Dan's Data letters #164Publication date: 4 April 2006.
Last modified 03-Dec-2011.
I read what you wrote about on the 30 inch Dell vs Apple monitors. My old man is going blind and wants a BIG screen at work so he can upsize the font. Let's just assume that someone up there is going to foot the bill. Thing is, all the work it is ever going to get is a large Excel spreadsheet. So is it worth going for the extra-hi-resolution monitors, or would something cheaper with smaller res suffice?
What video card would you recommend, given that it's not seeing much 3D work?
A thirty inch (or even larger) screen could therefore definitely be an idea, but there's no point paying for huge resolution.
Ideally, you'd want something like an LCD TV, or even a plasma screen, that can be driven from a PC with no fuss. A "720P" screen has 1280 by 720 resolution, which is perfectly adequate for desktop tasks, and you get some pretty hefty pixels out of that resolution with TV screen sizes. It's easy to find LCD HDTVs with 1366 by 768 resolution for prices considerably below those of the 30 inch super-monitors, and screens a bit bigger. Some of them even give you good old 15 pin analogue VGA inputs
I haven't looked into this field in detail, but to see what I'm talking about, check out the Acer AT3201W. It's about two thirds of the price of the 30 inch Dell.
These screens also have DVI inputs, but the "DVI" they want is DVI-HDTV, which has restricted resolutions and scan rates (I talk about it here).
It's becoming easier to make that happen with modern graphics cards, but you still want to be sure it'll work. Of course, there's nothing wrong with analogue VGA if a screen offers it; there may be a slightly visible quality difference from DVI, but that doesn't mean someone with lousy eyesight will be able to see it.
Any video adapter will work fine for you. Everything in the mainstream market - including adapters built into motherboards - has been more than fast enough for all 2D tasks for years now.
I have a story to share about purchasing memory cards in China that your readers might benefit from.
I just got back from my trip to China where I purchased a SanDisk 1Gb Ultra II CF card for my camera.
Or so I thought. Although it went into the camera and seemed to be fine, after about 250Mb of photos, I got errors writing to the CF and was forced to buy another CF card.
The photos seemed to still be on the camera, but when I got home I had to use an external card reader to access the images (not via the camera's USB cable).
I decided to format the card, and it came back as being only 256Mb!
It looks like I have been swindled. Somehow they had changed the CF partition info to read as a 1Gb (at least in the camera, a Canon IXUS 600).
The card had all the right stickers and even came in a nice SanDisk box!
This was from a small vendor in a crappy little shop just outside the Forbidden City in Beijing.
Just be wary of the cheap cheap CF, and I recommend all people buying from dodgy vendors, try to format the card in their camera before agreeing to buy it.
Not long after Joshua sent me this, a friend of mine showed me the "4Gb" card he bought in Hong Kong, that turned out to be 1Gb.
That card actually formatted in a camera just fine, because that's (usually) a "quick format", where you mark all of the card capacity as empty in the File Allocation Table but don't do anything to the actual memory cells the Table refers to. Three quarters of those cells, in this case, do not actually exist.
I presume these cards have some firmware chicanery going on to make them look bigger than they are. It's easy enough to make storage devices look misleadingly big, though; you can just put dud data in the partition table, for instance. It's often possible to do this with ordinary partitioning software, one way or another. People have been amazing their friends with 1Tb floppy disks for many, many years.
Anyway, you could probably work around the problem in the same way, by re-partitioning the card so that only the real memory is visible. That'd make it safely usable.
Of course, it's also possible that the flash chips on these things were swept off the factory floor after a flood, or something. Do you feel lucky?
(Which, as it turned out, were themselves fake!)
I liked your article about the USB drive death scenario. I have one, and usually keep it switched off not just to conserve the HDD life, but also because if it's on and connected when I get hit by the ravenous MP3/JPG/DOC eating virus of 200x, there isn't much point in having it.
Semi-offline storage - a.k.a. turning your USB drive off - is also half of a decent home or small business backup solution. Make your regular backups to the USB drive, make your weekly-or-less-often ones to DVD+/-R.
The LinkStation does indeed look like a great option, if you only need one drive, and don't mind being forced to pay too much for the one drive it comes with.
