These Phillips bulbs test well in heat. 200bulbs, 7000hours, 45C / 113F ambient, no failures. Stress tests to 136C / 278F (with humidity, 50g vibration, 258V square-wave voltage) no failures in 9 bulbs over 2 hours (and over 24 hours of similar but less extreme previous abuse)
Looking at the stress tests, these were unbelievably extreme - over 53 gravities of continuous vibration, voltage driving with square waves at 258V (more than double the rated voltage), while simultaneously running them at an ambient temperature of 137C. They also did the above with repeated cycling between +130C and -70C (40min of each, 10min ramp). All the CFL bulbs failed under those conditions, none of the Phillips LED bulbs failed. Humidity was also switched back and forth, with max humidity during the hot cycle. (So there is still some slight chance condensation might kill the bulbs.) In 10 stages of earlier stress testing, sine wave voltages with bad power factors were also used, as well as brownout voltages and other combinations of less extreme conditions.
Six of the starting group of nine CFL bulbs failed at least once at milder conditions, 3 replacements also failed, for a total of nine CFL failures in a starting group of nine bulbs. That was by stage 8: 96C -51C temperature cycle, 28g vibration, 54V-186V.
It's unbelievable how tough these Phillips bulbs are.
$60 is not a bad price for something that is practically impossible to kill, and may well last 35 years or more, 24hours a day before it loses 10% of its output.
Yep, these Phillips bulbs are god and reliable. Swiped from the comments on TFA:
The actual news, from the lightingprize.org 60-watt replacement lab testing page: Result: average of 200 bulbs 910 lumens 9.7 Watts 93.4 lumens/W 2727K color temperature 93 CRI
"With 95 percent confidence, lumen maintenance is predicted to be 99.3 percent at 25,000 hours." (7000 hours actual testing, 200 bulbs)
Color maintenance is nearly 7 times better than the prize requirement.
"stress testing consisted of a simultaneous combination of electrical, thermal, vibration, and humidity stresses which were increased over 14 stress levels. Tests were benchmarked against good-quality, commonly available 60W replacement compact fluorescent (CFL) lamps. Throughout the testing, photometric performance was conducted to assess any changes in performance as well as failures." ALL the CFLs failed. NONE of the Phillips LED lamps failed. (!)
Right, I was going to mention this. Hot resistors can work, too. Anyway, one can remove hidden order from the signal by using multiple diode sources and XORing them together, then running it through a "whitening" process, usually a bit-shift register with specific patterns of feedback taps (a pseudorandom number generator, PNR) which gives the pseudorandom part a repeat period of greater than the age of the universe, while its seed changes thousands or millions of times a second.
This is more than good enough. Possible attacks will be on reading the bitstream, either directly or through side channels such as power consumption or EM emissions, rather than directly on any lack of randomness. People too often get paranoid about one specific area while ignoring other areas that are easier for attackers. It's like insisting on shuffling the cards a hundred times while not noticing that you're doing it on a glass table with a high-speed camera mounted underneath.
With even a small FPGA and a few off-chip Zeners one should be able to get very high data rates, though the applications for such devices are not obvious. Usually real random numbers are needed only for the seeding of PNRs, or occasionally (and mostly historically) for one-time pads. (You need a secure channel for the distribution of the one-time pad, which is a chicken and egg problem. If it's secure enough for distribution of the pad, it's likely good enough for the message itself.)
Getting your random numbers off some web service, no matter how fancy its RNG, seems like a huge step backwards, requiring trust of the service and perhaps all the network nodes in between.
Well, how are programmers treated in the workplace In the culture at large? Do they have lots of respect, time off, money compared to other occupations with similar difficulty and requirements? Is the profession relatively secure against its jobs being shipped abroad or filled with H1Bs?
No, the culture in the US treats programmers like interchangeable nerds that can be replaced with foreigners making near minimum wage, who can be worked 80 hours per week until they become unemployable at 45. Yet few professions take more talent and effort. Unless programming itself is more attractive to you than anything else in the whole world - including family, friends, sex, money, and sleep - it's really a dumb thing to study.
I didn't link to any article. If you want to support your assertions you need to link to whatever it is that specifically supports them. I'm not going to comb through hundreds of posts and their links again to try and figure out what you were referring to.
