Posts by Nigel 11

Mars trips could blind astronauts

Maybe not quite so simples

I think you'd need a hub at the centre as well, for the communications dish that needs to stay pointed at Earth. That dish would have to be moved to the side when you re-join the three sections for orbital manouvering. There would be some weight penalty for the cable, and some fuel cost in generating and then shedding the angular momentum each time you re-join the modules into one.

I doubt any of these is more than an engineering challenge, especially if the G-force needed for keeping astronauts healthy is significantly less than 1G.

The biggest problem would be what to do in a solar storm. A rotating capsule would make it impossible to use the rest of the supplies between the atronauts and the sun as shielding. They could re-join into a single craft and orient it pointed at the sun with the crew quarters in the shadow - but there's that fuel cost every time you have to do it.

Self-planting plant discovered in Brazil

Wrong reason

That's not the why of fruit!

The fruit is intended to tempt an animal to eat it, seeds and all. The seed survives its trip through the animal's digestive system, and is deposited elsewhere, along with a nice little lump of manure. Plants particularly appreciate omnivores ... the manure of an omnivore will be richer in nitrogen.

Tomato seeds survive not just the human digestive tract but the metropolitan sewerage system, and cause serious problems when they arrive and germinate in the filter beds.

I've always wondered what sort of animal it is that eats a mango complete with the massive pip. Water buffalo?

Clones

Lots of plants spread themselves by extending tendrils or roots which then develop into another plant. Some even detach the clone and let it float away down a river. If there isn't a plant which grows clonelets with barbs to hook onto a passing animal's hair, I'll be very surprised.

But these offspring are all clones, and if that was the only form of reproduction, the parasites would catch up and wipe out the species. Plants, like animals, need sexual reproduction to shuffle their genes and keep ahead of the parasites.

Yes, ...

Peanuts do. That was my first thought. In fact the latin mame for the peanut is Arachis hypogaea, hypogaea means "under the earth".

Windows 8 secure boot would 'exclude' Linux

Hypervisors?

No-one has mentioned hypervisors yet.

If you can boot a hypervisor, then you can run LInux, Windows, whatever under it. If you can't, then you are cut off from a lot of technologies that I expect will break out of the datacentre onto the desktop, as network bandwidth and hard disk sizes increase.

To take just one example: if you want to secure your data in a corporate environment, you want the hard disk behind locked doors in the datacentre or a data-safe-closet. Given a Gbit or faster network, that's easy. Boot a hypervisor on the desktop, then boot the disk in the datacentre across the network.

Perhaps the BIOS of the future ought to BE a hypervisor? Just as long as it's open to all client O/Ses, of course.

Crooks push fake anti-virus via Skype calls

If only ....

"most people have wised up to crooks running unsolicited "security scans" that turn up a multitude of bogus problems on their machines."

If only they had.

and

If only they just turned up bogus problems, rather than actively creating very real ones!

Intel preps 15-Watt 'Sandy Bridge' for micro servers

Crap graphics?

I'd expect it to have completely crap SVGA 2D on-chip graphics. So not so much "not available", as "not suitable for". (Of course if it has enough PCI-X lanes you could integrate it with a third-party graphics chip, but you'd probably lose most of the power advantage by doing that)

Malware burrows deep into computer BIOS to escape AV

Offline AV

BIOS write-protect jumper or switch, seconded. Best design would be like the reset switch on desktops - you'd have to hold it down while you powered on the system to enable BIOS flashing, and you couldn't accidentally leave it enabled.

Another insanity is trying to run anti-virus software within a potentially infected and subverted operating system. The right approach would be to boot off a DVD-ROM, download up-to-date virus signatures from the vendor and then scan the disks. Since the on-disk operating system is not active, there is nowhere for a rootkit to hide (except maybe in the BIOS, hence the need for mechanical protection).

Hunt: We'll slightly inconvenience pirate sites

Samizdat

Samizdat - "subversive" literature hand-typed and passed on and re-typed, chain-letter style, played a big part in bringing down the former Soviet Union.

