Posts by Werner McGoole
101 posts • joined Tuesday 11th March 2008 20:33 GMT
Posted in Bacon sarnies can kill: Official
So how does bacon kill you, exactly?
I can't help wondering about the correlation with "all cause" mortality. Did they not have cause of death information? Could it be that bacon affects your reaction time and gets you killed in car accidents, or perhaps sausages weigh you down and make you more likely to drown?
If they did have cause of death information, then why not look at the diseases that all the commentators have immediately assumed are caused by the great bacon threat, like heart disease and cancer? It makes me think that perhaps the correlation disappears if you do that. They'd then have wasted their time because no-one is going to take seriously a paper that proposes a sausage-drowning theory.
It all looks very suspicious to me.
Re: coincedences
Of course, it might be that they are one and the same object, but they don't like to admit they got the calculations wrong.
Re: Some dude writes something up. Next: animal entrails, bone casting and the I Ching.
Actually, both systems are on similar shaky ground. In the mathematical case assumptions have to be made about the difficulty (or computational effort) required to reverse the encryption. For example, the difficulty of factoring the product of two large primes. I don't believe that has ever been proven to be difficult. It's just an empirically observed fact, not unlike the fact that QM seems to be a good description of nature.
Anyone could come along at any time with a better method/theory and break the encryption in either system.
Predictions can and do change
It's interesting that after a few years when temperatures haven't risen so much we have started to see models which predict less warming in future.
Of course, much of the science that goes into climate models is still quite uncertain. So it's unsurprising that science alone doesn't well-constrain the models (imagine creating a model locked in a room containing science books but unable to observe the world outside). So the actual past behavior of the climate is a major factor in defining what the models predict and, in fact, in clarifying the science in the process.
Or to put it another way, without a firm theoretical basis, the models are inevitably going to be a simple extrapolation of past events (and recent events in particular, since we have so much more data there). And that, of course, is exactly what we're seeing with the recent Met Office prediction and now this one.
It's also worth keeping in mind that low-confidence forecasts are likely to change. If you are 70% certain about a prediction (and that's the ball-park we're in for many aspects of climate), then there's a 30% chance you're wrong. That means that your next forecast, some years down the line, stands a reasonable chance of being different.
I think when tempted to be dogmatic about climate issues, it's worth keeping that in mind. Just because this prediction differs from an earlier one doesn't necessarily mean either was flawed. It just means we don't really know what's going to happen. In fact, you can reach that conclusion just by looking at a temperature graph.
Um, yeah, but...
The point about invisibility cloaking is that you can't disappear by hiding behind a picture of nothing, so you need to do some fancy light bending instead. But you can look like a tree just by hiding behind a picture of a tree - or even a real tree.
So how is this any better than a picture of a tree, or conventional camouflage?
Meh
I think people mostly appreciated him for his work in astronomy where he was an acknowledged expert. The fact he didn't contribute much to political correctness is pretty irrelevant as I don't think many people consulted him over that.
Re: Homo informaticus?
Could it have something to do with IT people enjoying using computers and doctors hating it, perhaps?
Constant regulatory FUD is what we need
If you have lots of laws to control the press, they will, at some point be abused.
If you don't have lots of laws to control the press, the press will, at some point, abuse that freedom.
So the solution? It's to have a hotch-potch of laws that no-one understands so they're all scared of doing anything even slightly risky. Most of the population already seems to be well-covered there, but obviously the press aren't.
So the best way forward, it seems to me, is to introduce huge amounts more FUD into press regulation, so they don't have a clue where they stand.
So far, everything seems to be on course to achieve that end perfectly. But of course, FUD only lasts so long. After a bit, people find the loopholes in any scheme, so new FUD is required. The bankers were a good example of that. Let's not make the same mistake with the press.
Java ain't so bad
I have number-crunching Java code running almost 24/7 and it's no slouch. Several of the inner loops have been re-written in C, unrolled, and all the optimiser flags tweaked to suit the CPUs it runs on... and... well after all that effort they do run about 30% faster. But given that the Java code required no optimisation effort whatsoever, I don't think it's doing too badly.
Dedicated hardware is what you want for real speed improvements. I happen to have some, but it's in use just now, thinking.
Re: What do they need a law for?
I think the idea of the law is to leech off Google to get some of their money. I could have got that wrong, but somehow I suspect not if human nature is the same in Germany as elsewhere.
Regardless of the merits of the case...
I'm still torn over which party I'd more like to see lose. On balance, I think probably both of them.
Re: MS should just post him a free sdcard
Or they could erase the OS for him.
