Never having been one to accept unsupported claims at face value, I just tested that assertion on a Pentim-M here, with a small C program that either calls a function to increment a counter, or directly increments the counter a number of times. I compiled with O0 to be sure gcc does not change around my code at all. Just the instructions, thanks. Funny thing? A hundred increments runs within 1% of the speed of 100 calls to a function to do the increment. And yes I unrolled those calls to isolate the cost of what I was measuring. So... rather surprisingly, the cost of these function calls is as close as doesn't matter, to exactly zero.
Except that (1) unless you look at the assembly code produced by the compiler, you don't actually know if what was generated matches your model of what was generated, and (2) this isn't a good test, since if you were to actually write code like this and allow a good optimizer to run on it, with a function call, you'd have the overhead of 100 function calls -- no matter how small -- and yet incrementing a variable 100 times in inline code will be optimzied out to a single assembler instruction on an x86 architecture (read value from memory, add constant, write new value back to the same location; that can be done in exactly one instruction). By putting the function call in the way, the compiler isn't going to be able to perform such a winning optimization.
So function calls, even with an extremely low run-time overhead, have real cost because they tend to be a barrier to optimization.
My experience is that loops are essentially free of overhead (at least with Microsoft's compiler). I have one beautiful example where the brain dead version of a loop in C code runs faster than my manually unrolled and aggressively optimized version. MS must have put a lot of effort into recognizing those brain-dead C idioms and optimizing the snot out of them. (If anyone reading this hasn't tried the MS C compiler recently, you definitely should; unlike their OS, their C compiler is damned good.)
In what way does the popular vote not count? As far as I understand, and bear in mind that I've been a US citizen for only 4 decades or so, and my only exposure is living here for that time and going through the primary, secondary, and tertiary educational system, including state-mandated civics classes, the popular vote is what determines which electors will vote and (by pledge) how the electors will vote. While there are some exceptions, and different states have different rules, the electors are understood to vote for the candidates indicated on the ballot, and are determined by, wait for it, popular vote.
Of course, you probably meant the national popular vote. And by focusing on that, you clearly have no understanding of why the electoral college was created in the first place.
Perhaps you've noticed that most presidential elections in the US are pretty close (maybe you're not old enough to have noticed, but it's true). We don't have 80% to 20% popular vote splits. A 5% margin is considered good. The 1972 landslide was barely 60-40. And yet Nixon won 49 of 50 states. (That should give you a clue right there.)
The standard story is that the electoral college was invented because at the time of the creation of the US as a nation, long-distance communication happened largely by horse. Sending results from each state to a central location to tally up meant sending a person in one form or another, to drive the horse carrying the results if nothing else, so instead of sending the votes, they sent people. Easy enough, not any slower, and it helped ensure that the votes weren't tampered with along the way.
But that's only part of the story.
The more important part is that the founding fathers were really, really smart. They saw how hard it was to organize and galvanize disparate peoples. They recognized that for leaders to be followed, they needed to be widely recognized by the larger populace as leaders. A nation, especially a younger nation, exists only because its citizens all agree it should. Broad dissent, particularly when the nation is still gaining its legs but also once it's strong, can be hugely deleterious. It leads to civil unrest and civil wars.
So, when most elections are close, barely much beyond 50-50, how do you convince the HALF of the population who voted for the losing candidate that they should give up and follow the winner? The answer, THE answer, is to arrange things so that elections are never close to 50-50. The electoral college is designed to do this, to amplify small differences, so that marginal elections become mandates. With a mandate, the winner can lead.
How does the electoral college do this? By taking the results from each state and, effectively, turning them into winner-take-all results. Not every state will vote for the nationally more popular candidate (except as was nearly true in 1972), so some states will vote for the ultimate winner, and some will vote for the ultimate loser, but by quantizing the results on a per-state basis, the small differences get amplified.
In our most recent election, Obama won the national popular vote 53-46. That's damned close to 50-50. Nearly half of the US population voted for the fellow who didn't win. They aren't happy with the results. And yet, Obama is called one of the most popular presidents ever. He has a clear mandate. Why? Because the electoral college results were 67-32, or over 2-to-1. Landslide. Mandate.
By taking the results from each state individually and turning them into winner-take-all, small differences (51-49 percent of the popular vote in a hypothetical example state like Kansas) are amplified into large differences (6-0 votes in the electoral college). And this creates a definitive result from the electoral college, and a mandate for the elected candidate.
Unfortunately, the OP forgot to include one bit of important information: are they being deployed as part of military service, or as part of a civilian effort?
While there are a few people on Slashdot who are or have been in the military (and I hope they speak up), I daresay the general Slashdot opinion will be worth about what the OP paid for it: squat. I haven't been in the service, but can imagine that there are a raft of security issues around communications back home and that they need to be done through approved channels.
For civilian deployments, however, the story is entirely different. For this, there is lots of worthwhile advice. Here's my bit...
1. I've yet to be in a town, even in remote parts of eastern Europe and the eastern Mediterranean, where there isn't some sort of internet cafe. Connectivity is available. Some intenet cafes even have headsets for Skype.
2. Cellular phone service is nearly ubiquitous. Seriously. You have to get very remote to not have some kind of mobile phone service. The US has terrible coverage compared to Europe and the Middle East. I've been on small, remote islands in the Aegean with 5 bars. And I've yet to find a country (including in the former Soviet bloc) where you can't get pay-as-you-go service that's heaploads cheaper than any US phone company's international roaming. Just make sure that your phone is (a) unlocked and (b) quad band GSM. Or buy one there.
3. Everything in the Middle East is negotiable. Everything. Negotiation and bartering is part of the culture.
From the web site: "Green Plug is the first developer of digital technology enabling real-time collaboration between electronic devices and their power sources..."
First class marketspeak, that is. Collaboration, you say, between electronic devices and power sources. In real time. Using digital technology. Do I need to worry about my devices and power sources becoming mutinous, then?
Or can I just stick with USB which seems to power almost every electronic device I carry around? I daresay I don't actually know WHERE the chargers that came with most of my devices are, since I just charge them all though USB. Green Plug has already lost its battle. The only hope they might have is to embrace and extend the USB standard.
Your university is a for profit organization. Guess from where they are getting the money to pay Microsoft for the university wide license.