All that power and only one drive... they're just teasing us, aren't they?
There's an ad on your site which causes Macromedia's Linux Flash player to completely crap the bed. The same ad seems to appear occasionally on dictionary.com, etc. As Firefox immediately dies, which you might guess is very, very annoying (especially as I have still not broken the "I'm bored, I'll check Dan's Data - crap, now all my tabs are forfeit" habit, which I guess is a compliment to you).
Sure enough, I fired up Ethereal with the intent of identifying the offending HTML-fragment, which sent that ad into hiding-- I've not seen it in dozens of refreshes!
Without knowing which ad or ads is or are the culprit, there is of course not a lot I can do to fix this problem. As usual, though, my advice is: Screw it. Just block the lot of 'em.
As I've said on previous occasions, I'm perfectly fine with my readers blocking some or all of my ads. I certainly block the darn things.
Whichever ad it is that's honking you off is brought to you by the fine people at
Burst Media. If you want to block their stuff via
ye olde hosts file entries, the hostnames to route to 127.0.0.1
or whatever are
It's not a criticality accident if you do it on purpose
OK, so I've got 60% of a critical mass of plutonium in my left hand, and another 60% in my right hand, and I smash them together.
You die, but not as spectacularly as you'd hoped.
And you don't take half of the city with you, although dudes in moonsuits will probably have a lot of work to do in the area.
Making a fission bomb is, I'm happy to say, difficult.
One of the many problems involved is keeping the lump of fissionable material together long enough for a nuclear explosion to actually take place. You only have to hold your lump of plutonium, or whatever, together for a fraction of a second, but that's harder than it sounds.
A-bombs (and the initiating A-bomb stages of H-bombs) work their magic by achieving prompt supercriticality, in which there are enough "prompt neutrons" emitted at the moment each atom splits (as opposed to the "delayed neutrons" emitted when the elements that're products of the fission decay to other elements) to, on average, split more than one other atom. Result, a mouse-traps-and-ping-pong-balls exponentially increasing chain reaction, in which a huge amount of energy is released in an extremely brief time. And there's your instant sunshine.
This sounds like a simple enough project for an aspirational psychopath with some time on his or her hands and a bunch of plutonium bought on the Internet, but there's a small problem and a big problem.
The small problem is the shape of the mass of fissionable material. Neutrons don't steer themselves towards fissionable atoms; they just shoot off in random directions when an atom splits. If they're on the surface of your fissionable lump and happen to shoot off away from it, they won't hit any of the nuclei you want them to hit. If you for some reason decided to beat your plutonium out into a millimetre-thick sheet, you wouldn't be able to achieve criticality even if you covered a football field with the stuff.
For this reason, the best bomb core shape is a sphere - minimum surface area to mass ratio. A-bombs also have neutron reflecting material, such as silicon carbide, around their core. The situation gets even more complex, when you consider materials like beryllium that emit neutrons when alpha particles hit them.
But a simple couple of large enough lumps of fissionable material will work.
The big problem standing between you and your Mushroom Cloud Cymbals is that even though it only takes a small fraction of a second for a nuclear explosion to happen, a simple prompt-critical lump of fissionable material will be blown apart by the energy it's starting to liberate before it really gets going. So much energy is released so quickly that some of the material flashes to vapour and blasts the lump into pieces, whereupon all of the bits are subcritical again and the excitement is over. You could, technically, say that this is indeed a nuclear explosion, but it's not nearly as impressive as the name suggests.
A-bombs use conventional explosives to blast the two pieces of the core together and hold them there by simple momentum and external pressure while the reaction starts. Bombs also have a dense uranium "tamper" around the core whose purpose is to help hold it together, by expanding in all directions when hit by the first wave of energy from the core; the expanding tamper pushes back in onto the core.
A-bombs also don't just stick together a big ol' lump of fissionable material and leave it to its own devices. Instead, they use a separate neutron-injecting "initiator" to kick the reaction off at exactly the right time for it to match the moment when the core's fused together, giving the kind of neutron flux that would otherwise only be possible when the core was already blowing itself apart.
Even then, though, most of the fissionable material doesn't fission. The "efficiency" of a nuclear bomb is the amount of its fuel that gets converted to energy. In the case of the Hiroshima bomb, that was a miserable 1.5 per cent.