Re: Science doesn't do equal weight
You don't have to report 49 items of evidence on one side and 49 items on the other side just to balance it. That's plain daft.
However, doing 100% of the evidence on one side and 100% of the evidence on the other isn't daft and is what we'd usually consider to be balanced reporting. This allows people to make up their own minds on the balance of evidence.
The problem is that the BBC is doing 100% on one side and 0% on the other side.
Re: Questionable credentials
I think this is the crux of the matter, actually. AGW may (or may not) be happening and scientific consensus may (or may not) exist, but can the BBC rely on the attendees at this meeting to inform them reliably on these matters? You have to remember that the BBC staff themselves aren't scientists, so the quality of their independent expert advice would be absolutely critical.
Apparently, the BBC thought they could (or so they say) - and were so convinced of the reliability of this group of experts that they changed editorial policy in an unprecedented manner on a topic of huge importance. However, pretty well any outsider can see the shortcomings in this meeting in terms of expertise, coverage, balance, etc., etc.
So it does look very much as if the BBC convened this meeting as a smoke screen to justify an editorial shift that they'd pretty much decided to implement anyway - which, of course explains their reluctance to have it examined.
This tendency to select the evidence to support their editorial preconceptions and not to seek out independent evidence is, of course, also what got Newsnight into such a lot of trouble.
He he...
Just waiting now for Newsnight to do an expose naming the individuals.
Re: Love these lawyers
I had a solicitor once who initialled stuff B.A.D.
Ah, but...
A civilisation advanced enough to build such a sphere might also be smart enough to adjust its infrared output to look like something innocuous, like a star surrounded by dust.
Re: Barrier tape?
You mean those mains outlets that look like a tangle of bare wires hanging down the wall?
Leafy suburbs demand... leaves
They should do what they do with phone masts (i.e. make them look like trees). I'm sure a few plastic leaves could make them look like a nice bit of suburban shrubbery, or a hedge, or topiary even. Depending on how posh the neighbourhood is.
Double-deckers for me
Bacon downstairs, fried egg upstairs. Pepper on the bacon, ketchup on the egg.
How long...
...before miniature Google drones, like humming birds, are flying in through open windows to map the interior of your property as they surreptitiously seek out USB ports to slurp your data nectar from...
Of course, this could only happen if one of their rogue programmers slipped up again.
I'm on VM
So I can gloat. For a change.
I guess...
...they momentarily forgot that the little girl was a customer. That sort of thing seems to happen a lot in local government.
Re: Indoctrinated
From what I can see, the American system does seem to have some serious flaws, but I'm not all that keen on the NHS here in the UK, either TBH. The idea is basically OK, but yes, there is an element of communism to it and the supposed "costs to the taxpayer" give the state too many excuses to interfere in personal lifestyle choices.
On top of that, standards of service in the NHS are often third-rate, which isn't all that surprising given that anything run by the government has this problem. If you want better care, you have to go 100% private. There's no way of "topping up" the basic NHS provision. The resulting step-change in costs is basically an ideological defence against the NHS being incrementally eroded by private suppliers who could potentially provide a better service.
But there are lots of health services in the world, and some of them look pretty good to me. I just wish we in the UK could eradicate that "not invented here" attitude and look around us a bit. The NHS is past its sell-by date IMHO, but that's not to say it can't be revived if we adopt a more open mind.
Meh
Yeh, I read Taleb's book. He has a talent for taking a half-true idea and pushing it beyond its limits. Pretty much exactly what he's complaining other do, in fact. There are ways of handling noise in data and a huge literature on the subject. I doubt if Taleb has read much of it, though.
Copyright is only a concession
Copyright is granted by society to encourage creativity. Does society have a lack of creativity (or at least the production of copyrighted works)? I don't think so. In fact rather the opposite in most fields.
So logically, we'd be better off with less protection for copyright holders and more of their work being available for the benefit of others. A greatly reduced copyright term would be a good place to start.
Publicity seeking scientists spin their result for the press
Actually, there's no need for infinitely rigid rods. You can demonstrate this "science" with two fingers. Raise one to a certain height and raise the other to a slightly lower height. Now quickly raise the second one to be higher than the first.
Did you see what happened? The finger height maximum moved instantaneously from one place to another.
OK, you can quibble about exactly how fast it moved, but separate the fingers by a few light years and there's no doubt it'd be a lot faster than light speed.
These guys will have more information about me than I keep about myself
So I demand a search too so I can use it too!
