Low quality plot too. Remember the eagles? Have them grab the ring and drop it into the volcano. Ta Da all done. Shrinks the trilogy down to about three pages.
I liked the series, but as a ultra loquacious fantasy version of Herodotus Histories or The Odyssey its not really all that great. The originals were better.
It should be cut in half, separating the provisioners from the content providers. One company runs the cable and another provides the tv channels. One runs the wire, and another provides the dial tone. One runs the fiberoptics, another provides the internet. One provides the cellular network, another provides the phones for it.
How would that create a confuseopoly where the megacorps can screw over the customers using their monopoly power and the laws they purchased thru election campaigns?
Its futile, like trying to find a more moral and ethical business plan for vampires or mosquitos or leaches. The way to win is not to play the game.
Yes, it is like selling a fuel-wasting car and then forcing the consumer to purchase fuel from you and only you. And advertising the fuel inefficiency as a feature. And rationing the fuel and switching from unlimited fuel to rationed fuel... ok maybe the analogy breaks down somewhere around there.
I have a better standard/. car analogy. WHAT IF my local car stealership's service dept intentionally had only one mechanic to make all warranty and recall repairs, so as to boost profits, so car service was excruciatingly slow, but as a PR move to avoid hiring more wrenches, they "discovered" that 1% of car owners made up to 90% of service appointments (because they have a lemon or whatever)? So now we can control the car owners as such: 1) they might be one of the 1% high users so they better shut up instead of complaining about slow service, or they might get cut off from all contracted service, or something similarly illogical. 2) we can get the users blaming each other for making service appointments instead of blaming the company for not hiring more wrenches. 3) The stockholm syndrome victims will blame themselves or their fellow drivers or anyone other than the stealership who is ripping them off 4) The guys on/. will complain, but since there is a govt controlled monopoly / confuseopoly, I guess they're just screwed and will have to bend over and take it anyway.
Oh and another one, are they actually using apps or is this apps that are updating? I used to always dread seeing Battle for Wesnoth update on my ipod touch because here comes a third of a gig each update. Are there any apps out there bigger than wesnoth? I know the xplane flight simulators are a bit on the large side.
Could it be a phone that is broken and continually downloading over the air updates over and over and over and over?
Doin what? Until you answer that you're just spinning wheels.
Is there some kind of spam sending virus out there? That would make sense and you could hope they'll fix it.
Are they spending a lot of time at websites? More than 10 or maybe 15 years ago now, Akamai fixed that, maybe the mobiles need that?
Is it one specific app, like google maps?
Is it tethering people trying to run an entire disaster recovery site over a phone?
Does it really matter? Supposedly 1% of the population, that being teen girls, made up most of the call volume at one time. So?
How does their battery survive this intense use? My new android phone barely lives thru the day with light use, so they must be living on a charger?
Why are they "monsters"? What a weird way to describe human beings. That means I should use my leet skyrim skills and cast an ice spear at them, right?
combines electrodes, silver nanoparticles, and salmon DNA.... DNA could turn out to be a less expensive alternative to traditional inorganic materials such as silicon.
could turn out = Weasel words. After I arrive at home, it could turn out that space aliens have swapped my wife out for a supermodel as part of an alien sociology research study regarding recreational human reproductive activities, but I'm thinking its unlikely.
Have you seen the price of salmon? I had a nice grilled slab last night wrapped in some herb leaves and lemon juice. I could buy quite the stack of I2C flash memory chips for that price. I'm not thinking that the salmon-flashdrive equivalent of the HHGTTG babel-fish is necessarily going to be profitable. And carrying around a dead fish with firefox installed on it sounds like some Stross Laundry series plot.
I don't mean for running the code, I mean for development work. I can't see many people sitting for hours coding on the Pi when they know they've got a proper computer on the next table. It will be just too painful.
Again, why? What development environment are you using that uses more than a thousandth the power of your desktop for educational-scale labs?
Will the vim editor magically know its on a pi and magically only work at 5 wpm? Don't think so...
The problem with dev on one box and run on another is now you have to debug it to the peculiar characteristics of two boxes and environments.
Further, if warming trends were to continue, the grain belt of the midwestern US would stretch up into Canada, potentially doubling the population support capacity of the farms of North America, to say nothing of those of Russia.
Please note that moving the "grain belt latitudes" into say, central america, would somewhat reduce our total grain production simply due to lack of land.
One requirement to having ice ages is having a lot of land at high latitudes, which means a huge food crunch during ice ages. There's just less biomass.
Being flooded sucks, but at least theoretically florida could be a nice fishery.
Starve to death or build another city... I'm going with the city.
Oh and I forgot that to a first approximation for thru hole a "soldering iron" is also a "desoldering iron" but for SMD that doesn't work. I hand solder but for desoldering its either elaborate machined tips to hit all pads at once or hot air time. Hot air is a PITA when it melts the "wrong" nearby components.
That's a hurdle that old time thru-holers need to get over, for SMD, until you get good at it, a soldering iron is NOT also a desoldering iron. I would still argue that noobs probably make 100 connections for every one they remove, so I think they still come out ahead even if desoldering is twice as hard for SMD. So 99 connections are twice as easy, and 1 removal is twice as hard, I think SMD overall wins.
In summary, the "flowchart" you need to teach thru-holers is much larger and has many more branches than the "flowchart" you need to teach noob SMDers, and conveniently for the old timers SMD is much faster than thruhole and conveniently it seems harder to make a poor SMD connection than thru-hole so average connection quality is higher. SMD for the win, pretty much.
I would with all respect claim its the other way around. You can teach a noob to be a good SMD hand solderer in about half the time as you can teach a noob to be a thru-hole solderer. Its just easier to do.