Wow. It's amazing that this was moderated insightful, as it is so deeply, fundamentally wrong.
While there are a few educational institutions that are operated as for-profit commercial entities, the vast majority are non-profit. (I'm in the midst of creating a non-profit foundation myself.) That does not mean they are not interested in income ("non-profit" does not mean "zero income" and if you thought that, I have an exciting investment opportunity to talk to you about).
Where does the money come from? Tuition is an important, but relatively small component of total income. There's also interest and dividends from the endowment, alumni donations, benefactors, licensing agreements for sports and memorabilia, licensing agreements for IP, conference hosting fees, catering services, and (hugely important) grants from private foundations and the federal government. If the university includes a medical or veterinary school, then it might also have significant income from its teaching hospital. Depending on the university, there might be other significant services they offer to commercial customers as well.
Sounds like a lot of money. It is. The expenses are huge as well, primarily salaries and infrastructure but, also, travel, supplies, capital expenditures, scholarships, grants, and so forth.
Depending on the university, and the skills of its negotiating team, the site-wide license for Windows might well be a donation from Redmond.
That's right, your tuition. I hope you are using Windows, as you are paying for it in any case.
That's a specious argument. You could just as easily say that your tuition pays for, oh, the football team, so I hope you're watching every game; or, the lights in the English Department's reading room, so I hope you're going there often; or, the salary of the professor who teaches Ethnic Cultures of the Pre-Mayan Americas, so I hope you're taking that class. To think that there's a direct line between tuition and any single expenditure -- real or imagined -- is so severely simple-minded that it's troublesome. To argue that since one small part of your tuition could be used to pay for some particular offering at the university you should therefore use that offering is absurd.
It didn't succeed because Negroponte wouldn't let anyone who wanted one buy it. It's that simple. Had he done that he would have sold enough of them to get them into the field and had money to continue development and produce them faster.
So what stopped Negroponte was....Negroponte.
Uhm, sources for this, please? According to the Wikipedia entry, there's an estimated 1,000,000 units sold http://en.wikipedia.org/wiki/One_Laptop_per_Child#Summary_of_laptop_orders and according to a recent written interview with Negroponte, they're about to deploy the 1,000,000th unit http://en.wikipedia.org/wiki/One_Laptop_per_Child#Summary_of_laptop_orders --- so I fail to see where your assertion holds together. You can't take orders for a million units and be all that selective about who buys them. Through the B1G1 / G1G1 programs anyone with a valid credit card could purchase. That certainly sounds like an open door.
Interesting design parameters you've set for yourself: a voltage source with varying output voltage, a load that prefers constant current, and make the whole thing adjustable.
I hadn't heard of the BuckPucks before, but they sure do look nice. The dimmable version claims to have full range dimming, and have a pretty wide range of input voltages. You sure it won't do the trick alone without additional PWM circuitry? Or am I missing something?
A reverse-biased (heavy-duty) zener diode might do the trick for shunting extra voltage/power from your alternator at higher speeds.
PWM works fine as long as the oscillating frequency is above what the visual system can respond to (IAAVN... I am a visual neuroscientist). The maximum frequency that the visual system responds to depends greatly on a large number of parameters, including contrast levels, and individual sensitivities, but, generally, the upper limit is about 150 Hz.
If your intention is to draw attention, then 5-to-10 Hz is excellent for this.
If your intention is to make even illumuniation, then oscillating at or above 200 Hz is certain to not appear to flicker.
These days, fewer and fewer people recall the horror of using a CRT computer monitor at 60 Hz refresh. It is painful. Seizure-inducing. But increase the refresh rate on a CRT to 85 Hz and most people don't see the flickering anymore. Increase it to 160 Hz, and it is nearly undetectable even in a laboratory setting. That's for a light that essentially goes full-on to full-off.
For light sources that are modulating at less extreme levels (like most LCD monitors, where each pixel is essentially constant in light output as long as the image does not change) a much lower refresh rate is necessary. 60 Hz is just fine.
The most recent laptops that have LED illumination are, unfortunately, modulating brightness with a typically 60 Hz PWM driver. This makes it much more like a CRT because the light source is pulsing on and off. I can see my laptop flicker easily, but it's only milding annoying (120 Hz, and it would not have been an issue -- and the thing is, a 120 Hz PWM driver isn't any harder to design than a 60 Hz one).
For a bike light, where you are using the illumination to see where you are going, I would recommend at least 100 Hz, if not 200 Hz. Do you have a choice in PWM drivers, or are you designing you own?
These are going to be awesome in an office environment. Especially since the ceilings are so high and nobody likes changing the lights. But I have yet to find truly warm non-tungsten/halogen/mercury/fire/quartz/evil light for home use. I could not picture LEDs (which are basically antennas radiating a frequency that we happen to see) overtaking the other lights (heat sources that coincidentally give off visible light) in terms of color richness.
Um, incandescent bulbs are also antennas that are radiating a band of frequencies we happen to see. Same with fluorescents. LEDs tend to have narrower spectra, but not the so-called white LEDs, which operate very much like fluorescents: they excite a phosphor.
You should try a high-quality fluorescent. They exist but are not inexpensive. However, they are very very good for doing color-accurate work. Even decent daylight fluorescents can be refreshing, if not entirely color accurate.
Not really; not like an incandescant. An incandescant dimmer is simply a rheostat (variable resistor), but Light Emitting Diodes are binary - either all the way on or all the way off. To dim an LED what you do is turn it off and on faster than the eye can see. The longer the off period, the dimmer it appears.
Dimmers for LED lamps will most likely come on the market, but the dimmer that's on you wall now won't work.
Half right. The last half. Sort-of.
LEDs have a quite nice linear lumens-vs-current regime as long as you're above threshold and below saturation (and you keep them temperature stabilized). OK, it isn't quite linear, but it's pretty close. It certainly isn't binary.
However, traditionally, LEDs were engineered so that the standard current for reasonably bright indicator use was 20 mA. Why 20 mA? Because that's the standard output current that a TTL gate can push. So, with a normal TTL logic chip, you can directly drive one LED, get good indicator use, and remain within the limits imposed by the TTL specification. This led to widespread use of LEDs as indicator lights -- thus an unwitting assumption that they only operate as on/off devices.
