Even soft water has enough mineral content to change the flavor, and I suspect that's part of the beer flavor. Brewer's tend to be very particular about their exact water source, which wouldn't be an issue if distilled water worked well.
...which goes back to the fundamental flaw in the brewers' argument. Either they are currently using distilled water (ie. "minerals dissolved in water" isn't a whitelisted ingredient according to the purity law so they have been using distilled water since the 16th century [haha, sure]), or they are using water that has additional dissolved molecules that "don't count" for the purposes of adhering to the law for whatever reason. If one set of dissolved molecules is allowed an implicit exception to the law, then why not another set?
My guess is that they aren't using distilled water, and therefore their entire line of argument is nullified. That's not to say I am in favor of added fracking byproducts in beer—merely that they need to find another rationale for opposing it than the beer purity law.
Then again, they probably realize this and are just attempting a publicity stunt.
Ah, no. Modern US pennies will not pick up with a magnet. Modern pennies(post 1982) are 97.5% Zinc with the remaining 2.5% being electrolytically plated copper.
The only pennies you could pick up with a magnet were made during World War II, due to copper shortages. They were zink coated steel.
Yes, you're correct in the common sense of the discussion, but you can have fun with Lenz's law.
Take a strong permanent magnet and a coin made of rather conductive metal (I like to use 90% silver quarters). Place the coin on a wooden table with the magnet on top. Rapidly lift the magnet and you will see that the induced magnetic field in the conductive coin will cause it to lift as well.
Even with powerful rare earth magnets and highly conductive silver the effect is minor. However, you should see it lift off the table if you perform the experiment correctly. My guess is that a pre-1982 penny would work too, though it might be more difficult to execute the technique.
The we that is currently running the country has changed in form if not function.
Please. You are deluding yourself. The United States is a hegemonic superpower... the only one remaining. As such, we meddle in the affairs of other countries in order to effect the outcomes we desire.
I see you have attempted to move the goalposts: you assert we have somehow repented of our hegemonic actions sometime in the past ~60 years despite no official apologies, policies, laws, or treaties that would suggest anything like this has happened.
Okay, I think it's naive to believe we had nothing to do with the 2002 coup attempt against Chavez. Is that "ancient history" as well? Furthermore, you should explain why you believe our invasion of Iraq was based on noble intentions. Frankly, I believe it was performed due to either a grudge W. had against Saddam for plotting to kill H.W. or, more likely, a desire to ensure that Iraqi's fungible petroleum reentered the global market in order to ease competitive pressure from China on US oil consumption and/or to ensure that the global oil trade remained denominated in US dollars. The ubiquitous petrodollar ensures that the world continues to have a demand for US dollars so they can buy oil. This causes countries to sell their exports to us for US dollars in order to be able to turn around and buy oil from the Middle Eastern nations. This is how the US props up the unending, perennial trade deficits and insane government spending.
I could continue, but by this point you are either convinced or you are resolved to cling to the idea that the US is somehow righteous. Don't bother trying to move the goalposts again: these specific examples aren't critical to my assertions (because there are more).
Do I believe the US is the "Evil Empire of the World"? No. I am just realistic enough to understand that we are willing to engage in cruel, dirty tricks—including actions that lead to the death of innocents—in order to perpetuate our global dominance.
Life is more pleasant if you just embrace this fact and release your cognitive dissonance that attempts to reconcile US actions with your idealistic perspective of US foreign policy. Someday you may be correct about the US, but that day will be after we have lost our superpower status. Imperial Britain was every bit as cruel as I assert the US is (and has been), but the modern UK was neutered by the US after WWII (cf. what the US threatened to do to the GBP if the UK didn't come to heel in the 1956 Suez Crisis). As such, the UK more likely to be "gentle" to other nations.
Just like the US will likely become once we are eclipsed by an ascendant superpower. Of course, that superpower will be cruel in order to maintain dominance. There will be apologists who will assert that superpower is righteous and noble, and there will be still others who will disagree based on the superpower's actions. Thus the cycle will perpetuate.
Poor filter design (though surely it wasn't mistakenly set up with a high/low-pass filter in a bandpass arrangement)? Flawed assumptions about the "thump" followed by a high frequency "tinkle" when glass breaks? Difficulty constructing a state machine for these conditions? Poor/nonlinear mic response over the frequency range required?
Was it subject solely to false negatives or were false positives a factor as well?
Yes: passive aggression is usually a indicator of a weak personality. People really should try regular aggression instead. Seems we share this perspective.
Though there is something to be said for the satisfaction of telling a blatant lie to someone you don't respect while looking them in the eye...
You just need to claim that you are a Mennonite, and for religious reasons you do not own a credit card. Generally employers know so little as to be unable to verify the check, and are usually too scare of state and federal laws around religious discrimination to mess with someone who has an education and a moderate level of affluence.
"Sorry, can't. I am a Mennonite... um, (software developer|computer engineer|adherent assisting in the sale of high-tech materialistic comforts via television infomercials). (*cough*) Right. Well, best be off to my buggy then..."
Bonus points for insinuating that you don't respect your employer enough to tell a plausible lie.
(Just as an aside, but poor people have access to $125+k in personal revolving lines of credit from Amex now? I guess I need welfare—or perhaps insulation from the inevitable credit bubble collapse that would cause)
Haha, awesome... "Help me, owner, I'm gonna die because I can't figure out how to stop draining my battery! I don't perceive the irony of consuming battery power to display this message, either!"
Haha, they also offered the option of having the car phone home with a GPS fix in order to dispatch the emergency "help you plug in your Roadster" crews too. All these "preventative measures" they tried appear to be a brain damaged, backward, insanely complex approach compared to the obvious solution of having the system simply shed all but trivial load/total shutdown/disconnect the battery via relay (had they designed one in)/whatever during a low-charge emergency.
Which apparently they figured out eventually, because they released an update to allow the vehicle to "sleep" now, right?
According to Battery University, it seems that a lithium ion battery at 40% charge level should retain 96% of its capacity after a year of storage at 25 C. So, it seems that simply removing *all* load (including state-of-charge monitors) after the capacity gets "low" (pick a value) and there is insufficient power to charge should prevent all damage to the battery pack. The further below the ideal storage charge level of 40%, the sooner a parked vehicle without charging power should enter emergency, total shutdown. This just cannot be that hard, and should have been patently obvious to anyone designing an EV's power system.