What chance that pirate sites can ever be suppressed, or even inconvenienced?

IBM, 3M glue chips into silicon skyscrapers

Reliability

When did you last have to replace a CPU or a DIMM?

Silicon chips are amazingly reliable. I see a DIMM that failed in service once or twice a year (8 or 16 chips per DIMM, 2 or 4 DIMMs per PC, about 400 PCs). And I suspect most of those failures are with the soldered joints onto the PCB, or with the connector.

I've seen a failed CPU twice in twenty-plus years. (Maybe a few of the old boxes that went straight to the scrap-heap were CPU failures rather than MoBo failures, but either way they'd lasted well into obsolescence).

At this level of reliability, a stack of 100 will still be acceptably reliable. Possibly, more so than the same 100 chips soldered onto a board (which you don't repair anyway in most cases).

Seagate pulls out the biggest hard one in the industry

Not crap

Seagate drives aren't crap. I have quite a number of them (ST31[05]00525AS mostly, 1Tb and 1.5Tb). No particular problems. I've experienced more failed WD drives, though probably not a greater percentage, and I don't look after enough drives to have statistically reliable numbers.

One Seagate drive started reporting an increasing number of errors through SMART (Reallocate Pending). It was in a RAID array so I swapped it out to be safe. Seagate's online warranty process was the best I've ever used, and I had a replacement drive in days.

ALL drive manufacturers have at times shipped batches of problem drives. If you bought many with near-consecutive serial numbers (or many same-spec PCs with near-consecutive serial numbers which amounts to the same thing) then you may experience a high failure rate that does not generalize to other customers whose drives were manufactured a month earlier or later. It's usually not the drive manufacturer's fault that they contain a faulty component.

ALL hard drives that you buy are in some sense prototypes. By the time they've been proved reliable in service (say five years) they are also obsolescent. Accelerated ageing tests can only get a manufacturer so far. Occasionally, the envelope may be pushed too far, and again all manufacturers have occasionally shipped a drive model with less than the hoped for reliability.

The worst experience I ever had was with a batch of Samsung 40Gb drives that went from working to brick in the blink of an eye (4-5 years back). Despite this, I'll gave Samsung the benefit of the doubt, and recent Spinpoints have been fine. Most hard drive problems I've experienced haven't been so suddenly terminal. They've shown signs of distress (I/O errors, or SMART reallocations) and I've managed to rescue all data off the failing drive.

Google are the only organisation I'm aware of that has published drive failure statistics for statistically significant numbers of drives. They said that they could find no evidence that any of the major manufacturers made drives that were significantly more or less reliable than the others. Their problems were with batches, not with manufacturers.

Whenever I'm constructing a RAID-1 (mirror) I always pair drives from two different manufacturers, to minimize the risk of common-mode failure (a bad batch of drives).

US judge tells Levi's to take its Euro problems to Europe

this sounds pathetic even by the merkin standards!

That's basically what the Merkin judge said in legalese, so make that extra-pathetic.

It's all a rip-off

Personally I question the sanity of anyone who thinks that an item of clothing that costs say £4 in generic form from some far-East sweatshop, becomes worth twice, ten times or a hundred times more if it has some bloody label prominently attached to it.

Anyone who thinks otherwise is free to part company with as much of their cash as it takes. Just bear in mind that the more the label says it cost, the more I think "you are a moron" when I see it!

Kernel.org Linux repository rooted in hack attack

Could happen to anyone.

Windows, Mac fans should ask themselves ... if the same thing happened to the kernel repository inside the corporate HQ of your favorite company, would you ever get to hear about it?

More likely, the company would keep the exploit a secret. For similar reasons, you don't know what procedures are in place to detect such an exploit and recover from it. What you do know is that the number of developers eyeballing the code is way smaller. You also know that there are documented cases of the company being informed about a security-realted issue, and choosing to do nothing about it for months or even years until the issue leads to real-world exploitation with malicious intent.