That said, the standard dimmer circuitry for incandescent bulbs, either the very old fashioned rheostat or the modern phase-adjusted SCR trigger, will work just fine with an LED load. The problem is that LEDs, unlike incandescents, have effectively zero emission inertia: you can switch them on an off very, very fast. Even a run-of-the-mill LED will switch on or off in microseconds, max. An incandescent bulb, in contrast, takes hundreds of milliseconds to turn on or off, and so the sharp switching waveform from SCR style dimmers is no big deal: the thermal inertial of an incandescent smooths them out. An LED on that kind of a dimmer is pure torture (I've tried it) since it's 50/60 or 100/120 Hz of full-field flickering light. Not pleasant in the least. New dimmer designs will have to be made for LEDs that include some smoothing within the dimmer, or the LED bulbs themselves will have to incorporate waveform smoothing.
But, the last suggestion -- that the apparent brightness of an LED can be modulated by changing the duty cycle -- is correct. If there are any designers reading this: PLEASE MAKE CERTAIN THE MODULATING FREQUENCY IS ABOVE 200 Hz (curious? I'd be happy to explain). The current rash of 60 Hz pulse-wave modulated LED illumination everywhere is driving me nuts.
Is this one of those words that has surreptitiously entered our language like "blog" or was the title just cut-off?
It's been part of the standard lexicon of photographic professionals and amateurs, and especially in the press, for as long as I can remember; and I've been shooting (yes, "shooting", and by gum, that's another idiomatic use of a term that's specific to a field different from computer geek stuff) for three decades now. However, "photog" it one of the odd terms that gets used far more in print than in verbalized conversation, probably because it lends itself to a more compact telegraphic style in headlines.
Personally, I use "photog" mostly as an indefinite noun, when describing a collection of photographers I don't know, or a collection of photographic things of uncertain size. Like, "the photog's convention," or, "that warehouse of photog equipment," or, "where is all my photog crap?"
Also aided in establishing "photog" as an accepted term (again, in a field I'm guessing you're not familiar with) is that Domke (a manufacturer of pro-grade photographic accessories; you might not be familiar with them, but they make good stuff and any pro will know about them) sold a "Photog Vest" for many years. They might still, for all I know; I bought mine in the mid 90s.
Oh, but there I go again, "pro" is an idiosyncratic term within the photog community for a professional photographer. In a previous life, I have been a pro, complete with (profitable!) one-man shows and paid assignments, but am now just an enthusiast photog.
The entire point of this, in case it hasn't become clear: there are terms that are widely accepted within other fields that are perfectly valid; just because you aren't familiar with the field doesn't mean you can't open your mind. The amount of jargon that's in the computer geek field is staggering: would you really expect someone, say an enthusiast photographer, to know it all?
p.s. A quick google search on "photog" would have shown that it gets used in headlines all over the place. That should probably have answered your question right there, but I like to be thorough.
What the world needs ...
on
Less Is Moore
·
· Score: 2, Insightful
We more-or-less got enough computing power for most things with the introduction of the PIII 1GHz CPU. You might not agree with this, but it's at least approximately true. A computer outfitted with that processor and reasonable RAM browses the web just fine, plays MP3s, reads email, shows videos from YouTube, etc. It doesn't do everything that you might want, but it does a lot.
If we took the amazing technology that has been used to create the 3 GHz multi-core monsters with massive on-chip cache memory in a power budget of 45W or so in some cases, and applied it to a re-implementation of the lowly PIII, we'd win big. We'd get a PIII 1GHz burning a paltry few watts.
And this is precisely why chips like the Intel Atom have been so successful. Reasonable computing power for almost no electricity. We don't necessarily need just MORE-FASTER-BIGGER-STRONGER, which is the path Intel and AMD have historically put the most effort into following, we also need more efficient.
The victorious strategist only seeks battle after the victory has been won, whereas he who is destined to defeat first fights and afterwards looks for victory. - Sun Tzu
My uncle's version of this is: only place a bet when you are certain of the outcome. He never loses a bet.
He used the words "data" and "statistics" in his inaugural address in a positive tone, without being the slightest bit derisive. He said that he would, "restore science to its rightful place." There is hope for the US.
Belkin does not participate in, nor does it endorse, unethical practices like this.
Given that the first part of this official statement is prima facia a lie -- Belkin DID participate in these unethical practices, as it was an employee, a well-placed employee high up in the managerial chain, who created the reviews in question -- it is reasonable to expect that the entire apologetic statement is not even worth the electrons spent to create it.
But why does it take more than a few milliseconds to discover a device? I've never understood this. We have had CPUs that have had sub-microsecond execution cycles for DECADES now, and yet the timeouts for communicating with devices are still measured in seconds. Why?
Device discovery should take no more than a few milliseconds for an entire machine, with the possible exception of disk drives which presumably need to spin up and verify correct operating speed to report back on a self-check.
No power supply needed for each machine. This removes a major point of failure. Instead, one would need to just step down voltages to the 5 and 12 volt rails.
Um, no. Stepping down a higher voltage to a lower voltage, say 48VDC to 5VDC and 12VDC from your example, would still require a power supply. That, precisely, is what a power supply does: voltage conversion. What you're talking about is a DC/DC power supply, rather than the traditional AC/DC ones. But, when it comes down to it, all of the modern computer power supplies are switching supplies that take AC, put it through a bridge rectifier, filter it a little, and then follow that with a DC/DC converter.
By providing 48VDC instead of 120VAC to the input to a power supply, you're eliminating the bridge rectifier (with typically less than 1% losses) and the filtering capacitor (with even lower losses, in a decent design at least). That's it.
Now, converting the noisy, ripply 170VDC that results from rectifying 120VAC to a low voltage is a harder task than converting relatively clean 48VDC to the same low voltage, and starting out with a input voltage that's almost the same as the output voltage means lower current losses, especially in a switching supply. But -- and this is a huge issue -- high current DC connectors are extremely difficult to design because of the inherent electromigration/electrolysis you get with DC current that is just not present at all with AC current. Where are said connectors? At the input to the power supply.
Because there is only one 48VDC power supply for a room, it has to be held up to a lot more rigorous standards than average mains current. It has to not just provide 48VDC, but provide it under extremely heavy load without the voltage dropping by much.