I guess that tips the balance for Telsa in favor of incompetence instead of malice. For whatever that's worth.
"Sorry, sir, it seems you left your Civic parked for several days with only 10% fuel left in the tank. I know that you're upset this caused $15k worth of damage to your car when the last of the fuel was finally burned while the car sat around parked... but if you refer to the 'sad face' icon next to the warning on page 327 of your manual, you see that this isn't covered under warranty. We warned you... and why would you leave your car like that in the first place? What were you thinking?"
It's not reasonable to engineer to "the happy use case"; the boundary conditions are more important—especially when those scenarios lead to damage! The fact remains that Tesla thought it was acceptable to allow the battery to fatally discharge itself which required a $40k battery replacement that wasn't covered under warranty. They shipped that as a known, documented bug when they could have (and later apparently did) prevented it via a (firmware?) update.
Anyway, it really seems to be either malice or incompetence on Tesla's part: malice, if they realized this would happen and just really wanted to ship before they could implement a fix; or incompetence, where they couldn't figure out how not to kill the battery while the car was sitting around unplugged. Either way, it's unacceptable. It doesn't matter if their happy use case involved the car being plugged in: there was no excuse for the car to damage itself if left unplugged.
You're the battery chemistry guru... tell me I'm wrong, and that there was some reason the Tesla Roadster had to kill its battery when it was left unplugged, and only some deep wizardry allowed Tesla to push a miraculous update that somehow amazingly prevented the battery pack from committing seppuku like the battery's chemistry naturally really wanted to force it to do.
I am not pleased with seemingly being forced to believe Tesla is malicious and/or incompetent, which is the only other possible conclusion. No one expects an uncharged EV to be able to drive, but they do legitimately expect the vehicle not to be permanently damaged by that condition.
If Tesla cut corners on this, then what else have they cut corners on?
Wow, that's not really a testament to Tesla's commitment to quality, because they shipped with that as a known issue (they documented it in the user manual as not covered by warranty). Their original workaround was to have the car phone home with a GPS report so that Tesla personnel could come out and help plug your car in for you. "We had to consume the battery in order to save it!" (*cough*)
Not exactly the pinnacle of sane engineering there. If your stat and Wikipedia's capacity specs are correct, then that's a vampire load of 175 W (8% of 53 kWh per day)—and they found that to be acceptable and reasonable? Even when the vehicle's battery was in dire threat of doom in terms of state of charge? There was nothing they felt like they could cut from that power budget, even in emergencies?
Yes, voltage drop over a long extension cord is expected when its under load. However, that's not the real issue: the actual problem was the "let's allow a daily 8% charge consumption while quiescent that leads to uncontrolled, fatal discharge, without understanding it is important to be able to shed load" engineering decision Tesla made. Furthermore, even the battery monitor IC's have a low-voltage cutout to protect the battery these days. I guess Tesla couldn't be bothered to follow industry standard practice in terms of power cell protection... better to just drive the $40k battery over the cliff, eh?
I mean, what's the expected, optimal, no-load, self-discharge rate for these cells under typical Tesla owner conditions (ie. garaged, or in California)? 10% per month?
Anyway, if what you said about devices needlessly killing their batteries while plugged in is true, then I hope that the engineers responsible for battery-powered products generally become less... "simpleminded" in terms of their perspective on battery circuits.
This all seems very ass-backwards in terms of design. Why drain from the battery while it's plugged in? My HTC Incredible has taken this brain damage even further: the battery completes the circuit and the phone literally can't boot unless a battery is in place. If you are referring to modern devices like the Nexus 10 whose drain exceeds the charger supply, that's one thing.
However, let's talk about laptops then: there's no excuse for a laptop power supply to be less than the worst case draw of the laptop. As for cutting the battery out of the circuit: can't a power MOSFET (serving as a solid state relay) handle that much current? I seem to recall that Rds has gotten low enough that all the laptop's current wouldn't pose an issue... and switching the FET should take milliseconds or less, so power interruption wouldn't be an issue.
Also, I'm not sure that citing Telsa as an example of battery engineering is a good idea... considering they self-destruct their batteries if they aren't constantly plugged in. Hell, one $40k Telsa battery pack died because the owner had plugged it in via a 100+ foot extension cord. Sure, I^2R losses, but exactly how much power does the damn thing need to keep its head above water while it is sitting around, doing nothing?
Seriously now, I thought modern lithium technology had excellent charge retention. So, are they honestly claiming that they can't figure out how to shed vampire load on the goddamn car in state-of-charge emergencies and therefore just have no choice but to deplete the cells below 2.7V (or whatever)? It just can't be that hard... stop charge balancing if state of charge gets low, stop powering all the telecomms/gps/logging/computer crap, and the quiescent battery should be fine for a Long Time(tm), right?
I guess I am still annoyed at my Thinkpad killing my battery even though I practically never used it unplugged. Then one day I found it dropped charge from 100% to 0% in 3 minutes whenever it was on battery. Basically, what you're saying is that this could largely have been avoided if the engineers had designed the circuit to be able to cut the battery out of the loop while it was charged and the laptop plugged into wall power, right?
...you know, like perhaps by the Dutch, who wouldn't have a country anymore. Oh, and maybe New York could start building a dome as some sort of post-apocalyptic bastion.
Ironically, Mitsubishi will probably be building and selling wind turbines as part of this deal.
...for certain, non-ironic values of "ironically".
Unless you count Ally Bank (nee GMAC) financing the sales of GM vehicles as irony. Or Toyota Finance offering loans to Toyota vehicle buyers.
There are countless examples. Hell, even the Marshall Plan was similar in concept—and that might be considered ironic (considering that US aid was given to reconstruct Japan after WWII).
You want a meritocratic, agile, decentralized Union that represents a Hacker/techno-utopian ethos, with blackjack and hookers and bitcoin micropayments? Then make that union!
Well, that would take about pi microseconds to collapse from the metastable state you describe into one of the binary stable states of either a "mediocracy tyranny" or Lord of the Flies.
Because if you give the unionized team the ability to choose who to hire and fire (hey, this is a meritocracy, right?) then the floors of many cubicle halls will run red with the blood of the unpopular developers. Even in a purely democratic collective, witch hunts leading to termination will become commonplace. I know, because I have witnessed developers "marked for death" in standard corporate environments where it's typically hard to convince the management to cull those considered "dead weight".