Still feeling smug, looking through your security-by-obscurity glasses?

BTW the greater risk by far is not corruption of the code base by penetrating the repository, but corruption of the code base by corrupting a contributor. Submission of a well-concealed backdoor in a legitimate patch. Again something that could happen to anyone, but probably easier to do to closed-source project with relatively few people having access to the code, than with an open-source project with many times more developers.

Outbound space probe looks back at tiny Earth and Moon

Black and White?

Is that a monochrome picture? If it's in colour, I'm surprised that the Earth doesn't look more blue-green-ish when you average it into a few pixels.

Russian rocket flub threatens to empty ISS

What's the point?

This'll be unpopular, but I have to ask what's the point of the ISS?

It wasn't a waste of money doing it. we've learned quite a lot from the process. But as far as I can tell, one of those things is that there's no particular point left keeping a few men in an orbiting laboratory.

I'd suggest that the money should be spent developing robotics and telepresence (Waldo-onics?), so we can still service useful things like orbital telescopes, without having to lug a human life-support system into orbit.

Sad, but it looks true to me.

Supercomputer and superboffins spot rare baby supernova

type 1a supernova

One way of looking at it, is that a big enough diamond isn't around forever, or even for very long.

Samsung refuses to buy HP's PC business

Spin-off the business ...

and call it Compaq?

Room-temperature brown dwarf spied just 9 light-years off

re: replies really don't need a title

This one doesn't really need a text

Jupiter is a very dark brown dwarf

Jupiter emits more radiation than it recieves from the sun. Fusion at a very low rate is the probable source of the excess heat. Jupiter's core is believed to be mostly hydrogen in its theoretically predicted high-pressure metallic form.

The Earth also emits more heat than it receives. In this case we have good reason to believe that the sources are radioactivity and tidal friction, and possibly also ongoing crystallisation of the Earth's solid inner core from its liquid outer core.

For the Earth, the excees heat may have been the difference between a living planet and a snowball, in the early days when the sun was somewhat cooler and the moon was a lot closer.

Water ...

We don't know much about what was on the surface of the earth before large parts of it got covered with liquid water, but there's no evidence that life can exist without water.

All terrestrial life still extant shares a common basic biochemistry, with features such as RNA coding for proteins built from a common set of amino-acids, ATP energy-transport, lipid membranes, and an aqueous support medium. There must have been simpler life-systems leading up to this system (think scaffolding), but we have no evidence of what it might have been. I think it extremely unlikely that whatever it was, it did not require liquid water to function. Water is a lowest-common-requirement for all the more complex subsystems.

I'm guessing that complex organisms find it hard or impossible to evolve in the atmospheres of gas giants or cool brown dwarfs. So they might remain at the single-celled stage "forever" until the brown dwarf no longer provides liquid water. Or until some exceptionally unlikely event happens, and multicellular or even intelligent life arises in the dark cold tail-end of a dying universe.

BTW if you envisage galaxies as having been "mined-out" by interstellar-scale intelligences, then think of a brown dwarf ejected from its galaxy and drifting forever alone and undetectable through one of the voids in intergalactic space. That would, in fact, be a more stable environment than one stil in a chaotic orbit around the centre of a galaxy.

The most complex form of life on the planet is surely some sort of insect. Butterfly: Egg, caterpillar, chrysalis ... complete dissolution of the caterpillar to a sort of living soup, and re-birth as a butterfly. Or spider-hunting solitary wasp. Somewhere in the egg is a program which allows it to hunt and paralyze spiders without becoming prey, dig a burrow, install the spider, lay an egg. I wish someone could tell me where and how.

Last refuge of life?

Interesting to note that such a "failed star" might become the last refuge of life in a dying universe, tens or hundreds of billions of years from now. Like the sun, they stay warm by nuclear fusion, but at such a low rate that they'll probably be the last places left where liquid water (and therefore life as we know it) can exist.

