
Next up on Brainiac:
Silicone breast, will they float or will they sink?
And are they really the best thing to grab hold of in case of office flood thanks to global warming?
Well, for the record, silicone is typically slightly denser than water so that'd make for a less-than-excellent flotation device.
And, as a strange aside, newly-filled saline implants sometimes have a bit of air in them, and they can audibly slosh, which is a bit weird.
I ask, because the other day some of our corporate overlords visited my site and one said to the other, as they were walking by the door to my office, "and you'll notice this is the only site company-wide where everyone has a separate office, and we think that's why their productivity is so high" along with some other stuff. It's true: we do have very high productivity, and I think a chunk of that is being able to close the door when it's crunch time, play loud music, and not have interruptions.
In the big crash I had I have no idea what exactly happened, since I don't remember that entire month, but according to the eyewitnesses, traffic was stopped, and I'd slowed down to a near-stop with, again, cars on either side of me, when the semi ran into the back of my car at 65 mph. Again: what do you think I could have done to avoid injury, when I had a car directly ahead of me (and in the crash, my car was smashed into hers, so answering "leave more space ahead of your car so you can move forwards" counts as a completely useless answer, since the amount of space required to do that was larger than the amount of space I had available) and cars on both sides.
I'd be really interested in knowing how I can avoid these crashes, since I've had the same exact accident four times in the last four years and I'd really like to have your amazing powers of accident-avoidance to prevent having another one. Please keep in mind the criteria: cars on one side, immediately beside me. Concrete barrier on the other, close enough there's no room for me to actually go that direction. Car immediately ahead of me. Less than 2 seconds to recognize the person behind me is not going to stop and do something.
That's a somewhat arbitrary distinction -- quadratic means a number raised to the exponent of two, right? so the class of quadratic equations is contained within the class of exponential equations. I don't think the popular usage of 'exponential' is inaccurate, just not as accurate as it could be -- but then again, almost nothing is described as accurately as it could be.
1. It's a well-known generalization that the more people you have sharing a space, the less care each person takes of equipment and of keeping the area tidy. Don't buy anything that's going to make you cringe when it gets dropped or has coffee spilled into it. An organizational scheme I learned from renovating a bathroom whilst living with a nice woman who had a startlingly large array of small bottles full of various weird colloidal materials was: if you want to collect a huge load of crap all over everything provide lots of open flat area, and if you don't, make your surfaces small enough to only contain what you're working on/with, and make everything else slanted. Specifically: figure out a way to mount your oscilloscopes and bench multimeters slanted upwards so people don't stack stuff on top of them, *particularly* containers full of liquid.
2. You can get a small, reasonably capable drillpress from the likes of Harbor Freight or Sears. The smallish model -- 1/2" drillbit, tilting table, working distance from the column of about 20 cm, seems to be exactly the same from any source, so get the cheapest one. Check out drills from a locked desk, because they *will* break and disappear. Double goes for taps. Chain the drillpress key to the drillpress because it's completely useless once that's gone and buying a key to match a no-name chuck is not actually very much fun.
3. While AVR and particularly Arduino are *fantastic* for their existing codebase, ST's microcontroller development kits, particularly the MSP430 LaunchPad, are truly amazing for what they deliver at the price they ask. Not as much code available, though, as for the PIC or AVR.
4. Buy test&measurement stuff off ebay. If you're doing this for people who are just dropping in, aren't paying for it, and you're not being paid to keep them happy, you can buy half-broken stuff for chump change. I bought a 125 MHz oscilloscope for $70 the other day because one of the two channels didn't work. Still, $70 is a good price for even one channel -- and I happen to know how to fix this scope, so it was a *great* price.
5. If you have the space, provide one workspace with an ESD mat and a grounding armband, and provide one workspace with an insulated surface, clearly marked "low voltage" and "high voltage". Trying to mix them means electronics are certainly going to get fried and it's quite possible people are, too.
I'm a big fan of the pegboard-up-one-wall plan. Another trick I learned, this one from a very bright ex-girlfriend: put all the tools up on the pegboard, and get some bright pink spraypaint, and spraypaint everything. That way you have a shadow in the shape of each tool so you know what went there, and all your tools are BRIGHT PINK which reduces their vapor pressure, if you know what I mean.
It's also nice if you can provide enough space for a bench vise and have a hacksaw handy.
1. The security theater surrounding searches and detectors isn't primarily designed to make *us* feel safer. It's designed to make the people who are responsible for aviation safety feel safer, because if and when something *does* go wrong, they're the ones who get in enormous amounts of trouble. As such, the minimal loss of civil liberties that they personally experience because of the increasingly intrusive security procedures they come up with, is dwarfed by the reduction in their risk of being on the front page of the New York Times in a story that says "ignored all the obvious warning signs" and "didn't enact simple safety procedures" and other hindsight-is-20/20 statements from scads of scared and upset people. There is a tremendous incentive, and no downside, to the people coming up with MOAR SECURITY NOW! ideas: if they propose processes that are immoral, unethical, and unconstitutional, so what? it's not their problem, and an airplane blowing up IS their problem, so they keep pushing all the draconian stuff they can.
2. People, in general, are really bad at return-on-investment calculations for very low-probability situations. I was reading an economics book last night that talked about this -- "Sex, Drugs, and Economics" by Diane Coyle, that I think everyone should read. One of the things mentioned in there was that if you ask people about the amount they'd pay to reduce the risk of something really nasty happening from 1:1000 to 1:10,000, and then from 1:100,000 to 1:1,000,000, they'll pay very close to the same amount. That's particularly the case if the 'really nasty' thing is described graphically, like a vivid description of dying from cancer from carcinogen-contaminated groundwater.
So, we can yell all day about security theater and the Constitution, but if we're actually going to try to slow down the reduction in our civil liberties, we're going to have to come up with ways of addressing both these issues.
Any ideas?