When management finally gave in, it led to a Reign of Terror akin to the French Revolution. There were multiple terminations on the "hit list" that were executed, before the purge finally turned inward and culminated in the termination of the highly-skilled developer who acted the part of Robespierre in organizing the terminations of the unworthy. He was eventually devoured by his own purge that had grown out of control, as it were.
Or, like I said, it could devolve into mediocracy where union developers threaten to file a grievance if someone in management makes their own Excel spreadsheet without a union "brother" developer involved. And no one gets fired, ever, except due to layoffs and then only in reverse seniority order. This is outwardly a very tranquil, boring, stable state, but that is merely the velvet glove over the iron fist of mediocrity that crushes developers' souls.
So, a union could start out like you describe, but ultimately I don't believe it is a stable state. But damn, maybe I'm just jaded by experience. Feel free to prove me wrong by setting up an awesome union like you describe and be sure to publicly mark the anniversaries of it running smoothly (in order to get the word out). It would be awesome to be wrong about human nature in this regard—I'm fairly certain I'm correct, though...
I agree with this, however it neglects the actual intrinsic value of Bitcoin, which is its putative secrecy for private transactions.
Putative or not, it is illusory. One might just as well extol Slashdot as a forum with putative anonymity for untraceable comments.
In both cases, practical attacks exist to unmask the parties involved unless special preventative measures are taken. Eg. for bitcoin, laundering; for Slashdot, proxies/VPN/Tor/etc.
Bear in mind that there is a limit to "secrecy" in any system that requires that the ledger of all history of every transaction be publicly available and downloaded by every client. This means that transactions can be traced from a known wallet address through the chain, much in the same way that the US ATF can putatively trace firearms based on serial numbers.
Last I checked (in February) the blockchain data was several GB on disk. Per client. And took days to sync. "Thanks, Satoshi Dice!" (ugh, talk about crapflooding the blockchain...)
But this is why I've blocked all Linux articles and won't be dealing with their shit anymore, have enough on my plate without wasting my time dealing with crazies.
Hey, I saw this while expanding the tree of our discussion.
I apologize if our long-running discussion (I thought it was rational rather than insane, but haha, okay?) was "the last straw" that soured you to the point of blocking an entire segment of Slashdot. I felt no acrimony toward you throughout the discussion, and I am actually sorry if you felt otherwise.
Best wishes for your health issues—I have no clue what you use to sleep, but personally, I find doxylamine succinate (aka. "Unisom 1") to be a powerful, benign OTC sleep aid vastly superior to the diphenhydramine types (aka. Benadryl, "Unisom 2", Simply Sleep, Tylenol PM [ugh, with acetaminophen, despite most users just wanting to use it as a sleep aid]). Incidentally, both drugs are antihistamines so they can double as anti-allergy drugs (though doxylamine is more powerful).
Dude if you want to run crypto that the E350 won't run and can't find the card? Fine get an A-Series, you can get one shipped for like $65 for a triple that'll run any damned thing you want. Hell if you don't mind getting an older model you can get E series ULV quad for $68.
Right, pretty much any desktop chip will perform adequately. However, those examples both increase the power budget to a 65 watt TDP for the CPU, which doesn't really fit into low-power system.
IIRC I found 'em by looking up "cell accelerator board"
Okay, using that search token I found the Mercury boards. Holy crap, you weren't kidding about blowing the power budget with those: the model I read about had a 210 watt TDP with 4 GB of RAM onboard, etc.
But again this whole thread started because somebody said the Pi could do it which surely to God even you will admit that isn't possible
Was this what you were upset about? I never contested that... it's an absurd notion. All I was trying to do with your initial comment was add a slight disclaimer to the effect of: * E-350 may not be sufficiently powerful to drive a simple crypto filesystem for encrypted storage while simultaneously processing a dashcam video source.
Though, this being Slashdot, perhaps someone else might come along and incorrectly interpret that to be an endorsement of the RPi for this usage. Haha, okay, so I will add a meta-disclaimer: ** E-350 may or may not be adequate, but the RPi, a pocket calculator, or a scurry of squirrels using abacuses are certainly not adequate.
(IIRC its designed primarily for AES but I'm sure others will run)
Current hardware acceleration for crypto typically revolves around AES (for symmetric/block), RSA (for asymmetric), and/or SHA-1/SHA-2 (for signing/hashing). The reason AES was selected for acceleration is probably due to the modified Rijndael algorithm that became AES having excellent algorithmic efficiency (that's one reason it won the competition). For example, the AES-NI acceleration in Intel/AMD CPUs only adds seven new instructions (all specifically related to fundamental AES operations). It is unlikely that those instructions could be used to accelerate, say, Blowfish, so other algorithms tend to require execution via generic CPU instructions and thus would achieve no speedup. This is also why the addition of AES-NI acceleration increases performance by 10x (or more) on the same silicon.
AES is certainly good enough for security, and is likely to have the best performance across the gamut of hardware, so it's typically the best choice unless someone really wants to go all crypto-nerd. And it's much easier to deploy than something using potentially-faster, obsolete crypto like DES, because none of the crypto filesystems offer that as an option anymore (if they ever did).
THAT solves ALL of your problems with the E350 since Via has hardware accelerated crypto baked into the chip and has since the C7. You can pick up one of those for less than $200, use its hardware acceleration to do any crypto you want and there ya go, Bob's your uncle.
Looks good. A comparative benchmark suggests that the C7 can probably hit 40 megabytes/sec for AES-128-CBC (ie. "lowest grade AES encryption") while accelerated. Crucially, having hardware acceleration instructions for AES means that it shouldn't require 100% of the CPU in order to achieve that throughput. So, one might expect a dashcam carputer with a C7 to be able to handle encrypted storage throughput sufficient to stream data to disk while also having enough capacity to do "other stuff" related to the dashcam video source. And not break the bank or be a space heater.
Cool. I'm glad you brought that up and that we continued the discussion long after everyone else left.
Hell this whole thread was started by some guy saying a fucking raspberry pi could do the job, why aren't you bitching at him?