Further, the supposed utility of Monsanto's patented seeds is their ability to survive being sprayed with Roundup. All the self-replicating bits of the seed were pre-existing and unpatentable. If the grower does not spray the crops with Roundup, the patented part has no utility to that grower and the patent should be held unenforceable against such defendants. Further, it is a big no-no to try to extend the coverage of a patent to cover things which are not patented and which have pre-existing uses or uses unrelated to the patented subject matter, which in this case means the rest of the plant. Monsanto attempts to do this with onerous licenses, but these do not apply to growers who have not agreed to such contracts. Monsanto tries to confuse its contract claims with its patent rights, but I hope they will finally get slapped down by the courts.
Yeah, for servers, it can be good, though I found the Sun/Naviscore stuff nicer for big network management. (Big, as in: "don't click that - you'll take down Florida". "But I suppose there's a downside?"
"Hmm, I see what you mean, but yes, for YOU there would be a downside".).
You have it backwards. Individuals get billed far higher amounts than insurance companies. The systemic effect of insurance, however, is to drive up costs - list price and demand are decoupled, and all the insurance paperwork, bureaucracy and profit have to be paid.
Trillian:"`You know they've reintroduced the death penalty for insurance company directors?' `Really?' said Arthur. `No I didn't. For what offense?' Trillian frowned. `What do you mean, offense?' `I see.'"
Eliminating malpractice claims is a bad idea, a real "no-brainer". Those who cause harm cannot be insulated from taking responsibility for their actions without creating perverse incentives that will kill hundreds of thousands of people over the course of a few decades. We need to go the other direction and eliminate immunity for government apparatchiks and insurance companies who presently have no incentive not to abuse their power. Malpractice suits are not a significant driver of costs, despite the mewling of insurance companies and their paid political hacks.
Your comment about "the problem of incompetent doctors relocating to new states and setting up practices" also shows that you know nothing about the safeguards already in place. A doctor cannot get a job in any US state without submitting references from every place he has ever practiced for a single month, plus much much more. For senior doctors who have worked in 20+ hospitals, applications for a new job can run to hundreds of pages.
You proposal for "the complete elimination of the emergency room unfunded mandate" would be a good idea if you mean that it would be a good idea to fund emergency rooms. If you mean to predicate emergency treatment on the ability to pay, that would be loathsome, unethical, criminal and a good argument for considering you an enemy of humanity.
Medicare also pays less per procedure than private insurance. Medicare has much lower administrative costs than private insurance. Medicare "fraud" is usually nothing of the kind, but rather failure to fill out paperwork to arbitrary and labyrinthine standards. Even failing to sufficiently screw indigent patients who aren't covered by government programs is considered "fraud". Since anyone can sue anyone for this "fraud" (qui tam bounties), there are many frivolous yet successful lawsuits, driving up costs.
Even so, the cost to benefit of Medicare is much higher than any US private insurance.
I propose a market-based solution for discussion based on our current medical market. I'll agree to read your septic drivel and bill you whatever I feel like after the fact, plus a markup for guild dues, licensure, administration, overhead, overhead on overhead, profits on overhead, bad debts, insurance against me not giving a fuck, plus three more layers of profit. If you don't want to pay that, you are free to declare bankruptcy, fuck off and die. Since I have already read your post, I believe you owe me $113,472.43. (Though if you had been responsible enough to pay protection mo- I mean insurance, we would accept $1.14 from Blue Crust/ Blue Peel to cover the bill.)
Welcome to the basic model of capitalism. It works in "health care", why not on Slashdot?
Back in the mid-90s I loaded Linux (Yggdrasil or Slackware, IIRC) on a 40MHz 386. I got Xwindows running in 8MB of RAM, which is half what the guy in the article needed. It was pretty useless.
Every couple of years I try Linux again, and it never seems to get any better - bad documentation, crappy programming, 20 different alternative programs for anything one might want to do, none of which works right. It's gotten prettier over the years, but ever more bloated and irritating. If you want to actually get some work done on a desktop or laptop instead of screwing around forever with drivers and configuration and sadistically incompetent interfaces, Linux is not competitive.
And that's the good stuff they keep behind the counter at MSDN. The actual locally-stored help files are generally worth fuck-all. I can't remember the last time the MS help file told me something that wasn't obvious by inspection - everything they write is clear, simple and utterly useless.
Not that linux is effectively much better - imagine you had a thinko and accidentally typed your encrypted volume password at the command prompt - now try to find how to delete (let alone secure delete) a specific entry in your bash command history in the bash man page and you'll appreciate a little better searching and organization. Config file settings for most things are also woefully documented. Actually, linux documentation pretty much sucks. Even the best books really don't help with the fundamental cruftiness of the whole thing. You can maybe do anything, given enough time and experimentation, but even the people doing the documentation aren't willing to undergo the pain of actually figuring out how to use most of the software's capabilities.