A major part of SMD soldering is the mechanical connection is the electrical connection. Unlike thruhole, a bad/missing/horrible soldering job means the part falls off the board, more or less. Also the attention to getting just the right amount of solder for a decent electrical connection means you must have a good mechanical connection. Surface tension / smooth fillet HAS to be good for SMD to work at all, unlike thru hole where any glob on the connection is sorta OK. You can foul up electrically or mechanically with thru-hole, so you gotta teach both. With SMD a good electrical connection is a good mechanical connection there is only one thing to teach.
Another area is dealing with solder bridges. I'm a fan of ye olde wipe technique for SOIC packages. Thru hole never learns to handle bridges or freaks out and/or doesn't get as much experience, whereas its a part of the SMD process so its always handled. A solder bridge on thru hole means an exotic rare failure mode failed to be corrected, so you can see how noobs fail there. A solder bridge on SMD is an inherent part of the process so noobs learn how to deal with it.
I would claim that thru hole is sort of analog in that joint quality is variable and it's hard for a total noob to catch on. Whereas SMD is pretty digital, either its dang near perfect or you are in miserable failure mode (like the chip cap stuck to the iron via surface tension or a failure like that)
I used a non-thermostat iron for SMD work from the 80s up till '09 and had no problems. I did get a ridiculous luxurious Hakko pro grade system around '09 that cost about one or two car payments (insert "ooh" "ahh" sounds from people who know who Hakko is and what they make) and it is slightly easier, and it heats up practically instantly like 20 seconds, and it never overheats so the flux never burns so the tips last forever it seems and never need retinning, and it seems weightless in my fingers because it is probably 1/10th the weight of a radio shack iron, and the cord is superflex like headphone wire so it never drags the iron, but its amazing ergonomic abilities are hardly required to make a decent connection.
You do need a good spudger, in a crisis a wood toothpick would probably work, again over the years I got a little ESD protected thing that cost more than a nice restaurant lunch, but I made do with toothpicks for about 20 years first. You initially stick exactly one pin or terminal to the pad, then reheat and spudge the device perfectly into position, then solder all the other pins/pads while cleaning up the first pad as your last job. Iron in right hand, a couple inches of solder between left pointer and index, spudger between left thumb and pointer. It takes less than 2 seconds once you know what you're doing. If I had 4 hands it would be a lot simpler to do and explain.
Also you need decent solder with a good flux... If its all corroded and icky looking its not gonna do. If its multicore that seems to help. Smaller diameter the better when it comes to controlling solder fillet size. If its got "radio shack" on the label its probably not the right stuff. And a little bottle / pen / can / whatever floats your boat of electrical-type flux helps. Finally this influences flux remover, I seem to end up using isopropyl as a cleaner even for psuedo-water cleanup fluxes. "no clean" is a marketing term doesn't really happen.
Sorry, but SMD is just easier, both for old timers and noobs.
I do agree there is no "hacking" reuse of smd parts. You desolder and throw away. One time use thats all you get. too much lead bending, too much thermal damage. If you try to reuse parts you'll just swear and swear and swear, unlike thru-hole.
Its right up there with "PL-259s are impossible to install"
Well the problem with PL-259s are most people have undersized soldering irons. Not to mention there are better connectors out there nowadays anyways.
The killer with PL-259s is/was two fold:
1) I use ladder line and change is too scary to contemplate so I'll use ladder line until I die. Lets face it most of those old timers have kicked the bucket so thats why we're not subjected to as much "pl-259 sucks" as we used to be.
2) The old ham adage that hams make their own antenna wire by grabbing each side of a penny and pulling really hard remains true. You really can't reuse a PL-259. Yeah you can try, but its agony. If its new and shiny and licks up solder its about 10 dB easier to attach than some tarnished corroded old POS that was salvaged using a torch.
Anybody who thinks it will compare well to the desktop PC they normally use is deluded.
Oh, of course it will compare well. Hmm so lets consider the stereotypical educational institution lab task of writing a embedded thermostat. This 10 ohm resistor epoxied to this thermistor will be heated up precisely to happy tropical fish temp... This sounds like a very stereotypical embedded lab, I think I did something like this in school back in '93 using a then new motorola microcontroller. This is a very stereotypical "3 hour lab" type of task.
The traditional microcontroller method would be to get down in the weeds and write your own floating point routines. A somewhat more modern MC method would be to link in the floating point routine and write it in C. Pi style would be to write a full featured PID control loop in octave or maybe python (or ruby, or scheme, or intercal, or perl, in rough order of readability) and then just run it interpreted style probably as once per minute cron job that sets the next minute's PWM percentage for the resistor based on the historical record of thermistor temperature. You'll probably spend more time fooling with the PID coefficients than doing hardware or software, well thats just life I guess.
My point is when you're doing something embedded its vital that you have at least 100% of the CPU power you require. But if the Pi has a million times the power you require, it really doesn't matter if in comparison your desktop has a billion times the power you require. You've got massive overkill either way.
If you are properly engineering to use the cheapest lowest power system that can still do the job, you'd be down in the weeds using a PIC 10F series, those things cost only a buck or so and are packaged in six lead SOT-23 packages and draw what amounts to no electrical power at all. (A SOT-23 is about half the size of an uncooked rice grain, depending on your rice, I guess). I've worked with the 10F series, they're certainly a trip, that's for sure, and even they are probably overkill for a lowly thermostat.
I can't think of any situation where Pi + Arduino really makes sense. Other than Gertboard for GPIO, using USB peripherals (e.g. servo driver, etc) with Pi makes more sense.