Any decent power supply works over a large range of input voltages. Designing for a precise input range (say, 48 to 48.5VDC) is shortsighted. Ever notice that the laptop power supply you have is rated for 100-240VAC? So are many (not all, but many) desktop power supplies. While the hypothetical room's 48V needs to be pretty good, it does not have to be held to any higher standards as you suggest. The circumstances where it would need to be held to higher standards would be if the delivered voltage were to be used directly without any local regulation / re-provisioning. And that would be just plain poor engineering.
The most amazing forensic work I read of was the Lockerbie bombing of a TWA flight while in midair. The debris was scattered over many square miles. Yet the investigators were able to reconstruct the bomb and find the bomb's timing circuit. A chip in the timing circuit was traced to the perpetrators.
That was pretty fucking cool, I thought.
It was incredibly cool. I believe the fragment traced to the perpetrators was a small part of a circuit board that was identified to be part of a Toshiba radio.
That alone, to me, was -- and still would be -- an astonishing feat of detective work.
Then you haven't bought a good CFL. Fluorescent bulbs with excellent, very white spectra are certainly available. Look for bulbs with color index (CRI) of 98 or over. They are not, however, inexpensive.
It is a myth that fluorescent bulbs with high-quality spectra are not available.
So-called white LED spectra are produced just like the spectra of fluorescent bulbs: white LEDs are actually UV LEDs that illuminate phosphors with different inherent spectra to cover the visible range.
For a brief while, triplet R/G/B LEDs were put in one package, but those are (a) too expensive, and (b) suffer from dramatic white-point shift (color shift) as the three chips age differentially.
you really didnt think that through to much did you...
tires last an avverage of 60k miles... so that would be approxamatly 2400 gallons of gas at 25 MPG. Oregon's current gas tax is 43.4 per gallon... so you are suggesting $1041.60 per tire in taxes?!?
Frankly, it's annoying when people accuse me of not thinking things through and then don't bother expending any intellectual effort themselves. How do you think a GPS mediated tax will work? Something like each year you'll go to a gas station that has a special instrument that reads out how far the chip in your car says it has traveled. Do you think the tax bill will be small at that point? The 10 to 20k miles of wear that each car on average places on the roads will be recovered one way or another. If the GPS-based tax is to completely replace the gas tax, then the bill at that point will be, indeed quite high. The State of Oregon is not going to give up this income stream.
But instead (and this is where you should have thought a little harder) of measuring distance traveled via GPS (with the concomitant invasion of privacy) the same service station could just measure tire wear. A simple mechanical tread depth gauge that every service station already owns would suffice. If the owner changed tires in the middle of the year, the depth measured would be pro-rated to the date of purchase. If the owner didn't have the receipt, then pro-rate the wear to 2 months (screwing most, and providing an incentive to keep the receipt; there would probably need to be an abatement procedure).
Or (gasp!), we could just use the odometer and factor in the published weight of the vehicle. But that's a less accurate measurement of the actual load each vehicle places on the road. Tire wear is about as good as it can get, but there are plenty of alternatives to GPS-based measurement if a simple gas tax is thought to be no longer sufficiently accurate.
A use tax for our roads and bridges is exactly what the gas tax is supposed to be. The deterioration of a road is directly related to the number of vehicles and their weight. When engines are all about equally efficient, then taxing fuel is a good a proxy as any for road usage.
But with the advent of hybrid technologies, and the push for new fuels which might not be gasoline, there is an inequality with use vs. amount of taxes paid for that use.
Taxing the number of miles traveled, however, is not the right answer. It's not even a good answer. It is, fundamentally, a bad answer. It is an answer with cloaked malicious intent, because it seeks to monitor the movement of the citizenry rather than recover taxes proportionate to use.
Far better would be to radically reduce or eliminate the gas tax and introduce a steeper tire tax. Tire wear is a far better proxy for road wear (that is, the amount of wear that the road has, which, mostly, determines when it needs to be resurfaced), since it is proportionate to vehicle weight and number of miles traveled. Road wear is also, not surprisingly, proportionate to vehicle weight and number of cars that pass a given point (ergo, miles traveled by each car).
Just tax tires. Problem solved. No need for additional gadgetry that requires a new infrastructure for inspection and reporting. Seriously, either this legislator is on crack or is being backed by some seriously questionable money.
When dealing with a pressurized body of molten rock with entrained gasses, I don't think one could ever say it is perfectly safe.
Absolutely. I go to Santorini, Greece every year or so. Santorini is one of the best-studied volcanoes in the world. A volcanologist friend often accompanies me and enjoys reeling of stats about magma chamber releases. Force of Nature level stuff. It's not the dissolved gases that matter quite so much as the dissolved water. Magma is typically a few weight percent (superheated) water. During an eruption, when the pressure is released the water turns to steam and the whole shebang expands in volume by 3 orders of magnitude. Until the upper part of the magma chamber is emptied, the ejection column flows at supersonic speeds. You do not screw around with Mother Nature. Intentionally drill into a magma chamber? No thank you.
Wow, thanks for the good info. Here's a quote from the article that you already linked:
SourceForge also said it "intends to take aggressive steps" to accelerate the growth of its core Internet properties: Slashdot, SourceForge.net and ThinkGeek.
"These sites represent a truly unique set of Internet assets and there is more we can be doing to make them better and more engaging for our users," said Neumeister. "We are focused on aggressively pursuing our plans for each of these sites and making opportunistic investments that will enable us to reach our objectives faster."
Thanks.
Why can't CEOs and such be satisfied with slow, even growth? Why does everything have to be aggressive leveraging of opportunity? Slashdot is never, ever going to be the next NYT or Washington Post. Why not just be satisfied with a job well done and a decent, reasonable profit?
The worst part about the SSC is mentioned in the parent comment's parenthetical comment about brain drain.
When the SSC was cancelled, there was a flood of high-energy physicists who were suddenly out of work. The US lost an entire generation of talent in physics. Instead of continuing on with a remarkable collection of centers of excellence, each themselves breeding excellence, and maintain the intellectual, scientific, technical, and economic advantages that the US Government prides itself on, the (pardon me) boneheads in Congress thought it better to continue the long slog toward mediocrity.
High-energy physics no longer happens in the US (my apologies to readers at LL, LANL, Brookhaven, Fermi, Argonne, Berkeley, and so forth). It happens in Europe and will continue to do so for the forseable future.