Slashdot Mirror
These Phillips bulbs test well in heat. 200bulbs, 7000hours, 45C / 113F ambient, no failures. Stress tests to 136C / 278F (with humidity, 50g vibration, 258V square-wave voltage) no failures in 9 bulbs over 2 hours (and over 24 hours of similar but less extreme previous abuse)
Looking at the stress tests, these were unbelievably extreme - over 53 gravities of continuous vibration, voltage driving with square waves at 258V (more than double the rated voltage), while simultaneously running them at an ambient temperature of 137C. They also did the above with repeated cycling between +130C and -70C (40min of each, 10min ramp). All the CFL bulbs failed under those conditions, none of the Phillips LED bulbs failed. Humidity was also switched back and forth, with max humidity during the hot cycle. (So there is still some slight chance condensation might kill the bulbs.) In 10 stages of earlier stress testing, sine wave voltages with bad power factors were also used, as well as brownout voltages and other combinations of less extreme conditions.
Six of the starting group of nine CFL bulbs failed at least once at milder conditions, 3 replacements also failed, for a total of nine CFL failures in a starting group of nine bulbs. That was by stage 8: 96C -51C temperature cycle, 28g vibration, 54V-186V.
It's unbelievable how tough these Phillips bulbs are.
$60 is not a bad price for something that is practically impossible to kill, and may well last 35 years or more, 24hours a day before it loses 10% of its output.
Yep, these Phillips bulbs are god and reliable.
Swiped from the comments on TFA:
Insightful?!
I thought he was doing a parody of Comic Book Guy!
At first I thought you said "raptored". Never mind.
Right, I was going to mention this. Hot resistors can work, too. Anyway, one can remove hidden order from the signal by using multiple diode sources and XORing them together, then running it through a "whitening" process, usually a bit-shift register with specific patterns of feedback taps (a pseudorandom number generator, PNR) which gives the pseudorandom part a repeat period of greater than the age of the universe, while its seed changes thousands or millions of times a second.
This is more than good enough. Possible attacks will be on reading the bitstream, either directly or through side channels such as power consumption or EM emissions, rather than directly on any lack of randomness. People too often get paranoid about one specific area while ignoring other areas that are easier for attackers. It's like insisting on shuffling the cards a hundred times while not noticing that you're doing it on a glass table with a high-speed camera mounted underneath.
With even a small FPGA and a few off-chip Zeners one should be able to get very high data rates, though the applications for such devices are not obvious. Usually real random numbers are needed only for the seeding of PNRs, or occasionally (and mostly historically) for one-time pads. (You need a secure channel for the distribution of the one-time pad, which is a chicken and egg problem. If it's secure enough for distribution of the pad, it's likely good enough for the message itself.)
Getting your random numbers off some web service, no matter how fancy its RNG, seems like a huge step backwards, requiring trust of the service and perhaps all the network nodes in between.
Excellent post.
Well, how are programmers treated in the workplace In the culture at large? Do they have lots of respect, time off, money compared to other occupations with similar difficulty and requirements? Is the profession relatively secure against its jobs being shipped abroad or filled with H1Bs?
No, the culture in the US treats programmers like interchangeable nerds that can be replaced with foreigners making near minimum wage, who can be worked 80 hours per week until they become unemployable at 45. Yet few professions take more talent and effort. Unless programming itself is more attractive to you than anything else in the whole world - including family, friends, sex, money, and sleep - it's really a dumb thing to study.
Maybe women are just being smart.
Sounds like Hollywood / Big Pharma accounting. Those expenses are padded to reduce taxes.
USA net assets / yearly income ~= [5,20] depending on what, how you count.
Assets are not the same as income, but about an order of magnitude larger. That is how a smaller percentage tax on assets would be needed.
I didn't link to any article. If you want to support your assertions you need to link to whatever it is that specifically supports them. I'm not going to comb through hundreds of posts and their links again to try and figure out what you were referring to.
Further, the supposed utility of Monsanto's patented seeds is their ability to survive being sprayed with Roundup. All the self-replicating bits of the seed were pre-existing and unpatentable. If the grower does not spray the crops with Roundup, the patented part has no utility to that grower and the patent should be held unenforceable against such defendants. Further, it is a big no-no to try to extend the coverage of a patent to cover things which are not patented and which have pre-existing uses or uses unrelated to the patented subject matter, which in this case means the rest of the plant. Monsanto attempts to do this with onerous licenses, but these do not apply to growers who have not agreed to such contracts. Monsanto tries to confuse its contract claims with its patent rights, but I hope they will finally get slapped down by the courts.