When it inevitably doesn't work, and I need to do the divide and conquer thing, I can attach the Arduino to my desktop linux box and run the exact same code to prove it's not a CPU power thing, or an obscure Pi USB bug, or an obscure Pi bug in general, and/or the heavy desktop has plenty of power and screen realestate to run dumps etc. Also watching/reverse engineering USB signals is very old and well understood so watching how the Arduino and Pi talk to each other should be childs play.
On the other hand the gert does GPIO, so I could emulate that with a bunch of switches and LEDs like a programmers front panel, or write something on a PIC to act like a Pi, but there's no easy way other than maybe parallel printer port to hook a gert up to a non-Pi PC.
Finally if you have the guts to do it you can offload most of the grunt work to the arduino processor. I really don't need the headache of synthesizing real time servo control pulses on the PC, any more than I need the headache of porting octave / R / gnuplot / perl / crontab / ntp into the arduino.
I would agree with you that the interface bus of the future is not the discontinued 1-wire, or SPI, or even the venerable I2C but the bus of the future for embedded work is USB. Theres some pretty good daq stuff out there, some decent motor controllers, USB is just the wave of the future for embedded work. I2C will probably never disappear entirely, but...
As parallel ports become rarer(if not extinct already), I could see this becoming the best way to control CNC/3d printers once somebody ports a RTOS version of Linux to it.
I have a CNC mill in the basement and your theory is pretty much on track for current applied work.
CNC mills are mostly PCI or motherboard printer port controlled right now, no they are not extinct although endangered species is probably appropriate phrase, and semi-intelligent USB attached controllers have been moving into the market for quite a few years.
The problem with the USB controllers is they are generally not too smart and don't handle lookahead/smoothing as well as printer port and real time PC based controllers. In other words on the PC your typical controller, like EMC2, will look ahead some distance and smooth the path so your specified feed rate remains constant, but the simpler USB controllers just try to make every move a smooth trapezoid of speed up, cruise, slow down. This is kind of bad for surface finish and endmill lifetime, in addition to being horribly slow.
If you still do the old fashioned paper magazine thing, there are a couple magazines I'd recommend reading the ads from, like digital machinist and home shop machinist and to a lesser extent machinists workshop.
The hard real time thing isn't as bad as you'd think until you get to exotic industrial levels of performance. If you can play mp3s on the box, you can generate 1s and 0s fast enough for any non-exotic cnc application. HOWEVER those skips and pauses in mp3 playback are not just "shrug shoulders" in CNC work, that means on CNC you just snapped off another $5 endmill and ruined another workpiece (or worse) so it becauses an exercise in how much overkill do you need to guarantee five nines performance... the answer seems to be lot of overkill indeed.
The fairly obvious answer of replacing a real time g-code interpreter with a g-code compiler has not been commercially successful, to the best of my knowledge, because adaptive speed controls, limit switch detection, e-stop detection are all more computationally intensive than simple g-code interpretation so there seems little gain.
LOL everyone else replied the opposite of a theoretical mathematician is an applied mathematician, which is true, but a pretty good synonym for "experimental mathematician" is physicist, and/or some of the more theoretical engineers. Hard science papers crossed over from "mostly text" to "mostly equations/graphs" more than a century ago.
Sagan was a great astronomer, not a physicist, although the crossover point between astrophysicist and astronomer and physicist is not in clear focus (bad pun). Sagan was a giant who stood on the shoulders of giants WRT to popular science writing... Check out Asimov's non-fiction back catalog, written before Sagan's time, its some good stuff.
Now, yes string theory makes predictions, which are currently, maybe even permanently, untestable / unfalsifiable. Too high of energy level, take too long to run, take too much technology or money, etc. That's why they get made fun of a little bit by the real physicists. Creationism with formulae, etc. A perfectly interesting mathematical playground that does not interact with the experimental world is definitely great math but also is not physics. An abstract math idea that does not as far as we currently know contradict currently known physics is not necessarily "truth" or "reality" just by nature of non-contradiction. If they ever find anything they can run a falsifiable predictable experiment on, and it works, hey I'll be the first to toast him/them and celebrate, but till then its just an interesting math idea not a physics idea. And there's nothing wrong with that at all, as long as you categorize it appropriately as exotic math, not physics.
No. It is an arc. If you could somehow measure with enough sensitivity, there would be a single day on which you are ever so slightly better than you will ever be, before or after. But so what? The same is true of anything else, such as your height.
No cognitive noise level is too high, probably the daily standard deviation exceeds 20 IQ points or equivalent.
Back when the kids were newborns and I got no sleep I pretty much just shambled thru the day in a haze. Another good one was back when I was young and drinking was new and exciting, hangover days. Another good one is when I have a cold or flu I may as well sleep all day because nothing's happening in my head.
The opposite effect would be doing something new is great mental exercise. My first microstripline microwave RF amplifier printed circuit board design. The first time I programmed a PIC microcontroller. etc
The sample age was 45-70 and they found that cognitive decline started at 45? Shouldn't they have started sampling people in their 30's to see a better bell curve?
A bell curve? What makes you think the rate of decline ever slows down?
Presumably the first derivative zeros out after death, unless you believe in all this zombie garbage, and we know that with little kids the 1st derivative is positive at least up to teenage years just via common sense, so a bell curve is not entirely unrealistic, if you assume a nice positive 1st d in youth, a leveling off and negative 1st d in adulthood and zeros at birth and death.
Personally I think its more of an impulse response function. School is intellectually challenging, but its mostly at the start of life, so a sharp spike up followed by a lifelong smooth decline to the much lower TV watching level.