Slashdot Mirror
Never having been one to accept unsupported claims at face value, I just tested that assertion on a Pentim-M here, with a small C program that either calls a function to increment a counter, or directly increments the counter a number of times. I compiled with O0 to be sure gcc does not change around my code at all. Just the instructions, thanks. Funny thing? A hundred increments runs within 1% of the speed of 100 calls to a function to do the increment. And yes I unrolled those calls to isolate the cost of what I was measuring. So... rather surprisingly, the cost of these function calls is as close as doesn't matter, to exactly zero.
Except that (1) unless you look at the assembly code produced by the compiler, you don't actually know if what was generated matches your model of what was generated, and (2) this isn't a good test, since if you were to actually write code like this and allow a good optimizer to run on it, with a function call, you'd have the overhead of 100 function calls -- no matter how small -- and yet incrementing a variable 100 times in inline code will be optimzied out to a single assembler instruction on an x86 architecture (read value from memory, add constant, write new value back to the same location; that can be done in exactly one instruction). By putting the function call in the way, the compiler isn't going to be able to perform such a winning optimization.
So function calls, even with an extremely low run-time overhead, have real cost because they tend to be a barrier to optimization.
My experience is that loops are essentially free of overhead (at least with Microsoft's compiler). I have one beautiful example where the brain dead version of a loop in C code runs faster than my manually unrolled and aggressively optimized version. MS must have put a lot of effort into recognizing those brain-dead C idioms and optimizing the snot out of them. (If anyone reading this hasn't tried the MS C compiler recently, you definitely should; unlike their OS, their C compiler is damned good.)
In what way does the popular vote not count? As far as I understand, and bear in mind that I've been a US citizen for only 4 decades or so, and my only exposure is living here for that time and going through the primary, secondary, and tertiary educational system, including state-mandated civics classes, the popular vote is what determines which electors will vote and (by pledge) how the electors will vote. While there are some exceptions, and different states have different rules, the electors are understood to vote for the candidates indicated on the ballot, and are determined by, wait for it, popular vote.
Of course, you probably meant the national popular vote. And by focusing on that, you clearly have no understanding of why the electoral college was created in the first place.
Perhaps you've noticed that most presidential elections in the US are pretty close (maybe you're not old enough to have noticed, but it's true). We don't have 80% to 20% popular vote splits. A 5% margin is considered good. The 1972 landslide was barely 60-40. And yet Nixon won 49 of 50 states. (That should give you a clue right there.)
The standard story is that the electoral college was invented because at the time of the creation of the US as a nation, long-distance communication happened largely by horse. Sending results from each state to a central location to tally up meant sending a person in one form or another, to drive the horse carrying the results if nothing else, so instead of sending the votes, they sent people. Easy enough, not any slower, and it helped ensure that the votes weren't tampered with along the way.
But that's only part of the story.
The more important part is that the founding fathers were really, really smart. They saw how hard it was to organize and galvanize disparate peoples. They recognized that for leaders to be followed, they needed to be widely recognized by the larger populace as leaders. A nation, especially a younger nation, exists only because its citizens all agree it should. Broad dissent, particularly when the nation is still gaining its legs but also once it's strong, can be hugely deleterious. It leads to civil unrest and civil wars.
So, when most elections are close, barely much beyond 50-50, how do you convince the HALF of the population who voted for the losing candidate that they should give up and follow the winner? The answer, THE answer, is to arrange things so that elections are never close to 50-50. The electoral college is designed to do this, to amplify small differences, so that marginal elections become mandates. With a mandate, the winner can lead.
How does the electoral college do this? By taking the results from each state and, effectively, turning them into winner-take-all results. Not every state will vote for the nationally more popular candidate (except as was nearly true in 1972), so some states will vote for the ultimate winner, and some will vote for the ultimate loser, but by quantizing the results on a per-state basis, the small differences get amplified.
In our most recent election, Obama won the national popular vote 53-46. That's damned close to 50-50. Nearly half of the US population voted for the fellow who didn't win. They aren't happy with the results. And yet, Obama is called one of the most popular presidents ever. He has a clear mandate. Why? Because the electoral college results were 67-32, or over 2-to-1. Landslide. Mandate.
By taking the results from each state individually and turning them into winner-take-all, small differences (51-49 percent of the popular vote in a hypothetical example state like Kansas) are amplified into large differences (6-0 votes in the electoral college). And this creates a definitive result from the electoral college, and a mandate for the elected candidate.
Unfortunately, the OP forgot to include one bit of important information: are they being deployed as part of military service, or as part of a civilian effort?
While there are a few people on Slashdot who are or have been in the military (and I hope they speak up), I daresay the general Slashdot opinion will be worth about what the OP paid for it: squat. I haven't been in the service, but can imagine that there are a raft of security issues around communications back home and that they need to be done through approved channels.
For civilian deployments, however, the story is entirely different. For this, there is lots of worthwhile advice. Here's my bit ...
1. I've yet to be in a town, even in remote parts of eastern Europe and the eastern Mediterranean, where there isn't some sort of internet cafe. Connectivity is available. Some intenet cafes even have headsets for Skype.
2. Cellular phone service is nearly ubiquitous. Seriously. You have to get very remote to not have some kind of mobile phone service. The US has terrible coverage compared to Europe and the Middle East. I've been on small, remote islands in the Aegean with 5 bars. And I've yet to find a country (including in the former Soviet bloc) where you can't get pay-as-you-go service that's heaploads cheaper than any US phone company's international roaming. Just make sure that your phone is (a) unlocked and (b) quad band GSM. Or buy one there.
3. Everything in the Middle East is negotiable. Everything. Negotiation and bartering is part of the culture.
From the web site: "Green Plug is the first developer of digital technology enabling real-time collaboration between electronic devices and their power sources..."
First class marketspeak, that is. Collaboration, you say, between electronic devices and power sources. In real time. Using digital technology. Do I need to worry about my devices and power sources becoming mutinous, then?
Or can I just stick with USB which seems to power almost every electronic device I carry around? I daresay I don't actually know WHERE the chargers that came with most of my devices are, since I just charge them all though USB. Green Plug has already lost its battle. The only hope they might have is to embrace and extend the USB standard.
Your university is a for profit organization. Guess from where they are getting the money to pay Microsoft for the university wide license.