Russian Progress space truck crashes in Siberia

Statistics

Can't remember what the appropriate statistics are for N experiments with a binary outcome (success | fail) but my gut feeling is that the above is insufficient data to prove (within 2 standard deviations) that the shuttle was significantly more reliable.

Anyone care to supply the maths?

US and Russia to give uranium to ANYONE

Natural Uranium not a risk

There's no risk attached to natural Uranium, above that of it being a somewhat toxic heavy metal like (say) Lead, and somewhat radioactive like Thorium (of which gas-lamp mantles are still made!) .Uranium is pretty useless stuff outside of the nuclear industry, although armour-piercing bullets and shells are often made of depleted uranium (what's left over when the enriched stuff is made) because it's nearly as dense as gold. Refining Uranium refers to getting rid of impurities and creating pure Uranium Oxide "yellowcake" or metal. Enrichment is the worry.

To make an A-bomb or nuclear fuel out of natural Uranium you have to concentrate the U235 isotope, which requires a huge investment in a centrifuge plant and associated technologies. A lesser degree of concentration creates LEU - lightly enriched uranium - which can fuel a reactor. A much greater degree of enrichment is needed to create HEU that can be turned into a bomb. There's no cheap or simple process that can perform this enrichment (thank God), so fears of nuclear terrorism revolve around theft of existing nuclear weapons or bits thereof.

Same reason not many chip companies have fabs.

It's the cost of establishing the facility to make LEU. Cheaper to buy it in from countries that have already built an enrichment plant, as well as less likely to cause your country to be perceived as a threat.

Threats aside, it's the same with high-tech chips using the latest 28nm and 20nm technologies. A fab costs many billions. Debugging a process likewise. It's cheaper and less risky to design your silicon and let one of a few fab operators make it for you, unless your volumes will be enormous.

Not so bad ...

Relying on one monopoly supplier of anything is a bad idea.

However, if there are a number of countries with LEU-production facilities, then it becomes much less of a risk. What chance of simultaneously alienating all of them to the point that they'll *all* renounce their treaty and contractual obligations? And if a state is worried that this might happen, well, what conclusion should one draw from them thinking that way?

Uranium is quite a common metal

There's a lot of Uranium out there. I doubt there's any country that couldn't feasibly mine Uranium within its own borders if it wanted or needed to. Any nation that's not land-locked could also extract it from seawater. No-one does that because it's cheaper to mine it, and many countries don't mine it because it's cheaper to buy it.

Unfortunately ...

Unfortunately, a uranium-fuelled reactor inevitably creates Plutonium as a by-product.

Fortunately, it creates not just the Pu isotope that can be made into bombs, but other isotopes of Pu that are more radioactive and less fissile. Pu chemically separated from a reactor isn't good for making nuclear bombs.

MOX made purely from dismantled warheads could have Pu-239 separated from it. Therefore, before giving it to suspect nations, MOX fuel would beed to be partially used in a reactor or include a blend of mixed-isotope Pu separated from used reactor fuel. Both ways the fuel would be rather "hot" and dangerous to transport. Best to burn up the unwanted cold-war warheads ourself, and offer only LEU to states wanting to build reactors.

Regardless, I'd much rather see warheads turned into reactor fuel, than see them sitting around in storage, waiting to be turned back into warheads.

Here lies /^v.+b$/i

For Buddhists

water.return(water)

Apple injunction startles Samsung

I'm adding Apple to my personal don't-buy list

Ok, Apple. That's it. You've now joined Sony on my personal list of tech companies whose products I will boycott on principle.

Anyone care to join me?

Heavily-looted mobile phone barns issue 'safety first' missives to staff

They ought to be shipped disabled!

A phone ought to be shipped disabled. The retailer could enter its IMEI into an activation database upon sale. At first network connect it would check its own status (sold | stolen) and take appropriate action if stolen. Something like waiting a few days for activation clearance to propagate, and then locking down its own firmware so as to require a return-to-factory reset, or even irreparably burning itself out.