I have seen this thru my life, that people with dumb hobbies seem to end up dumb even when they're not doing their hobby, and the opposite is true that people with hobbies that require some brains seem to end up smart in general. This isn't just a thinly veiled "I'm great because I post to/." claim, I'm thinking of relatives with wildly different interests that are technically challenging, like my hot rod engine blueprinter relative, or my relative who was something of an amateur fashion designer (the topological problems of covering a very 3-dimensional woman with 2-dimensional fabric, although expressed totally informally, are actually kind of challenging). Oh and my master welder and general fabricator great uncle at near 90 is more thoughtful than most 40 year olds I know. On the other hand the relative who spent decades behaving as if judge judy is too intellectual is not doing so well mentally. I'm using relatives because I've known them pretty well over a long term. The "my hobby is watching TV" type don't do as well after retirement as the "my hobby is tuning up chevy small blocks" type, even if both of them spend most of their time sitting on their butts.
"Cognitive function" in this instance isn't a measure of "raw processing power":
The Alice Heim 4-I (AH4-I) is composed of a series of 65 verbal and mathematical reasoning items of increasing difficulty.18 It tests inductive reasoning, measuring the ability to identify patterns and infer principles and rules. Participants had 10 minutes to do this section. Short term verbal memory was assessed with a 20 word free recall test. Participants were presented a list of 20 one or two syllable words at two second intervals and were then asked to recall in writing as many of the words in any order within two minutes.
We used two measures of verbal fluency: phonemic and semantic.19 Participants were asked to recall in writing as many words beginning with “S” (phonemic fluency) and as many animal names (semantic fluency) as they could. One minute was allowed for each test; the observed range on these tests was 0-35. Vocabulary was assessed with the Mill Hill vocabulary test,20 used in its multiple choice format, consisting of a list of 33 stimulus words ordered by increasing difficulty and six response choices.
Judgement, in particular, would suffer if one's ability to perform inductive reasoning was impaired.
Combine that with
Disturbingly enough, even the youngest participants started declining immediately
And you get the idea that "most people" do not do this, at all, as soon as they leave school. I'd be surprised if the result of a larger study would be anything other than decline begins at the graduation ceremony. I haven't done anything in that test for quite a few years other than the inductive reasoning, and thats only because I'm a weirdo; most Americans would rather die than think, so I'm sure they would do none of the above.
Ah but teaching, even teaching a cognition heavy top, does not necessarily require much cognition. The noncognition memory way to teach is "why I remember back in '63 another young man just like you making the same mistake with integration by parts and what I told him back in '63 was..." Then there is the non-cognition cheerleader way to teach which just amounts to telling you that you can do it. And the non-cognition drill sgt way to teach is just telling you that you will do it.
Calc hasn't changed much in a couple hundred years, at least at the undergrad level. Now a math teaching job that would require some cognition would be designing a "how to prove Fermats last theorem" class. So do you start with the full modularity theorem even though only the semistable elliptic curves are necessary for FLT and the full modularity theorem was proven after FLT, but maybe you should introduce the full theory as a concept and then go in depth into just semi-stable elliptic curves, or... Now experience does enter into this so you need to correct for that to test pure cognition.
Who are these people who keep on insisting on using through-hole components? That board could easily be the same size as the Raspberry-pi board itself simply by using SOIC packages as opposed to DIP for all of the ICs. Soldering a 1.27mm pitch SMT component is really easy, it takes about the same amount of time as a DIP component, and is much, much, smaller.
Its a meme that just won't die. As a guy who's been doing SMD at home on and off since the 80s for ham radio microwave gear, it gets tiring hearing for about three decades that what I find easy to do and enjoyable is "impossible" and will be the "death of homebrewing" and all that rot. Its right up there with "PL-259s are impossible to install" and "power poles are impossible to install", you only hear about it over and over from the 0.1% of the population who really can't do it.
I'm willing to bet there are some very young hardware hackers on/. right now emulating their elders by rambling about how impossible it is to do SMD at home, despite my experience doing it for years before they were born.
Arduino does not run linux. You can run the editor / compiler / loader on linux, but its just a bare metal controller. The pi literally runs linux, so you develop on board.
Stereotypically, you'd use something like a gert or a usb connected arduino as the hardware interface. Then you'd simply "apt-get install octave" or whatever on the pi, write a tiny little perl script or whatever to talk to your hardware interface, and write your PID loop or whatever in octave script on the pi. Along with probably running a web server on the pi to remotely monitor your PID loop, maybe R and gnuplot to make some pretty graphs and stats of how your PID loop is working.
Of course its more fun to do something much more complicated than a PID loop. Plenty of linux based neural network support, fuzzy logic libraries, etc, just an apt-get install away.
I looked at the pcb pix and saw the SPI interface lines and incorrectly guessed the whole thing runs off one SPI connection, which would be kind of cool, since pretty much every microcontroller made in the past 30 years is either has SPI hardware support or is at least easy to bit bang SPI. So it would not really be a pi board, but a generic board that works with everything that merely has support to directly plug a pi into it.
However I read the comments and the deal is the breakout board brings 18 GPIO ports from the pi, and you wire the GPIO ports however you want to various peripherals at the GPIO level, one of which is a SPI interface port expander, other things you could wire to are the motor drivers, etc.
So its really a mostly GPIO board with exactly one SPI part, not a board run entirely off just one SPI port. For example, if you have an old fashioned parallel port on your PC, plus or minus some level conversions you could wire that up to this board, etc.
The other interesting comment I read was something similar to "if you want arduino shield support on a pi, simply plug an arduino into the USB port and plug the shield into the arduino and talk to the Ardunino using the linux usb drivers", which is brilliantly simple.
LOL thats Hawking himself who wrote "Whether I'm a genius is more open to doubt". He's a humble guy despite all he's done. Basically an anti-politician. That's what would make him a great national leader, if he wanted to do that. Him being smart enough to not want to take a bite of that sh1t sandwich says a lot about the current world situation.
Slashdot Mirror
Tolkien's prose was viewed as low quality.