E=MC^2 tells us the energy liberated by total conversion of a given amount of mass to energy. Just one gram of mass converts to 89,875,517,873,682 joules.
Set off one gram of TNT and you get 4184 joules, a bit more than the spec-sheet muzzle energy of three shots from a .44 magnum handgun (or about 1.3 times the energy of one shot from the pistol Harry Callahan would be using today).
A titchy little one gram of mass completely converted into energy therefore equals 21.48 kilotons of TNT.
Getting back to your goodbye-cruel-world experiment, it's this problem that'll cause it to fizzle.
Plutonium is a hard and bouncy metal, and so slapping two chunks of it together by hand won't actually unite them very efficiently in the first place. This fact, plus the sudden energy release blowing them apart, could result in a really unexciting, from a weapons physicist's point of view, release of energy.
There would, however, probably still be a bang.
There would also definitely be a big nasty pulse of radiation that would very probably be enough to kill you, but not right away.
The radiation might be strong enough to similarly doom other people nearby, though the inverse-square law would mean that people wouldn't necessarily get a fatal dose just because they could see you.
Actually, you'd probably survive for a while, and might be able to deliver any last words you'd forgotten. You might even be able to smack your hemispheres together a few more times; they'd get very hot very quickly, but I doubt they'd just melt or be torn apart. You could numb your hands, or wear welding gauntlets or Oinky The Happy Pig oven mitts or something.
Some cleverer arrangement might work better than two simple Pu chunks. One chunk of plutonium and one of tungsten carbide, say, or a tungsten carbide shell you can drop your plutonium into, which could result in the Pu block shooting out like a bullet propelled by its own vapour. The feasibility studies wouldn't be easy, though.
Radiation poisoning, by the way, is one thing that really gives me the heebie-jeebies. Oh, sure, flesh-eating bacteria and hemorrhagic fevers and pretty much anything that causes crepitus are all very horrible, but not a lot of diseases let you think you've recovered before they smash your body like a grape.
(Associated note: Don't commit suicide with acetaminophen or something containing white phosphorus, like old-fashioned rat poison. They both kill your liver, which in turn kills you... eventually. Before that, you wake up in hospital feeling much better. Statistically, you're quite likely to decide you want to live if you survive a suicide attempt, or if someone talks you down from attempting it in the first place. But with a suitable dose of these poisons, your state of mind doesn't matter. You need a liver transplant, or you'll die.)
The fact that some poor schmuck can just pick up a lump of something, take it home and put it on the kitchen table, and give his whole family a lethal radiation dose in the next couple of days, even if none of them pick it up... yeech.
I've got this kilo of tungsten here on my desk. It's heavy enough to be a plausible Cylinder Of Death (plutonium is only 3% denser than tungsten). All I'd have to do is make it nice and warm, then leave it on a physicist's desk, to scare the living crap out of them.
Anyway, surprisingly enough, you probably wouldn't have dropped dead from radiation poisoning from just handling your plutonium, putting it in your smiley-face backpack, carrying it to the orphanage, and so on. Subcritical lumps of plutonium are pretty creepy things - as with other substantially-radioactive materials, they're constantly warm to the touch - but plutonium is primarily an alpha emitter, and alpha radiation can't make it through the layer of dead cells on the surface of your skin.
Plutonium's also a toxic heavy metal, but if you handle it with rubber gloves and don't inhale or ingest any of it, that's not an issue. If you're machining the stuff to get it into the shape you want then you could easily end up internally loaded with it, or start a really funky metal fire. But Pu is not at all the horrifying-invisible-death-stuff you'd expect it to be.
On the plus side, though (from the point of view of the glassy-eyed dude with the lead briefcase, not the point of view of the other people at the Kenny G concert), the bare-sphere (no neutron reflector or other tricks) critical mass for plutonium-239 (the standard good-for-bombs isotope) is only around ten kilograms. So you actually could physically perform this stunt, in theory at least, with a couple of hemispheres that a person of unremarkable strength could lift.
The critical masses for many fissionable elements are much higher. The critical mass for an unreflected sphere of americium, for instance, is around sixty kilograms.
(You'll need about three hundred billion smoke detectors to get that much, by the way.)