The reason I'm not "bitching" at anyone else is due to your response to my original comment. To refresh your memory, this was my initial comment: "Just an FYI: the maximum throughput of an E-350 to an SSD encrypted with AES-128 CBC (4K sectors) is on the order of 30 megabytes/sec... and that's with full CPU use for the crypto (i.e. the system is doing nothing else)."
This is a simple fact derived from repeated testing (by myself and others). You superciliously attacked this as a crypto geek overthinking the problem, which led to my attempts to explain. I really don't know what your original disconnect was on that, because eventually you suggested using TrueCrypt (a crypto filesystem with stronger encryption than what I cited above) as a "simple" solution...
As I said you have options, there are several E350 boards that give you a full X16 PCI-E that is more than enough to support a drop in cell card or FPGA and while I haven't looked at FPGA crypto in ages IIRC the cell card (which FYI can be had for around $100-$150 last i looked but its been awhile, you might have to ebay it) supported up to 1024bit in just about every format you can name OOTB so then you are only using the E350 for basic tasks like memory management and letting the cell do the heavy lifting.
Okay, I just spent 15 minutes scouring eBay and elsewhere looking for a crypto card like you describe. Needs to be a PCI-Express card with AES support, and have an AES throughput better than 250 Mbps (which is approximately what you can get with the E-350 CPU by itself). Oh, and it can't draw more than, say, another 17 watts or cost more than, say, $150 (used on eBay is fine). If you have a link to something like you describe then that would be appreciated; I might actually buy it.
Like I said, you can buy a $200 (brand new), 17 watt TDP Xeon E3-1220LV2 *today* that has >1 Gbps AES crypto filesystem performance without having to blow your money & power budget on something expensive like a crypto accelerator that does NSA-grade crypto at multigigabit speeds.
if you went to the Open-CL forums and told them what you needed it to do you could probably code it yourself to run on a $65 AMD A4-3400
Jesus, and you accused me of overcomplicating things? You are actually proffering the idea that it is reasonable to code your own GPU-accelerated crypto filesystem (of which zero examples exist in the real world) for a project like this? Even knowing that a sub-$300 Xeon-based system can handle everything in a ~30 watt complete system using widely-available, well-tested software like TrueCrypt, LUKS, geli, etc, etc?
Oh, hey, "Hoss"... I found this pithy XKCD comic that reminds me of you: XKCD: Security... you know, some crypto geeks out there really overthink security. Haha, don't overthink this.
Oh and just FYI there won't be any consumer Jaguar chips, at least not for another 3-4 years
That's too bad... the damn rumor mills had piqued my interest. I guess I will spec my next low-power builds to use a low-power Xeon instead. That's probably only about $75 more—at ~$250 for the Xeon CPU+motherboard—than I paid for my E35MI-I Deluxe when it came out. Graphics performance is irrelevant for all of my uses, so that's not a comparison point I weight.
You get a lot of mileage of your E-350 installs; that's great for you. However, they may not be fit for purpose in the case of needing to use crypto filesystems (TrueCrypt, etc) with even low-grade encryption... that's what my original comment was about. This is a valid consideration if someone were planning to build/use an E-350 with the intention of using even simple, widely-available crypto, which fits with this overall subject. Or if someone stumbled across your orig
Ah, but you aren't thinking degenerately enough. Have you seen the Pulse Audio diagram? ALSA is a source *and* a sink in the PulseAudio abstraction layer—what a great design!
In the spirit of PulseAudio, I imagine a typical Unix design approach for a de novo "universal" framework to allow application remoting would be a meta-layer on *top* of both X and Wayland (and/or any other framework that comes along). And, for platform portability, it could render widgets using Java Swing.
As long as I'm guaranteed that my local display can talk to remote applications, no matter what the flavor of the month is, I'll be happy. There needs to be a universal standard.
You do realize you probably prompted some brain-addled dev someplace to come up with the equivalent of PulseAudio for display technology, right?
All problems in computer science can be solved by another level of indirection—except for the problem of too many layers of indirection. —David Wheeler
But when you are talking about running something off a car you better be VERY picky on the power budget or have a hell of a battery, that is all I'm saying.
Per TFS, this whole context is about dashcams, so the entire topic is implicitly about vehicular use. FYI, in case you ever consider building a vehicle-based computer, these tiny DC-DC vehicle PC power supplies support being wired to toggle based on the ignition state and usually come with a low-battery cutoff protection for the vehicle. So, no, you don't need a hell of a battery or an expensive setup. A vehicle can power a 30 watt system like that with zero risk.
Hell a Truecrypt volume can do THAT, that ain't brain surgery.
Do you want to know what my beef is, "Hoss"? It's that when I originally commented about the performance of crypto filesystems (what do you think TrueCrypt is?) on the E-350, you mocked the concept as a crytpo geek overthinking the issue. So, what's this you're saying here? Should I link you to that XKCD comic, "Hoss"? What I was proposing was using a simple crypto filesystem package with *lower* grade encryption than TrueCrypt can do, because, as we agree, it doesn't need to keep out the NSA.
Let me put it very simply for you: on the lowest encryption grade TrueCrypt offers, you can expect a TrueCrypt volume on an E-350 to have about a 20 MB/sec write throughput while TrueCrypt pegs the CPU. Maybe that's good enough for something like a dashcam system, maybe not. If the E-350 had AES-NI (like its successor probably will), then the same configuration would have ~90 MB/sec throughput with 20% CPU usage with the same 17 watt TDP.
It's a false dichotomy to suggest that this is a choice between an inexpensive, low-TDP system with minimally-acceptable crypto filesystem performance vs a hellishly-expensive quad-core "space heater" type CPU plus a crypto FPGA PCIe card.
Slashdot Mirror
Distilled water also tastes like crap.
I beg to differ, but this is a tangent anyway.
Even soft water has enough mineral content to change the flavor, and I suspect that's part of the beer flavor. Brewer's tend to be very particular about their exact water source, which wouldn't be an issue if distilled water worked well.
...which goes back to the fundamental flaw in the brewers' argument. Either they are currently using distilled water (ie. "minerals dissolved in water" isn't a whitelisted ingredient according to the purity law so they have been using distilled water since the 16th century [haha, sure]), or they are using water that has additional dissolved molecules that "don't count" for the purposes of adhering to the law for whatever reason. If one set of dissolved molecules is allowed an implicit exception to the law, then why not another set?