Low quality plot too. Remember the eagles? Have them grab the ring and drop it into the volcano. Ta Da all done. Shrinks the trilogy down to about three pages.
I liked the series, but as a ultra loquacious fantasy version of Herodotus Histories or The Odyssey its not really all that great. The originals were better.
It should be cut in half, separating the provisioners from the content providers. One company runs the cable and another provides the tv channels. One runs the wire, and another provides the dial tone. One runs the fiberoptics, another provides the internet. One provides the cellular network, another provides the phones for it.
How would that create a confuseopoly where the megacorps can screw over the customers using their monopoly power and the laws they purchased thru election campaigns?
Its futile, like trying to find a more moral and ethical business plan for vampires or mosquitos or leaches. The way to win is not to play the game.
Yes, it is like selling a fuel-wasting car and then forcing the consumer to purchase fuel from you and only you. And advertising the fuel inefficiency as a feature. And rationing the fuel and switching from unlimited fuel to rationed fuel... ok maybe the analogy breaks down somewhere around there.
I have a better standard /. car analogy. WHAT IF my local car stealership's service dept intentionally had only one mechanic to make all warranty and recall repairs, so as to boost profits, so car service was excruciatingly slow, but as a PR move to avoid hiring more wrenches, they "discovered" that 1% of car owners made up to 90% of service appointments (because they have a lemon or whatever)? /. will complain, but since there is a govt controlled monopoly / confuseopoly, I guess they're just screwed and will have to bend over and take it anyway.
So now we can control the car owners as such:
1) they might be one of the 1% high users so they better shut up instead of complaining about slow service, or they might get cut off from all contracted service, or something similarly illogical.
2) we can get the users blaming each other for making service appointments instead of blaming the company for not hiring more wrenches.
3) The stockholm syndrome victims will blame themselves or their fellow drivers or anyone other than the stealership who is ripping them off
4) The guys on
Oh and another one, are they actually using apps or is this apps that are updating? I used to always dread seeing Battle for Wesnoth update on my ipod touch because here comes a third of a gig each update. Are there any apps out there bigger than wesnoth? I know the xplane flight simulators are a bit on the large side.
Could it be a phone that is broken and continually downloading over the air updates over and over and over and over?
And I bet the top 1% of slashdot posters live in 80% of all basements.
Top 1% of posters get 80% of all +5 articles. This is true.
Doin what? Until you answer that you're just spinning wheels.
Is there some kind of spam sending virus out there? That would make sense and you could hope they'll fix it.
Are they spending a lot of time at websites? More than 10 or maybe 15 years ago now, Akamai fixed that, maybe the mobiles need that?
Is it one specific app, like google maps?
Is it tethering people trying to run an entire disaster recovery site over a phone?
Does it really matter? Supposedly 1% of the population, that being teen girls, made up most of the call volume at one time. So?
How does their battery survive this intense use? My new android phone barely lives thru the day with light use, so they must be living on a charger?
Why are they "monsters"? What a weird way to describe human beings. That means I should use my leet skyrim skills and cast an ice spear at them, right?
combines electrodes, silver nanoparticles, and salmon DNA. ... DNA could turn out to be a less expensive alternative to traditional inorganic materials such as silicon.
could turn out = Weasel words. After I arrive at home, it could turn out that space aliens have swapped my wife out for a supermodel as part of an alien sociology research study regarding recreational human reproductive activities, but I'm thinking its unlikely.
Have you seen the price of salmon? I had a nice grilled slab last night wrapped in some herb leaves and lemon juice. I could buy quite the stack of I2C flash memory chips for that price. I'm not thinking that the salmon-flashdrive equivalent of the HHGTTG babel-fish is necessarily going to be profitable. And carrying around a dead fish with firefox installed on it sounds like some Stross Laundry series plot.
I don't mean for running the code, I mean for development work. I can't see many people sitting for hours coding on the Pi when they know they've got a proper computer on the next table. It will be just too painful.
Again, why? What development environment are you using that uses more than a thousandth the power of your desktop for educational-scale labs?
Will the vim editor magically know its on a pi and magically only work at 5 wpm? Don't think so...
The problem with dev on one box and run on another is now you have to debug it to the peculiar characteristics of two boxes and environments.
Further, if warming trends were to continue, the grain belt of the midwestern US would stretch up into Canada, potentially doubling the population support capacity of the farms of North America, to say nothing of those of Russia.
Please note that moving the "grain belt latitudes" into say, central america, would somewhat reduce our total grain production simply due to lack of land.
One requirement to having ice ages is having a lot of land at high latitudes, which means a huge food crunch during ice ages. There's just less biomass.
Being flooded sucks, but at least theoretically florida could be a nice fishery.
Starve to death or build another city... I'm going with the city.
Oh and I forgot that to a first approximation for thru hole a "soldering iron" is also a "desoldering iron" but for SMD that doesn't work. I hand solder but for desoldering its either elaborate machined tips to hit all pads at once or hot air time. Hot air is a PITA when it melts the "wrong" nearby components.
That's a hurdle that old time thru-holers need to get over, for SMD, until you get good at it, a soldering iron is NOT also a desoldering iron. I would still argue that noobs probably make 100 connections for every one they remove, so I think they still come out ahead even if desoldering is twice as hard for SMD. So 99 connections are twice as easy, and 1 removal is twice as hard, I think SMD overall wins.
In summary, the "flowchart" you need to teach thru-holers is much larger and has many more branches than the "flowchart" you need to teach noob SMDers, and conveniently for the old timers SMD is much faster than thruhole and conveniently it seems harder to make a poor SMD connection than thru-hole so average connection quality is higher. SMD for the win, pretty much.
I would with all respect claim its the other way around. You can teach a noob to be a good SMD hand solderer in about half the time as you can teach a noob to be a thru-hole solderer. Its just easier to do.