Wow. It's amazing that this was moderated insightful, as it is so deeply, fundamentally wrong.
While there are a few educational institutions that are operated as for-profit commercial entities, the vast majority are non-profit. (I'm in the midst of creating a non-profit foundation myself.) That does not mean they are not interested in income ("non-profit" does not mean "zero income" and if you thought that, I have an exciting investment opportunity to talk to you about).
Where does the money come from? Tuition is an important, but relatively small component of total income. There's also interest and dividends from the endowment, alumni donations, benefactors, licensing agreements for sports and memorabilia, licensing agreements for IP, conference hosting fees, catering services, and (hugely important) grants from private foundations and the federal government. If the university includes a medical or veterinary school, then it might also have significant income from its teaching hospital. Depending on the university, there might be other significant services they offer to commercial customers as well.
Sounds like a lot of money. It is. The expenses are huge as well, primarily salaries and infrastructure but, also, travel, supplies, capital expenditures, scholarships, grants, and so forth.
Depending on the university, and the skills of its negotiating team, the site-wide license for Windows might well be a donation from Redmond.
That's right, your tuition. I hope you are using Windows, as you are paying for it in any case.
That's a specious argument. You could just as easily say that your tuition pays for, oh, the football team, so I hope you're watching every game; or, the lights in the English Department's reading room, so I hope you're going there often; or, the salary of the professor who teaches Ethnic Cultures of the Pre-Mayan Americas, so I hope you're taking that class. To think that there's a direct line between tuition and any single expenditure -- real or imagined -- is so severely simple-minded that it's troublesome. To argue that since one small part of your tuition could be used to pay for some particular offering at the university you should therefore use that offering is absurd.
It didn't succeed because Negroponte wouldn't let anyone who wanted one buy it. It's that simple. Had he done that he would have sold enough of them to get them into the field and had money to continue development and produce them faster.
So what stopped Negroponte was....Negroponte.
Uhm, sources for this, please? According to the Wikipedia entry, there's an estimated 1,000,000 units sold http://en.wikipedia.org/wiki/One_Laptop_per_Child#Summary_of_laptop_orders and according to a recent written interview with Negroponte, they're about to deploy the 1,000,000th unit http://en.wikipedia.org/wiki/One_Laptop_per_Child#Summary_of_laptop_orders --- so I fail to see where your assertion holds together. You can't take orders for a million units and be all that selective about who buys them. Through the B1G1 / G1G1 programs anyone with a valid credit card could purchase. That certainly sounds like an open door.
Interesting design parameters you've set for yourself: a voltage source with varying output voltage, a load that prefers constant current, and make the whole thing adjustable.
I hadn't heard of the BuckPucks before, but they sure do look nice. The dimmable version claims to have full range dimming, and have a pretty wide range of input voltages. You sure it won't do the trick alone without additional PWM circuitry? Or am I missing something?
A reverse-biased (heavy-duty) zener diode might do the trick for shunting extra voltage/power from your alternator at higher speeds.
PWM works fine as long as the oscillating frequency is above what the visual system can respond to (IAAVN ... I am a visual neuroscientist). The maximum frequency that the visual system responds to depends greatly on a large number of parameters, including contrast levels, and individual sensitivities, but, generally, the upper limit is about 150 Hz.
If your intention is to draw attention, then 5-to-10 Hz is excellent for this.
If your intention is to make even illumuniation, then oscillating at or above 200 Hz is certain to not appear to flicker.
These days, fewer and fewer people recall the horror of using a CRT computer monitor at 60 Hz refresh. It is painful. Seizure-inducing. But increase the refresh rate on a CRT to 85 Hz and most people don't see the flickering anymore. Increase it to 160 Hz, and it is nearly undetectable even in a laboratory setting. That's for a light that essentially goes full-on to full-off.
For light sources that are modulating at less extreme levels (like most LCD monitors, where each pixel is essentially constant in light output as long as the image does not change) a much lower refresh rate is necessary. 60 Hz is just fine.
The most recent laptops that have LED illumination are, unfortunately, modulating brightness with a typically 60 Hz PWM driver. This makes it much more like a CRT because the light source is pulsing on and off. I can see my laptop flicker easily, but it's only milding annoying (120 Hz, and it would not have been an issue -- and the thing is, a 120 Hz PWM driver isn't any harder to design than a 60 Hz one).
For a bike light, where you are using the illumination to see where you are going, I would recommend at least 100 Hz, if not 200 Hz. Do you have a choice in PWM drivers, or are you designing you own?
These are going to be awesome in an office environment. Especially since the ceilings are so high and nobody likes changing the lights. But I have yet to find truly warm non-tungsten/halogen/mercury/fire/quartz/evil light for home use. I could not picture LEDs (which are basically antennas radiating a frequency that we happen to see) overtaking the other lights (heat sources that coincidentally give off visible light) in terms of color richness.
Um, incandescent bulbs are also antennas that are radiating a band of frequencies we happen to see. Same with fluorescents. LEDs tend to have narrower spectra, but not the so-called white LEDs, which operate very much like fluorescents: they excite a phosphor.
You should try a high-quality fluorescent. They exist but are not inexpensive. However, they are very very good for doing color-accurate work. Even decent daylight fluorescents can be refreshing, if not entirely color accurate.
standard LEDs dim just fine
Not really; not like an incandescant. An incandescant dimmer is simply a rheostat (variable resistor), but Light Emitting Diodes are binary - either all the way on or all the way off. To dim an LED what you do is turn it off and on faster than the eye can see. The longer the off period, the dimmer it appears.
Dimmers for LED lamps will most likely come on the market, but the dimmer that's on you wall now won't work.
Half right. The last half. Sort-of.
LEDs have a quite nice linear lumens-vs-current regime as long as you're above threshold and below saturation (and you keep them temperature stabilized). OK, it isn't quite linear, but it's pretty close. It certainly isn't binary.
However, traditionally, LEDs were engineered so that the standard current for reasonably bright indicator use was 20 mA. Why 20 mA? Because that's the standard output current that a TTL gate can push. So, with a normal TTL logic chip, you can directly drive one LED, get good indicator use, and remain within the limits imposed by the TTL specification. This led to widespread use of LEDs as indicator lights -- thus an unwitting assumption that they only operate as on/off devices.