My guess is that they aren't using distilled water, and therefore their entire line of argument is nullified. That's not to say I am in favor of added fracking byproducts in beer—merely that they need to find another rationale for opposing it than the beer purity law.
Then again, they probably realize this and are just attempting a publicity stunt.
Ah, no. Modern US pennies will not pick up with a magnet.
Modern pennies(post 1982) are 97.5% Zinc with the remaining 2.5% being electrolytically plated copper.
The only pennies you could pick up with a magnet were made during World War II, due to copper shortages. They were zink coated steel.
Yes, you're correct in the common sense of the discussion, but you can have fun with Lenz's law.
Take a strong permanent magnet and a coin made of rather conductive metal (I like to use 90% silver quarters). Place the coin on a wooden table with the magnet on top. Rapidly lift the magnet and you will see that the induced magnetic field in the conductive coin will cause it to lift as well.
Even with powerful rare earth magnets and highly conductive silver the effect is minor. However, you should see it lift off the table if you perform the experiment correctly. My guess is that a pre-1982 penny would work too, though it might be more difficult to execute the technique.
The we that is currently running the country has changed in form if not function.
Please. You are deluding yourself. The United States is a hegemonic superpower... the only one remaining. As such, we meddle in the affairs of other countries in order to effect the outcomes we desire.
I see you have attempted to move the goalposts: you assert we have somehow repented of our hegemonic actions sometime in the past ~60 years despite no official apologies, policies, laws, or treaties that would suggest anything like this has happened.
Okay, I think it's naive to believe we had nothing to do with the 2002 coup attempt against Chavez. Is that "ancient history" as well? Furthermore, you should explain why you believe our invasion of Iraq was based on noble intentions. Frankly, I believe it was performed due to either a grudge W. had against Saddam for plotting to kill H.W. or, more likely, a desire to ensure that Iraqi's fungible petroleum reentered the global market in order to ease competitive pressure from China on US oil consumption and/or to ensure that the global oil trade remained denominated in US dollars. The ubiquitous petrodollar ensures that the world continues to have a demand for US dollars so they can buy oil. This causes countries to sell their exports to us for US dollars in order to be able to turn around and buy oil from the Middle Eastern nations. This is how the US props up the unending, perennial trade deficits and insane government spending.
I could continue, but by this point you are either convinced or you are resolved to cling to the idea that the US is somehow righteous. Don't bother trying to move the goalposts again: these specific examples aren't critical to my assertions (because there are more).
Do I believe the US is the "Evil Empire of the World"? No. I am just realistic enough to understand that we are willing to engage in cruel, dirty tricks—including actions that lead to the death of innocents—in order to perpetuate our global dominance.
Life is more pleasant if you just embrace this fact and release your cognitive dissonance that attempts to reconcile US actions with your idealistic perspective of US foreign policy. Someday you may be correct about the US, but that day will be after we have lost our superpower status. Imperial Britain was every bit as cruel as I assert the US is (and has been), but the modern UK was neutered by the US after WWII (cf. what the US threatened to do to the GBP if the UK didn't come to heel in the 1956 Suez Crisis). As such, the UK more likely to be "gentle" to other nations.
Just like the US will likely become once we are eclipsed by an ascendant superpower. Of course, that superpower will be cruel in order to maintain dominance. There will be apologists who will assert that superpower is righteous and noble, and there will be still others who will disagree based on the superpower's actions. Thus the cycle will perpetuate.
the USA ... still willing to kill people over access to natural resources
No, the USA is not.
Here's one proof to the contrary: we were willing to overthrow a democratically-elected government merely to ensure our access to cheap bananas.
1954 Guatemalan coup d'état
Bananas, Mandrake? Children's Chiquita bananas?
So: planning to move the goalposts now?
Then, inexplicably, they turned around and just threw most of it away, with the ribbon interface and then Windows 8.
Blame Julie Larson-Green.
Metro's mother to replace defenestrated Windows boss Sinofsky:
"...the fact she was behind the Ribbon and the Windows 8 Metro interface will not fill some folk with confidence"
O rly.
Okay, I'll bite: what was wrong with the design?
Poor filter design (though surely it wasn't mistakenly set up with a high/low-pass filter in a bandpass arrangement)? Flawed assumptions about the "thump" followed by a high frequency "tinkle" when glass breaks? Difficulty constructing a state machine for these conditions? Poor/nonlinear mic response over the frequency range required?
Was it subject solely to false negatives or were false positives a factor as well?
without the euro, [...] the average german would have been about 1000 euro's poorer
...I see what you did there.
Yes: passive aggression is usually a indicator of a weak personality. People really should try regular aggression instead. Seems we share this perspective.
Though there is something to be said for the satisfaction of telling a blatant lie to someone you don't respect while looking them in the eye...
You just need to claim that you are a Mennonite, and for religious reasons you do not own a credit card. Generally employers know so little as to be unable to verify the check, and are usually too scare of state and federal laws around religious discrimination to mess with someone who has an education and a moderate level of affluence.
"Sorry, can't. I am a Mennonite... um, (software developer|computer engineer|adherent assisting in the sale of high-tech materialistic comforts via television infomercials).
(*cough*) Right. Well, best be off to my buggy then..."
Bonus points for insinuating that you don't respect your employer enough to tell a plausible lie.
(Just as an aside, but poor people have access to $125+k in personal revolving lines of credit from Amex now? I guess I need welfare—or perhaps insulation from the inevitable credit bubble collapse that would cause)
Haha, awesome... "Help me, owner, I'm gonna die because I can't figure out how to stop draining my battery! I don't perceive the irony of consuming battery power to display this message, either!"
Haha, they also offered the option of having the car phone home with a GPS fix in order to dispatch the emergency "help you plug in your Roadster" crews too. All these "preventative measures" they tried appear to be a brain damaged, backward, insanely complex approach compared to the obvious solution of having the system simply shed all but trivial load/total shutdown/disconnect the battery via relay (had they designed one in)/whatever during a low-charge emergency.
Which apparently they figured out eventually, because they released an update to allow the vehicle to "sleep" now, right?
According to Battery University, it seems that a lithium ion battery at 40% charge level should retain 96% of its capacity after a year of storage at 25 C. So, it seems that simply removing *all* load (including state-of-charge monitors) after the capacity gets "low" (pick a value) and there is insufficient power to charge should prevent all damage to the battery pack. The further below the ideal storage charge level of 40%, the sooner a parked vehicle without charging power should enter emergency, total shutdown. This just cannot be that hard, and should have been patently obvious to anyone designing an EV's power system.