A major part of SMD soldering is the mechanical connection is the electrical connection. Unlike thruhole, a bad/missing/horrible soldering job means the part falls off the board, more or less. Also the attention to getting just the right amount of solder for a decent electrical connection means you must have a good mechanical connection. Surface tension / smooth fillet HAS to be good for SMD to work at all, unlike thru hole where any glob on the connection is sorta OK. You can foul up electrically or mechanically with thru-hole, so you gotta teach both. With SMD a good electrical connection is a good mechanical connection there is only one thing to teach.
Another area is dealing with solder bridges. I'm a fan of ye olde wipe technique for SOIC packages. Thru hole never learns to handle bridges or freaks out and/or doesn't get as much experience, whereas its a part of the SMD process so its always handled. A solder bridge on thru hole means an exotic rare failure mode failed to be corrected, so you can see how noobs fail there. A solder bridge on SMD is an inherent part of the process so noobs learn how to deal with it.
I would claim that thru hole is sort of analog in that joint quality is variable and it's hard for a total noob to catch on. Whereas SMD is pretty digital, either its dang near perfect or you are in miserable failure mode (like the chip cap stuck to the iron via surface tension or a failure like that)
I used a non-thermostat iron for SMD work from the 80s up till '09 and had no problems. I did get a ridiculous luxurious Hakko pro grade system around '09 that cost about one or two car payments (insert "ooh" "ahh" sounds from people who know who Hakko is and what they make) and it is slightly easier, and it heats up practically instantly like 20 seconds, and it never overheats so the flux never burns so the tips last forever it seems and never need retinning, and it seems weightless in my fingers because it is probably 1/10th the weight of a radio shack iron, and the cord is superflex like headphone wire so it never drags the iron, but its amazing ergonomic abilities are hardly required to make a decent connection.
You do need a good spudger, in a crisis a wood toothpick would probably work, again over the years I got a little ESD protected thing that cost more than a nice restaurant lunch, but I made do with toothpicks for about 20 years first. You initially stick exactly one pin or terminal to the pad, then reheat and spudge the device perfectly into position, then solder all the other pins/pads while cleaning up the first pad as your last job. Iron in right hand, a couple inches of solder between left pointer and index, spudger between left thumb and pointer. It takes less than 2 seconds once you know what you're doing. If I had 4 hands it would be a lot simpler to do and explain.
Also you need decent solder with a good flux... If its all corroded and icky looking its not gonna do. If its multicore that seems to help. Smaller diameter the better when it comes to controlling solder fillet size. If its got "radio shack" on the label its probably not the right stuff. And a little bottle / pen / can / whatever floats your boat of electrical-type flux helps. Finally this influences flux remover, I seem to end up using isopropyl as a cleaner even for psuedo-water cleanup fluxes. "no clean" is a marketing term doesn't really happen.
Sorry, but SMD is just easier, both for old timers and noobs.
I do agree there is no "hacking" reuse of smd parts. You desolder and throw away. One time use thats all you get. too much lead bending, too much thermal damage. If you try to reuse parts you'll just swear and swear and swear, unlike thru-hole.
Well the problem with PL-259s are most people have undersized soldering irons. Not to mention there are better connectors out there nowadays anyways.
The killer with PL-259s is/was two fold:
1) I use ladder line and change is too scary to contemplate so I'll use ladder line until I die. Lets face it most of those old timers have kicked the bucket so thats why we're not subjected to as much "pl-259 sucks" as we used to be.
2) The old ham adage that hams make their own antenna wire by grabbing each side of a penny and pulling really hard remains true. You really can't reuse a PL-259. Yeah you can try, but its agony. If its new and shiny and licks up solder its about 10 dB easier to attach than some tarnished corroded old POS that was salvaged using a torch.
Anybody who thinks it will compare well to the desktop PC they normally use is deluded.
Oh, of course it will compare well. Hmm so lets consider the stereotypical educational institution lab task of writing a embedded thermostat. This 10 ohm resistor epoxied to this thermistor will be heated up precisely to happy tropical fish temp... This sounds like a very stereotypical embedded lab, I think I did something like this in school back in '93 using a then new motorola microcontroller. This is a very stereotypical "3 hour lab" type of task.
The traditional microcontroller method would be to get down in the weeds and write your own floating point routines. A somewhat more modern MC method would be to link in the floating point routine and write it in C. Pi style would be to write a full featured PID control loop in octave or maybe python (or ruby, or scheme, or intercal, or perl, in rough order of readability) and then just run it interpreted style probably as once per minute cron job that sets the next minute's PWM percentage for the resistor based on the historical record of thermistor temperature. You'll probably spend more time fooling with the PID coefficients than doing hardware or software, well thats just life I guess.
My point is when you're doing something embedded its vital that you have at least 100% of the CPU power you require. But if the Pi has a million times the power you require, it really doesn't matter if in comparison your desktop has a billion times the power you require. You've got massive overkill either way.
If you are properly engineering to use the cheapest lowest power system that can still do the job, you'd be down in the weeds using a PIC 10F series, those things cost only a buck or so and are packaged in six lead SOT-23 packages and draw what amounts to no electrical power at all. (A SOT-23 is about half the size of an uncooked rice grain, depending on your rice, I guess). I've worked with the 10F series, they're certainly a trip, that's for sure, and even they are probably overkill for a lowly thermostat.
I can't think of any situation where Pi + Arduino really makes sense. Other than Gertboard for GPIO, using USB peripherals (e.g. servo driver, etc) with Pi makes more sense.