That said, the standard dimmer circuitry for incandescent bulbs, either the very old fashioned rheostat or the modern phase-adjusted SCR trigger, will work just fine with an LED load. The problem is that LEDs, unlike incandescents, have effectively zero emission inertia: you can switch them on an off very, very fast. Even a run-of-the-mill LED will switch on or off in microseconds, max. An incandescent bulb, in contrast, takes hundreds of milliseconds to turn on or off, and so the sharp switching waveform from SCR style dimmers is no big deal: the thermal inertial of an incandescent smooths them out. An LED on that kind of a dimmer is pure torture (I've tried it) since it's 50/60 or 100/120 Hz of full-field flickering light. Not pleasant in the least. New dimmer designs will have to be made for LEDs that include some smoothing within the dimmer, or the LED bulbs themselves will have to incorporate waveform smoothing.
But, the last suggestion -- that the apparent brightness of an LED can be modulated by changing the duty cycle -- is correct. If there are any designers reading this: PLEASE MAKE CERTAIN THE MODULATING FREQUENCY IS ABOVE 200 Hz (curious? I'd be happy to explain). The current rash of 60 Hz pulse-wave modulated LED illumination everywhere is driving me nuts.
Is this one of those words that has surreptitiously entered our language like "blog" or was the title just cut-off?
It's been part of the standard lexicon of photographic professionals and amateurs, and especially in the press, for as long as I can remember; and I've been shooting (yes, "shooting", and by gum, that's another idiomatic use of a term that's specific to a field different from computer geek stuff) for three decades now. However, "photog" it one of the odd terms that gets used far more in print than in verbalized conversation, probably because it lends itself to a more compact telegraphic style in headlines.
Personally, I use "photog" mostly as an indefinite noun, when describing a collection of photographers I don't know, or a collection of photographic things of uncertain size. Like, "the photog's convention," or, "that warehouse of photog equipment," or, "where is all my photog crap?"
Also aided in establishing "photog" as an accepted term (again, in a field I'm guessing you're not familiar with) is that Domke (a manufacturer of pro-grade photographic accessories; you might not be familiar with them, but they make good stuff and any pro will know about them) sold a "Photog Vest" for many years. They might still, for all I know; I bought mine in the mid 90s.
Oh, but there I go again, "pro" is an idiosyncratic term within the photog community for a professional photographer. In a previous life, I have been a pro, complete with (profitable!) one-man shows and paid assignments, but am now just an enthusiast photog.
The entire point of this, in case it hasn't become clear: there are terms that are widely accepted within other fields that are perfectly valid; just because you aren't familiar with the field doesn't mean you can't open your mind. The amount of jargon that's in the computer geek field is staggering: would you really expect someone, say an enthusiast photographer, to know it all?
p.s. A quick google search on "photog" would have shown that it gets used in headlines all over the place. That should probably have answered your question right there, but I like to be thorough.
We more-or-less got enough computing power for most things with the introduction of the PIII 1GHz CPU. You might not agree with this, but it's at least approximately true. A computer outfitted with that processor and reasonable RAM browses the web just fine, plays MP3s, reads email, shows videos from YouTube, etc. It doesn't do everything that you might want, but it does a lot.
If we took the amazing technology that has been used to create the 3 GHz multi-core monsters with massive on-chip cache memory in a power budget of 45W or so in some cases, and applied it to a re-implementation of the lowly PIII, we'd win big. We'd get a PIII 1GHz burning a paltry few watts.
And this is precisely why chips like the Intel Atom have been so successful. Reasonable computing power for almost no electricity. We don't necessarily need just MORE-FASTER-BIGGER-STRONGER, which is the path Intel and AMD have historically put the most effort into following, we also need more efficient.
The victorious strategist only seeks battle after the victory has been won, whereas he who is destined to defeat first fights and afterwards looks for victory.
- Sun Tzu
My uncle's version of this is: only place a bet when you are certain of the outcome. He never loses a bet.
He used the words "data" and "statistics" in his inaugural address in a positive tone, without being the slightest bit derisive. He said that he would, "restore science to its rightful place." There is hope for the US.
Belkin does not participate in, nor does it endorse, unethical practices like this.
Given that the first part of this official statement is prima facia a lie -- Belkin DID participate in these unethical practices, as it was an employee, a well-placed employee high up in the managerial chain, who created the reviews in question -- it is reasonable to expect that the entire apologetic statement is not even worth the electrons spent to create it.
But why does it take more than a few milliseconds to discover a device? I've never understood this. We have had CPUs that have had sub-microsecond execution cycles for DECADES now, and yet the timeouts for communicating with devices are still measured in seconds. Why?
Device discovery should take no more than a few milliseconds for an entire machine, with the possible exception of disk drives which presumably need to spin up and verify correct operating speed to report back on a self-check.
Pros:
No power supply needed for each machine. This removes a major point of failure. Instead, one would need to just step down voltages to the 5 and 12 volt rails.
Um, no. Stepping down a higher voltage to a lower voltage, say 48VDC to 5VDC and 12VDC from your example, would still require a power supply. That, precisely, is what a power supply does: voltage conversion. What you're talking about is a DC/DC power supply, rather than the traditional AC/DC ones. But, when it comes down to it, all of the modern computer power supplies are switching supplies that take AC, put it through a bridge rectifier, filter it a little, and then follow that with a DC/DC converter.
By providing 48VDC instead of 120VAC to the input to a power supply, you're eliminating the bridge rectifier (with typically less than 1% losses) and the filtering capacitor (with even lower losses, in a decent design at least). That's it.
Now, converting the noisy, ripply 170VDC that results from rectifying 120VAC to a low voltage is a harder task than converting relatively clean 48VDC to the same low voltage, and starting out with a input voltage that's almost the same as the output voltage means lower current losses, especially in a switching supply. But -- and this is a huge issue -- high current DC connectors are extremely difficult to design because of the inherent electromigration/electrolysis you get with DC current that is just not present at all with AC current. Where are said connectors? At the input to the power supply.
Because there is only one 48VDC power supply for a room, it has to be held up to a lot more rigorous standards than average mains current. It has to not just provide 48VDC, but provide it under extremely heavy load without the voltage dropping by much.