If taken outside the UK it would never be able to check its activation status, so it would never work.

Linus Torvalds dubs GNOME 3 'unholy mess'

Burnt bridges, again

If anyone is criticicising them for trying something new, they shouldn't be.

What they deserve to be flamed for, is telling the world that the completely different and (IMO) horrible new thing is merely a new release of the old thing, and inplementing it in such a way that one cannot run the new thing and the old thing on the same system.

I can install KDE, XFCE, and one release of Gnome on the same system until I decide which I like the best. On a multi-user system I can install them all and let each user make that choice for him/her self.

But I cannot install Gnome 2 and Gnome 3 on the same system. The Gnome folks have deemed the change to be an upgrade, not a new product, even though in effect they've taken away a car and given us a boat. Once a distribution embraces Gnome 3, Gnome 2 becomes un-usable thereon. Which is why I now hate Gnome developers.

The problem is burnt bridges.

What annoyed me most, and presumably what's annoyed Linus most, is that the Gnome developers created Gnome 3 in such a way that you can't install both Gnome 2 and Gnome 3 on the same (probably multi-user) system. They presented the utterly different UI of Gnome 3 as if it were a mere new release of Gnome 2. It was exactly the same as what Microsoft did with Vista - except Microsoft had a financial reason for shafting experienced XP users, whereas with Gnome it must have been something like arrogance and pride.

So yes, I really hope that someone goes back a good release of Gnome 2 and renames all the entities that clash with Gnome 3, creating a "Gnome classic" fork which can then be maintained indefinitely, while Gnome 3 developers carry on pleasuring themselves. "Maintained" shouldn't be a lot of work, because we don't want any radical changes. In particular, if any UI changes are introduced, they should be small and incremental, so that whatever way you are used to working, carries on working.

As for the big picture, one of the strengths of Linux is multiple UIs that you can install and choose between at login. I'm sure Linus isn't flaming the Gnome people because they've created something utterly different, which he hated. He's flaming them because they smashed and burned the old UI that he liked while they were doing it.

I'll use XFCE if I have to - at least it has workspaces - but I'll miss Gnome 2 if it does die rather than getting reincarnated.

SMART unveils smarter, faster, fatter SSDs

Hard to clean?

It *ought* to be simplicity itself. SATA has boasted a drive-erase command for quite some time, that logically writes the whole drive to zero in one command. Because of the nature of the beast, it takes the drive firmware quite a while to actually do that ... but it's non-interruptible by power-failing it, the erasure resumes as soon as power comes back. Also one has to take it on trust that the drive firmware really does also zero the bad sectors that have been reallocated (or not care, since it's beyond most ordinary hackers to read them).

So implement that command on SSD firmware (or use it, if it's there). I'd expect that drive-erase on flash memory ought to be an awful lot faster. If I remember right, a flash erase operation is much faster than either read or write.

12% of UK don't carry cash

Count me out

There's a huge difference between putting most of my transactions (by value) on my credit card (I do) and not carrying any cash at all. I find it very hard to believe that anyone really doesn't.

How do you do an office whip-around with plastic? How do you give a bit of pocket-money to a child? How do kids do sponsored things for charity in future?

And then there's the elephant in the room. The audit trail. Dumb cash doesn't have one, so there's nothing for journalists, private investigators, jealous spouses, bunny-boiling exes, etc. to make trouble with. Say hello to e-cash without any alternative, and (whatever they say), say goodbye to financial privacy.

Yes, I know it's possible to do trail-less e-cash. I'm just quite certain that a few years after the old sort of cash disappears, the new sort will be modified to introduce a trail, to "protect the children" or "help fight terrorism" or any of the other usual suspects.