When it inevitably doesn't work, and I need to do the divide and conquer thing, I can attach the Arduino to my desktop linux box and run the exact same code to prove it's not a CPU power thing, or an obscure Pi USB bug, or an obscure Pi bug in general, and/or the heavy desktop has plenty of power and screen realestate to run dumps etc. Also watching/reverse engineering USB signals is very old and well understood so watching how the Arduino and Pi talk to each other should be childs play.
On the other hand the gert does GPIO, so I could emulate that with a bunch of switches and LEDs like a programmers front panel, or write something on a PIC to act like a Pi, but there's no easy way other than maybe parallel printer port to hook a gert up to a non-Pi PC.
Finally if you have the guts to do it you can offload most of the grunt work to the arduino processor. I really don't need the headache of synthesizing real time servo control pulses on the PC, any more than I need the headache of porting octave / R / gnuplot / perl / crontab / ntp into the arduino.
I would agree with you that the interface bus of the future is not the discontinued 1-wire, or SPI, or even the venerable I2C but the bus of the future for embedded work is USB. Theres some pretty good daq stuff out there, some decent motor controllers, USB is just the wave of the future for embedded work. I2C will probably never disappear entirely, but...
As parallel ports become rarer(if not extinct already), I could see this becoming the best way to control CNC/3d printers once somebody ports a RTOS version of Linux to it.
I have a CNC mill in the basement and your theory is pretty much on track for current applied work.
CNC mills are mostly PCI or motherboard printer port controlled right now, no they are not extinct although endangered species is probably appropriate phrase, and semi-intelligent USB attached controllers have been moving into the market for quite a few years.
The problem with the USB controllers is they are generally not too smart and don't handle lookahead/smoothing as well as printer port and real time PC based controllers. In other words on the PC your typical controller, like EMC2, will look ahead some distance and smooth the path so your specified feed rate remains constant, but the simpler USB controllers just try to make every move a smooth trapezoid of speed up, cruise, slow down. This is kind of bad for surface finish and endmill lifetime, in addition to being horribly slow.
If you still do the old fashioned paper magazine thing, there are a couple magazines I'd recommend reading the ads from, like digital machinist and home shop machinist and to a lesser extent machinists workshop.
The hard real time thing isn't as bad as you'd think until you get to exotic industrial levels of performance. If you can play mp3s on the box, you can generate 1s and 0s fast enough for any non-exotic cnc application. HOWEVER those skips and pauses in mp3 playback are not just "shrug shoulders" in CNC work, that means on CNC you just snapped off another $5 endmill and ruined another workpiece (or worse) so it becauses an exercise in how much overkill do you need to guarantee five nines performance ... the answer seems to be lot of overkill indeed.
The fairly obvious answer of replacing a real time g-code interpreter with a g-code compiler has not been commercially successful, to the best of my knowledge, because adaptive speed controls, limit switch detection, e-stop detection are all more computationally intensive than simple g-code interpretation so there seems little gain.
LOL everyone else replied the opposite of a theoretical mathematician is an applied mathematician, which is true, but a pretty good synonym for "experimental mathematician" is physicist, and/or some of the more theoretical engineers. Hard science papers crossed over from "mostly text" to "mostly equations/graphs" more than a century ago.
Sagan was a great astronomer, not a physicist, although the crossover point between astrophysicist and astronomer and physicist is not in clear focus (bad pun). Sagan was a giant who stood on the shoulders of giants WRT to popular science writing... Check out Asimov's non-fiction back catalog, written before Sagan's time, its some good stuff.
Now, yes string theory makes predictions, which are currently, maybe even permanently, untestable / unfalsifiable. Too high of energy level, take too long to run, take too much technology or money, etc. That's why they get made fun of a little bit by the real physicists. Creationism with formulae, etc. A perfectly interesting mathematical playground that does not interact with the experimental world is definitely great math but also is not physics. An abstract math idea that does not as far as we currently know contradict currently known physics is not necessarily "truth" or "reality" just by nature of non-contradiction. If they ever find anything they can run a falsifiable predictable experiment on, and it works, hey I'll be the first to toast him/them and celebrate, but till then its just an interesting math idea not a physics idea. And there's nothing wrong with that at all, as long as you categorize it appropriately as exotic math, not physics.
No. It is an arc. If you could somehow measure with enough sensitivity, there would be a single day on which you are ever so slightly better than you will ever be, before or after. But so what? The same is true of anything else, such as your height.
No cognitive noise level is too high, probably the daily standard deviation exceeds 20 IQ points or equivalent.
Back when the kids were newborns and I got no sleep I pretty much just shambled thru the day in a haze. Another good one was back when I was young and drinking was new and exciting, hangover days. Another good one is when I have a cold or flu I may as well sleep all day because nothing's happening in my head.
The opposite effect would be doing something new is great mental exercise. My first microstripline microwave RF amplifier printed circuit board design. The first time I programmed a PIC microcontroller. etc
The sample age was 45-70 and they found that cognitive decline started at 45? Shouldn't they have started sampling people in their 30's to see a better bell curve?
A bell curve? What makes you think the rate of decline ever slows down?
Presumably the first derivative zeros out after death, unless you believe in all this zombie garbage, and we know that with little kids the 1st derivative is positive at least up to teenage years just via common sense, so a bell curve is not entirely unrealistic, if you assume a nice positive 1st d in youth, a leveling off and negative 1st d in adulthood and zeros at birth and death.
Personally I think its more of an impulse response function. School is intellectually challenging, but its mostly at the start of life, so a sharp spike up followed by a lifelong smooth decline to the much lower TV watching level.