Any decent power supply works over a large range of input voltages. Designing for a precise input range (say, 48 to 48.5VDC) is shortsighted. Ever notice that the laptop power supply you have is rated for 100-240VAC? So are many (not all, but many) desktop power supplies. While the hypothetical room's 48V needs to be pretty good, it does not have to be held to any higher standards as you suggest. The circumstances where it would need to be held to higher standards would be if the delivered voltage were to be used directly without any local regulation / re-provisioning. And that would be just plain poor engineering.
The most amazing forensic work I read of was the Lockerbie bombing of a TWA flight while in midair. The debris was scattered over many square miles. Yet the investigators were able to reconstruct the bomb and find the bomb's timing circuit. A chip in the timing circuit was traced to the perpetrators.
That was pretty fucking cool, I thought.
It was incredibly cool. I believe the fragment traced to the perpetrators was a small part of a circuit board that was identified to be part of a Toshiba radio.
That alone, to me, was -- and still would be -- an astonishing feat of detective work.
Then you haven't bought a good CFL. Fluorescent bulbs with excellent, very white spectra are certainly available. Look for bulbs with color index (CRI) of 98 or over. They are not, however, inexpensive.
It is a myth that fluorescent bulbs with high-quality spectra are not available.
So-called white LED spectra are produced just like the spectra of fluorescent bulbs: white LEDs are actually UV LEDs that illuminate phosphors with different inherent spectra to cover the visible range.
For a brief while, triplet R/G/B LEDs were put in one package, but those are (a) too expensive, and (b) suffer from dramatic white-point shift (color shift) as the three chips age differentially.
you really didnt think that through to much did you...
tires last an avverage of 60k miles... so that would be approxamatly 2400 gallons of gas at 25 MPG. Oregon's current gas tax is 43.4 per gallon... so you are suggesting $1041.60 per tire in taxes?!?
Frankly, it's annoying when people accuse me of not thinking things through and then don't bother expending any intellectual effort themselves. How do you think a GPS mediated tax will work? Something like each year you'll go to a gas station that has a special instrument that reads out how far the chip in your car says it has traveled. Do you think the tax bill will be small at that point? The 10 to 20k miles of wear that each car on average places on the roads will be recovered one way or another. If the GPS-based tax is to completely replace the gas tax, then the bill at that point will be, indeed quite high. The State of Oregon is not going to give up this income stream.
But instead (and this is where you should have thought a little harder) of measuring distance traveled via GPS (with the concomitant invasion of privacy) the same service station could just measure tire wear. A simple mechanical tread depth gauge that every service station already owns would suffice. If the owner changed tires in the middle of the year, the depth measured would be pro-rated to the date of purchase. If the owner didn't have the receipt, then pro-rate the wear to 2 months (screwing most, and providing an incentive to keep the receipt; there would probably need to be an abatement procedure).
Or (gasp!), we could just use the odometer and factor in the published weight of the vehicle. But that's a less accurate measurement of the actual load each vehicle places on the road. Tire wear is about as good as it can get, but there are plenty of alternatives to GPS-based measurement if a simple gas tax is thought to be no longer sufficiently accurate.
A use tax for our roads and bridges is exactly what the gas tax is supposed to be. The deterioration of a road is directly related to the number of vehicles and their weight. When engines are all about equally efficient, then taxing fuel is a good a proxy as any for road usage.
But with the advent of hybrid technologies, and the push for new fuels which might not be gasoline, there is an inequality with use vs. amount of taxes paid for that use.
Taxing the number of miles traveled, however, is not the right answer. It's not even a good answer. It is, fundamentally, a bad answer. It is an answer with cloaked malicious intent, because it seeks to monitor the movement of the citizenry rather than recover taxes proportionate to use.
Far better would be to radically reduce or eliminate the gas tax and introduce a steeper tire tax. Tire wear is a far better proxy for road wear (that is, the amount of wear that the road has, which, mostly, determines when it needs to be resurfaced), since it is proportionate to vehicle weight and number of miles traveled. Road wear is also, not surprisingly, proportionate to vehicle weight and number of cars that pass a given point (ergo, miles traveled by each car).
Just tax tires. Problem solved. No need for additional gadgetry that requires a new infrastructure for inspection and reporting. Seriously, either this legislator is on crack or is being backed by some seriously questionable money.
When dealing with a pressurized body of molten rock with entrained gasses, I don't think one could ever say it is perfectly safe.
Absolutely. I go to Santorini, Greece every year or so. Santorini is one of the best-studied volcanoes in the world. A volcanologist friend often accompanies me and enjoys reeling of stats about magma chamber releases. Force of Nature level stuff. It's not the dissolved gases that matter quite so much as the dissolved water. Magma is typically a few weight percent (superheated) water. During an eruption, when the pressure is released the water turns to steam and the whole shebang expands in volume by 3 orders of magnitude. Until the upper part of the magma chamber is emptied, the ejection column flows at supersonic speeds. You do not screw around with Mother Nature. Intentionally drill into a magma chamber? No thank you.
Wow, thanks for the good info. Here's a quote from the article that you already linked:
SourceForge also said it "intends to take aggressive steps" to accelerate the growth of its core Internet properties: Slashdot, SourceForge.net and ThinkGeek.
"These sites represent a truly unique set of Internet assets and there is more we can be doing to make them better and more engaging for our users," said Neumeister. "We are focused on aggressively pursuing our plans for each of these sites and making opportunistic investments that will enable us to reach our objectives faster."
Thanks.
Why can't CEOs and such be satisfied with slow, even growth? Why does everything have to be aggressive leveraging of opportunity? Slashdot is never, ever going to be the next NYT or Washington Post. Why not just be satisfied with a job well done and a decent, reasonable profit?
The suits kill everything.
The worst part about the SSC is mentioned in the parent comment's parenthetical comment about brain drain.
When the SSC was cancelled, there was a flood of high-energy physicists who were suddenly out of work. The US lost an entire generation of talent in physics. Instead of continuing on with a remarkable collection of centers of excellence, each themselves breeding excellence, and maintain the intellectual, scientific, technical, and economic advantages that the US Government prides itself on, the (pardon me) boneheads in Congress thought it better to continue the long slog toward mediocrity.
High-energy physics no longer happens in the US (my apologies to readers at LL, LANL, Brookhaven, Fermi, Argonne, Berkeley, and so forth). It happens in Europe and will continue to do so for the forseable future.