I guess that tips the balance for Telsa in favor of incompetence instead of malice. For whatever that's worth.
Really?
"Sorry, sir, it seems you left your Civic parked for several days with only 10% fuel left in the tank. I know that you're upset this caused $15k worth of damage to your car when the last of the fuel was finally burned while the car sat around parked... but if you refer to the 'sad face' icon next to the warning on page 327 of your manual, you see that this isn't covered under warranty. We warned you... and why would you leave your car like that in the first place? What were you thinking?"
It's not reasonable to engineer to "the happy use case"; the boundary conditions are more important—especially when those scenarios lead to damage! The fact remains that Tesla thought it was acceptable to allow the battery to fatally discharge itself which required a $40k battery replacement that wasn't covered under warranty. They shipped that as a known, documented bug when they could have (and later apparently did) prevented it via a (firmware?) update.
Anyway, it really seems to be either malice or incompetence on Tesla's part: malice, if they realized this would happen and just really wanted to ship before they could implement a fix; or incompetence, where they couldn't figure out how not to kill the battery while the car was sitting around unplugged. Either way, it's unacceptable. It doesn't matter if their happy use case involved the car being plugged in: there was no excuse for the car to damage itself if left unplugged.
You're the battery chemistry guru... tell me I'm wrong, and that there was some reason the Tesla Roadster had to kill its battery when it was left unplugged, and only some deep wizardry allowed Tesla to push a miraculous update that somehow amazingly prevented the battery pack from committing seppuku like the battery's chemistry naturally really wanted to force it to do.
I am not pleased with seemingly being forced to believe Tesla is malicious and/or incompetent, which is the only other possible conclusion. No one expects an uncharged EV to be able to drive, but they do legitimately expect the vehicle not to be permanently damaged by that condition.
If Tesla cut corners on this, then what else have they cut corners on?
Wow, that's not really a testament to Tesla's commitment to quality, because they shipped with that as a known issue (they documented it in the user manual as not covered by warranty). Their original workaround was to have the car phone home with a GPS report so that Tesla personnel could come out and help plug your car in for you. "We had to consume the battery in order to save it!" (*cough*)
Not exactly the pinnacle of sane engineering there. If your stat and Wikipedia's capacity specs are correct, then that's a vampire load of 175 W (8% of 53 kWh per day)—and they found that to be acceptable and reasonable? Even when the vehicle's battery was in dire threat of doom in terms of state of charge? There was nothing they felt like they could cut from that power budget, even in emergencies?
Yes, voltage drop over a long extension cord is expected when its under load. However, that's not the real issue: the actual problem was the "let's allow a daily 8% charge consumption while quiescent that leads to uncontrolled, fatal discharge, without understanding it is important to be able to shed load" engineering decision Tesla made. Furthermore, even the battery monitor IC's have a low-voltage cutout to protect the battery these days. I guess Tesla couldn't be bothered to follow industry standard practice in terms of power cell protection... better to just drive the $40k battery over the cliff, eh?
I mean, what's the expected, optimal, no-load, self-discharge rate for these cells under typical Tesla owner conditions (ie. garaged, or in California)? 10% per month?
Anyway, if what you said about devices needlessly killing their batteries while plugged in is true, then I hope that the engineers responsible for battery-powered products generally become less... "simpleminded" in terms of their perspective on battery circuits.
This all seems very ass-backwards in terms of design. Why drain from the battery while it's plugged in? My HTC Incredible has taken this brain damage even further: the battery completes the circuit and the phone literally can't boot unless a battery is in place. If you are referring to modern devices like the Nexus 10 whose drain exceeds the charger supply, that's one thing.
However, let's talk about laptops then: there's no excuse for a laptop power supply to be less than the worst case draw of the laptop. As for cutting the battery out of the circuit: can't a power MOSFET (serving as a solid state relay) handle that much current? I seem to recall that Rds has gotten low enough that all the laptop's current wouldn't pose an issue... and switching the FET should take milliseconds or less, so power interruption wouldn't be an issue.
Also, I'm not sure that citing Telsa as an example of battery engineering is a good idea... considering they self-destruct their batteries if they aren't constantly plugged in. Hell, one $40k Telsa battery pack died because the owner had plugged it in via a 100+ foot extension cord. Sure, I^2R losses, but exactly how much power does the damn thing need to keep its head above water while it is sitting around, doing nothing?
Seriously now, I thought modern lithium technology had excellent charge retention. So, are they honestly claiming that they can't figure out how to shed vampire load on the goddamn car in state-of-charge emergencies and therefore just have no choice but to deplete the cells below 2.7V (or whatever)? It just can't be that hard... stop charge balancing if state of charge gets low, stop powering all the telecomms/gps/logging/computer crap, and the quiescent battery should be fine for a Long Time(tm), right?
I guess I am still annoyed at my Thinkpad killing my battery even though I practically never used it unplugged. Then one day I found it dropped charge from 100% to 0% in 3 minutes whenever it was on battery. Basically, what you're saying is that this could largely have been avoided if the engineers had designed the circuit to be able to cut the battery out of the loop while it was charged and the laptop plugged into wall power, right?
...our planet is heating up so fast that by some estimates, in another 20 years we won't have ice in antarctica.
That's a fairly alarmist claim. Where did you read that? Because, you know, it seems like the risk of 200+ foot rise in sea levels due to *all* the antarctic ice melting within 20 years might be getting a little more attention.
Ironically, Mitsubishi will probably be building and selling wind turbines as part of this deal.
...for certain, non-ironic values of "ironically".
Unless you count Ally Bank (nee GMAC) financing the sales of GM vehicles as irony. Or Toyota Finance offering loans to Toyota vehicle buyers.
There are countless examples. Hell, even the Marshall Plan was similar in concept—and that might be considered ironic (considering that US aid was given to reconstruct Japan after WWII).
we're not THAT far away (50 years? 100 years?) from an era in which people could credibly create their own nuclear or bioweapons.
I'm sure that in 2143 plutonium is available in every corner drug store, but here in 2113 it's a little hard to to come by...
You want a meritocratic, agile, decentralized Union that represents a Hacker/techno-utopian ethos, with blackjack and hookers and bitcoin micropayments? Then make that union!