Patriot Wildfire 120GB Sata 3 SSD

Gbps not the only useful measure

It all depends on what you are doing with your drive, as to whether the maximum transfer rate is a significant issue. If you are doing lots of small transfers then the IOPS measure is the significant one, and the amount of data being transferred won't flatten even a 1.5Gbps SATA.

Booting an OS is quite likely to be an example of the latter. Can't we have time to boot an OS as a real-world benchmark number? (Obviously it has to be the exact same O/S image every time).

Boffins shine 800Mbps wireless network from flashlight

A in ADSL

We all use a broadcast-mostly network. the A in ADSL stands for Asymmetric. Downloads are much faster than uploads. Mostly, that's what people want.

And there's no problem having another transmitter-LED on the device receiving the data for the back-channel. Where multiple devices are sharing one access point, you're probably back to the original CSMA-CD Ethernet design. Sense transmit collisions, back off by a random amount, exponentially increase back-off time after exach repeat collision. It's been done with radio (ALOHA and successors), and in a coaxial pipe (original thickwire and thinwire multi-tap Ethernet).

It sounds perfect for aircraft cabins. There's already a spotlight for each passenger ... replace or augment it with a LED and use it for comms as well, to the entertainment display on the seat-back and the passenger's notebook PC (via a USB lightwave-comms adapter). Deploy a long-distance (focussed beam) variant in the airspace at the top of the cabin as the network backbone. Save a small fortune every flight because a considerable weight of copper wire is eliminated. It might even be retro-fittable.

LOHAN spaceplane project starting to shape up nicely

Use acidified water, and a battery

You can make Hydrogen as needed, by electrolysis of water (and throw the Oxygen away into the atmosphere). Don't store the stuff (apart from in the balloon).

Outdoors, with any breeze, this is safe. There's nowhere for any leaking hydrogen to accumulate and mix with air to an explosive concentration. Indoors, this is a recipe for an explosion.

Hydrogen lift

Slightly better than Helium. H2 molecular weight = 2 (plus a smidgeon for the Deuterium). He molecular / atomic weight 4 (less a smidgeon for the Helium-3). Air average molecular weight about 29. So 2/29 extra lift.

I still think that a smallish rubber or polymer balloon filled with hydrogen in an open outdoors location is safe. Extreme incompetence or a lightning strike would result in a fire burning upwards, ie away from the people. Just put a small exclusion zone around the balloon during the initial fill (until it's bouyant enough to be tethered a fair number of feet above any human heads). Or get it to no-load bouyancy with Helium, float it upwards, switch the fill to Hydrogen.

PC chip sales up, shipments flat in Q2

Which came first?

I'm almost sure I heard "flatlined" first in biz-speak, meaning zero growth (neither expansion nor contraction). Medical use of the term, in TV dramas, for a flat hear monitor trace (=dead) came later. Maybe the medics were using it that way a long time before the TV dramas, and maybe not.

Anyway, what's wrong with a word with two or more meanings?

Murdoch's PIE BOY jailed for six weeks

Ridiculous

A lot of people don't go to jail for causing actual bodily harm, and this guy gets six weeks for injuring Murdoch's pride and messing up his clothes, with clear intent NOT to cause actual injury.

Even so, he's probably laughing all the way to the bank. They've made him into a martyr. If this can't make his career as a comedian take off, nothing can.

The loser? Us, the taxpayers and victims of crime, paying to keep a harmess idiot in jail. And to make space for him, there's a rapist or armed robber again prowling our streets.

Good news: A meltdown would kill fewer than we thought

Three?

What are you counting?

The two (or three?) at Fukushima were all caused by power loss. Three Mile Island is one that wasn't. Chernobyl wasn't a melt-down, it was an explosion and fire (and is irrelevant because only the Soviets were crazy enough to build that sort of reactor).

Anyway, the more important thing is containment. Three Mile Island was at least 99% successful. Chernobyl was 0% by design. Fukushima failed quite badly because active cooling was required, and failed. Any new reactor will have passive cooling.