Posted in Sitting down all day is killing you
Yes, living a good life is pretty dangerous
It's only better to stop and make tea if tea-making is safe. Have you studied the statistics for tea-related mortality (or even just burning yourself on the kettle)? Enquiring minds need to know these things.
If things go on like this, we'll soon be doing nothing but running around (assuming running is proved safe) to find the least-hazardous activity at each point in the day and never achieving anything. Remember that quite a few important break-thoughs in science have required substantial personal risk-taking. Marie Curie and Benjamin Franklin spring immediately to mind.
We all know you don't get anywhere by never taking risks. I think this sort of study is now showing that you can never enjoy yourself without taking risks either.
These tests are often pretty meaningless
I think it's widely known that anti-spam products perform rather better in tests than they do in the field. The testers have to choose a "standard" sample of spam messages to test with. If these don't exactly match the mix of spam arriving at your company, then the performance is going to differ. The relative performance of different products may also differ in the field.
Generally, the producers of anti-spam products can more easily guess what spam the testers will be using than they can guess what spam a typical customer will be getting. So they can tune their products accordingly. This is especially true when you consider that spammers' tactics are changing all the time. The "standard" test set will be spam collected in the past, which the products will be adapted to cope with. In the field, they also have to cope with new stuff that no-one has ever seen before.
So take the test figures with a pinch of salt. It could just be that Virus Bulletin has changed the way it collects spam or made some other change that makes it a little harder for the producers to tune their products for the test.
They had it coming
Maybe there's something to this "security though obscurity" thing after all. It seems it's hard to be famous and anonymous at the same time. I can't say it surprises me.
Maybe
This could become an olympic sport: synchronised mountain measuring. I can't think of any other reason for countries to compete for the pointless goal of having measured Everest most accurately.
Posted in Melting Arctic leads to snowy winters
What would be more convincing
Is a series of formal climate forecasts. Maybe some figures for snowfall in particular places, or frequency of winds above a certain strength, or annual rainfall, or temperatures, or whatever - over a good number of years (let's say at least 10 but ideally much longer). After the appropriate number of years have elapsed, a proper analysis could then be performed to see if any agreement with observations is statistically significant.
To really convince, you would need to make predictions that can't be made simply by extrapolating recent trends. They would have to involve detailed climate modelling in situations where this predicts something we haven't yet seen. And, of course, since there are multiple climate models available, a *consensus* on the forecast would need to be reached before the observations are made (because choosing your model afterwards is cheating).
These are the standards to which proper science is held. Explaining past events is all very well, and obviously necessary, but making successful predictions is what really convinces. I'd be interested to hear details if anyone thinks climate science has already passed this test.
Public electricity
I guess they'll just throw him in jail for a bit, then. Where they'll be able to use a bit more public leccy lighting his cell, cooking his food, etc.
Maybe he'll even be allowed to charge his phone there too.
Never mind
Perhaps the nuclear power plants will need some really big fans for cooling something or other. We have lots of expertise in really big fans on poles, so that'll put us in a strong position. Perhaps we'll even be able to export the technology around the world in future.
Thank goodness we have strategic thinkers in this country.
A vascular outcome?
Still trying to get my head around that one.
Perhaps a cerebral outcome will be along soon...
Spam spam spam...
My Google privacy email message came to the account I reserve for junk use, expecting it to get spammed, and not to any of my other accounts.
I think Google knows that email address because I've used the account to pay for things through Google (I'd have used that account because I don't trust Google, with good cause as it turns out). So I'm not altogether sure they've blindly spammed all their addresses.
OTOH, maybe they tried to be smart and targeted the account with the most traffic (i.e. spam).
Hmm
So it's faster than a FFT because it doesn't do all the transform. Why didn't I think of that!
Seriously, wouldn't a better comparison be with other lossy compression algorithms? In which case the advantage of using sin & cos as the basis functions isn't altogether obvious.
Posted in DIY virtual machines: Rigging up at home
Still to be persuaded
I use VirtualBox on Linux and run VMs for a few tasks that require a degree of isolation or experimentation with the OS (the ability to roll back to earlier states is very useful). I have considered a home virtualisation setup several times, but a number of problems always deter me:
1) Windows licences: why would I want to pay for these to run on my VM when I can use Wine to run the few legacy apps that are still windows-specific? When buying a new machine, you get a free windows licence (well, not free, but it's usually hard not to pay for it anyway), so why not use that on the machine it was bought with if you really need windows?
2) You still need to keep all your VM-installed software up to date. At a basic level this is for security, but in practice also for compatibility with external stuff that keeps on changing. The more VMs you have, the more hassle that entails (except possibly if they all use identical software). Eventually, the OS installed on a VM needs upgrading because support stops. That's a whole new load of hassle too. If you installed it to run legacy apps, you may find they don't run after the upgrade, negating the whole point of using the VM in the first place.