Well, that would take about pi microseconds to collapse from the metastable state you describe into one of the binary stable states of either a "mediocracy tyranny" or Lord of the Flies.
Because if you give the unionized team the ability to choose who to hire and fire (hey, this is a meritocracy, right?) then the floors of many cubicle halls will run red with the blood of the unpopular developers. Even in a purely democratic collective, witch hunts leading to termination will become commonplace. I know, because I have witnessed developers "marked for death" in standard corporate environments where it's typically hard to convince the management to cull those considered "dead weight".
When management finally gave in, it led to a Reign of Terror akin to the French Revolution. There were multiple terminations on the "hit list" that were executed, before the purge finally turned inward and culminated in the termination of the highly-skilled developer who acted the part of Robespierre in organizing the terminations of the unworthy. He was eventually devoured by his own purge that had grown out of control, as it were.
Or, like I said, it could devolve into mediocracy where union developers threaten to file a grievance if someone in management makes their own Excel spreadsheet without a union "brother" developer involved. And no one gets fired, ever, except due to layoffs and then only in reverse seniority order. This is outwardly a very tranquil, boring, stable state, but that is merely the velvet glove over the iron fist of mediocrity that crushes developers' souls.
So, a union could start out like you describe, but ultimately I don't believe it is a stable state. But damn, maybe I'm just jaded by experience. Feel free to prove me wrong by setting up an awesome union like you describe and be sure to publicly mark the anniversaries of it running smoothly (in order to get the word out). It would be awesome to be wrong about human nature in this regard—I'm fairly certain I'm correct, though...
I agree with this, however it neglects the actual intrinsic value of Bitcoin, which is its putative secrecy for private transactions.
Putative or not, it is illusory. One might just as well extol Slashdot as a forum with putative anonymity for untraceable comments.
In both cases, practical attacks exist to unmask the parties involved unless special preventative measures are taken. Eg. for bitcoin, laundering; for Slashdot, proxies/VPN/Tor/etc.
Bear in mind that there is a limit to "secrecy" in any system that requires that the ledger of all history of every transaction be publicly available and downloaded by every client. This means that transactions can be traced from a known wallet address through the chain, much in the same way that the US ATF can putatively trace firearms based on serial numbers.
Last I checked (in February) the blockchain data was several GB on disk. Per client. And took days to sync. "Thanks, Satoshi Dice!" (ugh, talk about crapflooding the blockchain...)
It's not like the FORTY-FIVE year old Ponce has a nuclear reactor on it.
The laser will only last as long as they can provide power to it, which I am sure is substantial.
* Unlimited ammunition subject to Fair & Acceptable Use policy, which is available online and subject to revision at any time without notice
But this is why I've blocked all Linux articles and won't be dealing with their shit anymore, have enough on my plate without wasting my time dealing with crazies.
Hey, I saw this while expanding the tree of our discussion.
I apologize if our long-running discussion (I thought it was rational rather than insane, but haha, okay?) was "the last straw" that soured you to the point of blocking an entire segment of Slashdot. I felt no acrimony toward you throughout the discussion, and I am actually sorry if you felt otherwise.
Best wishes for your health issues—I have no clue what you use to sleep, but personally, I find doxylamine succinate (aka. "Unisom 1") to be a powerful, benign OTC sleep aid vastly superior to the diphenhydramine types (aka. Benadryl, "Unisom 2", Simply Sleep, Tylenol PM [ugh, with acetaminophen, despite most users just wanting to use it as a sleep aid]). Incidentally, both drugs are antihistamines so they can double as anti-allergy drugs (though doxylamine is more powerful).
Good luck, hope you heal soon.
Dude if you want to run crypto that the E350 won't run and can't find the card? Fine get an A-Series, you can get one shipped for like $65 for a triple that'll run any damned thing you want. Hell if you don't mind getting an older model you can get E series ULV quad for $68.
Right, pretty much any desktop chip will perform adequately. However, those examples both increase the power budget to a 65 watt TDP for the CPU, which doesn't really fit into low-power system.
IIRC I found 'em by looking up "cell accelerator board"
Okay, using that search token I found the Mercury boards. Holy crap, you weren't kidding about blowing the power budget with those: the model I read about had a 210 watt TDP with 4 GB of RAM onboard, etc.
But again this whole thread started because somebody said the Pi could do it which surely to God even you will admit that isn't possible
Was this what you were upset about? I never contested that... it's an absurd notion. All I was trying to do with your initial comment was add a slight disclaimer to the effect of:
* E-350 may not be sufficiently powerful to drive a simple crypto filesystem for encrypted storage while simultaneously processing a dashcam video source.
Though, this being Slashdot, perhaps someone else might come along and incorrectly interpret that to be an endorsement of the RPi for this usage. Haha, okay, so I will add a meta-disclaimer: .
** E-350 may or may not be adequate, but the RPi, a pocket calculator, or a scurry of squirrels using abacuses are certainly not adequate
(IIRC its designed primarily for AES but I'm sure others will run)
Current hardware acceleration for crypto typically revolves around AES (for symmetric/block), RSA (for asymmetric), and/or SHA-1/SHA-2 (for signing/hashing). The reason AES was selected for acceleration is probably due to the modified Rijndael algorithm that became AES having excellent algorithmic efficiency (that's one reason it won the competition). For example, the AES-NI acceleration in Intel/AMD CPUs only adds seven new instructions (all specifically related to fundamental AES operations). It is unlikely that those instructions could be used to accelerate, say, Blowfish, so other algorithms tend to require execution via generic CPU instructions and thus would achieve no speedup. This is also why the addition of AES-NI acceleration increases performance by 10x (or more) on the same silicon.
AES is certainly good enough for security, and is likely to have the best performance across the gamut of hardware, so it's typically the best choice unless someone really wants to go all crypto-nerd. And it's much easier to deploy than something using potentially-faster, obsolete crypto like DES, because none of the crypto filesystems offer that as an option anymore (if they ever did).
THAT solves ALL of your problems with the E350 since Via has hardware accelerated crypto baked into the chip and has since the C7. You can pick up one of those for less than $200, use its hardware acceleration to do any crypto you want and there ya go, Bob's your uncle.