I have seen this thru my life, that people with dumb hobbies seem to end up dumb even when they're not doing their hobby, and the opposite is true that people with hobbies that require some brains seem to end up smart in general. This isn't just a thinly veiled "I'm great because I post to /." claim, I'm thinking of relatives with wildly different interests that are technically challenging, like my hot rod engine blueprinter relative, or my relative who was something of an amateur fashion designer (the topological problems of covering a very 3-dimensional woman with 2-dimensional fabric, although expressed totally informally, are actually kind of challenging). Oh and my master welder and general fabricator great uncle at near 90 is more thoughtful than most 40 year olds I know. On the other hand the relative who spent decades behaving as if judge judy is too intellectual is not doing so well mentally. I'm using relatives because I've known them pretty well over a long term. The "my hobby is watching TV" type don't do as well after retirement as the "my hobby is tuning up chevy small blocks" type, even if both of them spend most of their time sitting on their butts.
"Cognitive function" in this instance isn't a measure of "raw processing power":
The Alice Heim 4-I (AH4-I) is composed of a series of 65 verbal and mathematical reasoning items of increasing difficulty.18 It tests inductive reasoning, measuring the ability to identify patterns and infer principles and rules. Participants had 10 minutes to do this section. Short term verbal memory was assessed with a 20 word free recall test. Participants were presented a list of 20 one or two syllable words at two second intervals and were then asked to recall in writing as many of the words in any order within two minutes.
We used two measures of verbal fluency: phonemic and semantic.19 Participants were asked to recall in writing as many words beginning with “S” (phonemic fluency) and as many animal names (semantic fluency) as they could. One minute was allowed for each test; the observed range on these tests was 0-35. Vocabulary was assessed with the Mill Hill vocabulary test,20 used in its multiple choice format, consisting of a list of 33 stimulus words ordered by increasing difficulty and six response choices.
Judgement, in particular, would suffer if one's ability to perform inductive reasoning was impaired.
Combine that with
Disturbingly enough, even the youngest participants started declining immediately
And you get the idea that "most people" do not do this, at all, as soon as they leave school. I'd be surprised if the result of a larger study would be anything other than decline begins at the graduation ceremony. I haven't done anything in that test for quite a few years other than the inductive reasoning, and thats only because I'm a weirdo; most Americans would rather die than think, so I'm sure they would do none of the above.
Use it or lose it.
Ah but teaching, even teaching a cognition heavy top, does not necessarily require much cognition. The noncognition memory way to teach is "why I remember back in '63 another young man just like you making the same mistake with integration by parts and what I told him back in '63 was..." Then there is the non-cognition cheerleader way to teach which just amounts to telling you that you can do it. And the non-cognition drill sgt way to teach is just telling you that you will do it.
Calc hasn't changed much in a couple hundred years, at least at the undergrad level. Now a math teaching job that would require some cognition would be designing a "how to prove Fermats last theorem" class. So do you start with the full modularity theorem even though only the semistable elliptic curves are necessary for FLT and the full modularity theorem was proven after FLT, but maybe you should introduce the full theory as a concept and then go in depth into just semi-stable elliptic curves, or ... Now experience does enter into this so you need to correct for that to test pure cognition.
Who are these people who keep on insisting on using through-hole components? That board could easily be the same size as the Raspberry-pi board itself simply by using SOIC packages as opposed to DIP for all of the ICs. Soldering a 1.27mm pitch SMT component is really easy, it takes about the same amount of time as a DIP component, and is much, much, smaller.
Its a meme that just won't die. As a guy who's been doing SMD at home on and off since the 80s for ham radio microwave gear, it gets tiring hearing for about three decades that what I find easy to do and enjoyable is "impossible" and will be the "death of homebrewing" and all that rot. Its right up there with "PL-259s are impossible to install" and "power poles are impossible to install", you only hear about it over and over from the 0.1% of the population who really can't do it.
I'm willing to bet there are some very young hardware hackers on /. right now emulating their elders by rambling about how impossible it is to do SMD at home, despite my experience doing it for years before they were born.
Arduino does not run linux. You can run the editor / compiler / loader on linux, but its just a bare metal controller.
The pi literally runs linux, so you develop on board.
Stereotypically, you'd use something like a gert or a usb connected arduino as the hardware interface. Then you'd simply "apt-get install octave" or whatever on the pi, write a tiny little perl script or whatever to talk to your hardware interface, and write your PID loop or whatever in octave script on the pi. Along with probably running a web server on the pi to remotely monitor your PID loop, maybe R and gnuplot to make some pretty graphs and stats of how your PID loop is working.
Of course its more fun to do something much more complicated than a PID loop. Plenty of linux based neural network support, fuzzy logic libraries, etc, just an apt-get install away.
I looked at the pcb pix and saw the SPI interface lines and incorrectly guessed the whole thing runs off one SPI connection, which would be kind of cool, since pretty much every microcontroller made in the past 30 years is either has SPI hardware support or is at least easy to bit bang SPI. So it would not really be a pi board, but a generic board that works with everything that merely has support to directly plug a pi into it.
However I read the comments and the deal is the breakout board brings 18 GPIO ports from the pi, and you wire the GPIO ports however you want to various peripherals at the GPIO level, one of which is a SPI interface port expander, other things you could wire to are the motor drivers, etc.
So its really a mostly GPIO board with exactly one SPI part, not a board run entirely off just one SPI port. For example, if you have an old fashioned parallel port on your PC, plus or minus some level conversions you could wire that up to this board, etc.
The other interesting comment I read was something similar to "if you want arduino shield support on a pi, simply plug an arduino into the USB port and plug the shield into the arduino and talk to the Ardunino using the linux usb drivers", which is brilliantly simple.
LOL thats Hawking himself who wrote "Whether I'm a genius is more open to doubt".
He's a humble guy despite all he's done. Basically an anti-politician. That's what would make him a great national leader, if he wanted to do that. Him being smart enough to not want to take a bite of that sh1t sandwich says a lot about the current world situation.