The alternative is solar power

The world has plenty of barren desert. Covering a very small fraction of this with solar panels would generate all the electricity currently generated in nuclear plants. It's still a small fraction to replace all other electricity generation.

Yes, there are problems yet to solve. All that I've seen boil down to economics, not physics. Energy supply between sunset and sunrise is the hardest. Possible solutions include pumped gravity storage - best integrated with a tidal barrage; molten salt thermal storage, flow-battery storage (do we have enough Vanadium?), flywheel storage, electrolytic hydrogen storage, underground salt-cavern compressed-air storage.

If we can get 20% efficiency from a thin-film solar panel manufacturable by a continuous process, that doesn't require impossible amounts of rare elements, the economics will start to look attractive, even compared to burning coal. Another ten years and we may be there.

Depending on who does the calculating, solar economics may be ahead of nuclear already. (It certainly is for daylight hours, but an all-solar solution requires energy storage to be costed in. That's hard to do because the present electrical infrastructure does not include any massive energy-storage facilities. )

A fair result, not a good one

It showed that even in a case worse than the worst-case scenario, the deaths and damage from the reactor meltdowns were small in comparison to the same caused by a natural disaster of unprecedented size.

It also showed up two massive failings. One, that the reactor relied on active cooling. (All current reactor designs are passive-cooled). Two, that the emergency power generators were vulnerable to flooding. (The reactors could have survived, if there had been emergency generators on a site at greater altitude or several miles inland, connected by armoured, floodproofed, underground cables).

Acer turns to trains for imports

Final assembly in the UK?

I wonder whether they have ever considered the option of doing final assembly in the UK (or EU). Ship components rather than complete laptops, and then configure CPU / hard disk size / RAM / Case colour in the UK as orders are received.

It works for cars.

It's official: IE users are dumb as a bag of hammers

No, not a Mac exile

She's never used a Mac. The "killer app" for her work needs Windows. I asked her what she thought about the other browsers. IE: "gives you viruses". Firefox: "slow". (based on 3.5 and her slow old laptop). Opera: "What?". Chrome: "too new" "don't want Google spying on me".

And she likes that accursed ribbon interface in Office 2007. Sigh.

IQ does mean something

IQ is a measure of how good one is at doing IQ tests.

As for how strong or otherwise is the correlation between someone's test result and what you perceive as their level of intelligence, that's a hard thing to study. However, I think it's unlikely that many of the folks who can't score more than 85 in an IQ test are going to strike you as any kind of super-smart dude when you meet them. In other words, there's a fairly good corellation with something that you'd like to measure, but can't.

As for self-selection bias, it's hard to see how there is any in this study, which gives the browser chioce by percentage, for people testing with similar IQs. For there to be a bias, you have to think of a reason why an IE user with an IQ of (say) 115 is less likely to take the test than a Firefox with that same IQ, whereas an IE user with an IQ of 85 is more likely to take the test than a Firefox user with an IQ of 85. That a dim lightbulb may be more likely to select himself than a bright one doesn't affect the results.

The missing detail is the number of people in the various buckets - it it's too small then the error bars (not shown) will be large and the evidence concerning those at the extremes of the

IQ distribution may not be enough to exclude the null hypothesis. The trends are pretty persuasive, though. Modulo self-selection bias, you can estimate the number of people in each bucket from the shape of the IQ distribution curve (more or less a normal distribution) because they give the total number of test results analyzed.

one from me further down ...

... pointing out that if you filtered out people responding from work computers and internet cafes, then you might lose a disproportionate part of the high-IQ IE users. Just as you say, they don't choose to use it, they are given no choice in those environments.

Ever been bored?

Ever wanted to fill a few minutes? Ever done a (non-cryptic) crossword puzzle? Ever played Angry Birds?

Anecdotal, but ...

Over the week-end I was helping an artrs graduate set up a new laptop (a very bright cookie but completely IT-clueless, or so I thought). First thing she did after I'd done the tech stuff was to download and install Safari.