Looks good. A comparative benchmark suggests that the C7 can probably hit 40 megabytes/sec for AES-128-CBC (ie. "lowest grade AES encryption") while accelerated. Crucially, having hardware acceleration instructions for AES means that it shouldn't require 100% of the CPU in order to achieve that throughput. So, one might expect a dashcam carputer with a C7 to be able to handle encrypted storage throughput sufficient to stream data to disk while also having enough capacity to do "other stuff" related to the dashcam video source. And not break the bank or be a space heater.
Cool. I'm glad you brought that up and that we continued the discussion long after everyone else left.
Hell this whole thread was started by some guy saying a fucking raspberry pi could do the job, why aren't you bitching at him?
The reason I'm not "bitching" at anyone else is due to your response to my original comment. To refresh your memory, this was my initial comment: "Just an FYI: the maximum throughput of an E-350 to an SSD encrypted with AES-128 CBC (4K sectors) is on the order of 30 megabytes/sec... and that's with full CPU use for the crypto (i.e. the system is doing nothing else)."
This is a simple fact derived from repeated testing (by myself and others). You superciliously attacked this as a crypto geek overthinking the problem, which led to my attempts to explain. I really don't know what your original disconnect was on that, because eventually you suggested using TrueCrypt (a crypto filesystem with stronger encryption than what I cited above) as a "simple" solution...
As I said you have options, there are several E350 boards that give you a full X16 PCI-E that is more than enough to support a drop in cell card or FPGA and while I haven't looked at FPGA crypto in ages IIRC the cell card (which FYI can be had for around $100-$150 last i looked but its been awhile, you might have to ebay it) supported up to 1024bit in just about every format you can name OOTB so then you are only using the E350 for basic tasks like memory management and letting the cell do the heavy lifting.
Okay, I just spent 15 minutes scouring eBay and elsewhere looking for a crypto card like you describe. Needs to be a PCI-Express card with AES support, and have an AES throughput better than 250 Mbps (which is approximately what you can get with the E-350 CPU by itself). Oh, and it can't draw more than, say, another 17 watts or cost more than, say, $150 (used on eBay is fine). If you have a link to something like you describe then that would be appreciated; I might actually buy it.
Like I said, you can buy a $200 (brand new), 17 watt TDP Xeon E3-1220LV2 *today* that has >1 Gbps AES crypto filesystem performance without having to blow your money & power budget on something expensive like a crypto accelerator that does NSA-grade crypto at multigigabit speeds.
if you went to the Open-CL forums and told them what you needed it to do you could probably code it yourself to run on a $65 AMD A4-3400
Jesus, and you accused me of overcomplicating things? You are actually proffering the idea that it is reasonable to code your own GPU-accelerated crypto filesystem (of which zero examples exist in the real world) for a project like this? Even knowing that a sub-$300 Xeon-based system can handle everything in a ~30 watt complete system using widely-available, well-tested software like TrueCrypt, LUKS, geli, etc, etc?
Oh, hey, "Hoss"... I found this pithy XKCD comic that reminds me of you: XKCD: Security... you know, some crypto geeks out there really overthink security. Haha, don't overthink this.
Oh and just FYI there won't be any consumer Jaguar chips, at least not for another 3-4 years
That's too bad... the damn rumor mills had piqued my interest. I guess I will spec my next low-power builds to use a low-power Xeon instead. That's probably only about $75 more—at ~$250 for the Xeon CPU+motherboard—than I paid for my E35MI-I Deluxe when it came out. Graphics performance is irrelevant for all of my uses, so that's not a comparison point I weight.
You get a lot of mileage of your E-350 installs; that's great for you. However, they may not be fit for purpose in the case of needing to use crypto filesystems (TrueCrypt, etc) with even low-grade encryption... that's what my original comment was about. This is a valid consideration if someone were planning to build/use an E-350 with the intention of using even simple, widely-available crypto, which fits with this overall subject. Or if someone stumbled across your orig
Ah, but you aren't thinking degenerately enough. Have you seen the Pulse Audio diagram? ALSA is a source *and* a sink in the PulseAudio abstraction layer—what a great design!
In the spirit of PulseAudio, I imagine a typical Unix design approach for a de novo "universal" framework to allow application remoting would be a meta-layer on *top* of both X and Wayland (and/or any other framework that comes along). And, for platform portability, it could render widgets using Java Swing.
As long as I'm guaranteed that my local display can talk to remote applications, no matter what the flavor of the month is, I'll be happy. There needs to be a universal standard.
You do realize you probably prompted some brain-addled dev someplace to come up with the equivalent of PulseAudio for display technology, right?
All problems in computer science can be solved by another level of indirection—except for the problem of too many layers of indirection.
—David Wheeler
But when you are talking about running something off a car you better be VERY picky on the power budget or have a hell of a battery, that is all I'm saying.
Per TFS, this whole context is about dashcams, so the entire topic is implicitly about vehicular use. FYI, in case you ever consider building a vehicle-based computer, these tiny DC-DC vehicle PC power supplies support being wired to toggle based on the ignition state and usually come with a low-battery cutoff protection for the vehicle. So, no, you don't need a hell of a battery or an expensive setup. A vehicle can power a 30 watt system like that with zero risk.
Hell a Truecrypt volume can do THAT, that ain't brain surgery.
Do you want to know what my beef is, "Hoss"? It's that when I originally commented about the performance of crypto filesystems (what do you think TrueCrypt is?) on the E-350, you mocked the concept as a crytpo geek overthinking the issue. So, what's this you're saying here? Should I link you to that XKCD comic, "Hoss"? What I was proposing was using a simple crypto filesystem package with *lower* grade encryption than TrueCrypt can do, because, as we agree, it doesn't need to keep out the NSA.
Let me put it very simply for you: on the lowest encryption grade TrueCrypt offers, you can expect a TrueCrypt volume on an E-350 to have about a 20 MB/sec write throughput while TrueCrypt pegs the CPU. Maybe that's good enough for something like a dashcam system, maybe not. If the E-350 had AES-NI (like its successor probably will), then the same configuration would have ~90 MB/sec throughput with 20% CPU usage with the same 17 watt TDP.
It's a false dichotomy to suggest that this is a choice between an inexpensive, low-TDP system with minimally-acceptable crypto filesystem performance vs a hellishly-expensive quad-core "space heater" type CPU plus a crypto FPGA PCIe card.
But do what you like.