I don't see anything in that link that supports your assertion, and many things that refute it.
Yeah, for servers, it can be good, though I found the Sun/Naviscore stuff nicer for big network management.
(Big, as in: "don't click that - you'll take down Florida".
"But I suppose there's a downside?"
"Hmm, I see what you mean, but yes, for YOU there would be a downside".).
$5800!!
I pay less than that for rent!
You have it backwards. Individuals get billed far higher amounts than insurance companies. The systemic effect of insurance, however, is to drive up costs - list price and demand are decoupled, and all the insurance paperwork, bureaucracy and profit have to be paid.
Insightful.
Tyr fails at both humanity and business.
-Douglas Adams
Eliminating malpractice claims is a bad idea, a real "no-brainer". Those who cause harm cannot be insulated from taking responsibility for their actions without creating perverse incentives that will kill hundreds of thousands of people over the course of a few decades. We need to go the other direction and eliminate immunity for government apparatchiks and insurance companies who presently have no incentive not to abuse their power. Malpractice suits are not a significant driver of costs, despite the mewling of insurance companies and their paid political hacks.
Your comment about "the problem of incompetent doctors relocating to new states and setting up practices" also shows that you know nothing about the safeguards already in place. A doctor cannot get a job in any US state without submitting references from every place he has ever practiced for a single month, plus much much more. For senior doctors who have worked in 20+ hospitals, applications for a new job can run to hundreds of pages.
You proposal for "the complete elimination of the emergency room unfunded mandate" would be a good idea if you mean that it would be a good idea to fund emergency rooms. If you mean to predicate emergency treatment on the ability to pay, that would be loathsome, unethical, criminal and a good argument for considering you an enemy of humanity.
Medicare also pays less per procedure than private insurance. Medicare has much lower administrative costs than private insurance. Medicare "fraud" is usually nothing of the kind, but rather failure to fill out paperwork to arbitrary and labyrinthine standards. Even failing to sufficiently screw indigent patients who aren't covered by government programs is considered "fraud". Since anyone can sue anyone for this "fraud" (qui tam bounties), there are many frivolous yet successful lawsuits, driving up costs.
Even so, the cost to benefit of Medicare is much higher than any US private insurance.
I propose a market-based solution for discussion based on our current medical market. I'll agree to read your septic drivel and bill you whatever I feel like after the fact, plus a markup for guild dues, licensure, administration, overhead, overhead on overhead, profits on overhead, bad debts, insurance against me not giving a fuck, plus three more layers of profit. If you don't want to pay that, you are free to declare bankruptcy, fuck off and die. Since I have already read your post, I believe you owe me $113,472.43. (Though if you had been responsible enough to pay protection mo- I mean insurance, we would accept $1.14 from Blue Crust/ Blue Peel to cover the bill.)
Welcome to the basic model of capitalism. It works in "health care", why not on Slashdot?
Or maybe the other way around.
Back in the mid-90s I loaded Linux (Yggdrasil or Slackware, IIRC) on a 40MHz 386. I got Xwindows running in 8MB of RAM, which is half what the guy in the article needed. It was pretty useless.
Every couple of years I try Linux again, and it never seems to get any better - bad documentation, crappy programming, 20 different alternative programs for anything one might want to do, none of which works right. It's gotten prettier over the years, but ever more bloated and irritating. If you want to actually get some work done on a desktop or laptop instead of screwing around forever with drivers and configuration and sadistically incompetent interfaces, Linux is not competitive.
And that's the good stuff they keep behind the counter at MSDN. The actual locally-stored help files are generally worth fuck-all. I can't remember the last time the MS help file told me something that wasn't obvious by inspection - everything they write is clear, simple and utterly useless.
Not that linux is effectively much better - imagine you had a thinko and accidentally typed your encrypted volume password at the command prompt - now try to find how to delete (let alone secure delete) a specific entry in your bash command history in the bash man page and you'll appreciate a little better searching and organization. Config file settings for most things are also woefully documented. Actually, linux documentation pretty much sucks. Even the best books really don't help with the fundamental cruftiness of the whole thing. You can maybe do anything, given enough time and experimentation, but even the people doing the documentation aren't willing to undergo the pain of actually figuring out how to use most of the software's capabilities.