3) The guest tools that provide integration with the host often don't run on old or obscure guest OSes so, again, VMs aren't an ideal solution for running really old software.
4) Historically, virtualisation software has had its fair share of bugs. Compatibility between old VMs and new versions of the virtualisation server hasn't always been 100%. Some older OSes won't run on certain versions of certain VMs. Integration between host and guest can break when versions change. Sometimes you can live with the consequences (although they're less acceptable to family members who aren't so computer-savvy), but sometimes they are show-stoppers too.
5) Access to hardware from VMs, as others have noted, is usually fairly pitiful, typically with serious performance issues. Even if your chosen VM allows direct access to devices, this will remove the isolation that the VM provides, so you'll have issues when the hardware changes (and that was what the VM was supposed to prevent).
I think the goal of separating the OS, the personal work environment and the hardware is a laudable one, but I'm not sure that virtualisation really pays off in a home environment.
Posted in Why your tech CV sucks
Not much of a recruiter are you?
I've processed sackloads of CVs in my time, mostly for hard-core coding positions.
If you have that responsibility you should realise that it's not your job to critique the style of the CV writer. It's your job to hire the best possible employees. If Albert Einstein applied, do you think his CV would stand on a pinnacle above the rest? I rather doubt it.
The skill you need is to be able to visualise the person behind the CV and their suitability for the job from the information you've been given. If you can't be bothered to read through the longer CVs then you're not exactly a shining example of an employee yourself - because that's what you're being paid to do!
And what if you do get a job somewhere where a CV gets just a cursory glance? Chances are your colleagues will turn out to be the type who are full of BS with no substance to back it up. Lots of jobs in IT are hard to do and demand staff with real talent. That makes recruiting them pretty demanding too. I hope there are still a few companies around who put the effort in.
What'll happen is this...
The rich euro countries must realise that it's a choice between them hanging on to their money for now but having a financial disaster resulting in no euro, no banks and probably eventually no money. Or, they can bail out the debtors in the eurozone, resulting in a continuation of the status quo, albeit with less money for themselves and an urgent need to reform how the euro works.
The decision really rests on how much Germany and France want to keep the euro (i.e. how much they're willing to pay to keep it). When push comes to shove, if they decide it has to be rescued (and I think they will), the fact that there's no legal way won't matter. These are politicians - they make the laws (or they re-interpret them as necessary).
IMHO, Germany and France shouldn't be that bothered about having to bail out the PIIGS. The PIIGS have suffered from having an over-valued currency (the euro), while France and Germany have profited from having an under-valued currency (also the euro). If they want to continue profiting, it's only fair some of that gain goes back to the PIIGS who are keeping the exchange rate down for them.
So a transfer union of some sort, but the reason that hasn''t already happened is because the rich euro countries don't want to sign a blank cheque. While they can, they're determined to make the PIIGS reform their economies as that will make the eventual bill as small as possible. But I think they know the bill will eventually fall on them, probably through some form of QE. That'd be the fairest way, as all sectors of the euro economy would then pay via the effects of inflation.
I know the eurozone politicians are acting like they're stupid, but I think most of them know how this must play out. They just don't want to commit before they've squeezed as much as possible out of the PIIGS. Similarly, China, Russia, Brazil, US, UK, et al. aren't going to commit to helping the eurozone via the EFSF until they've squeezed as much as they can out of Germany and co.
It's all politics of course, but in this situation it's also good economics. Those who have made mistakes must pay, but only as much as they can afford to pay without wrecking the whole financial system. That's what's being negotiated at the moment via a process of brinkmanship.
Natural selection in miniature
You can do a nice demonstration of the power of natural selection if you have, say, ten dice. Throwing them all and getting 10 sixes will take a long time, as any creationist will happily point out. But if you throw them and just keep the sixes that turn up (let's call them successful mutations) and then throw the rest again and repeat... then you pretty soon have 10 sixes.
I guess this experiment is along the same lines. Analysed logically, its rapid success is obvious. But it's still amazing how many people just can't see the difference between this and throwing all the dice at once.
So I think it's a useful reminder that there is a world of difference.
Educational standards
Unfortunately, there are quite a lot of things that should be taught to kids before we get to anything as complicated as writing computer code.
I was siphoning some homebrew the other day and my daughter took a great interest because she'd never seen a siphon working before. I had to explain to her how it worked.
She's only got a masters degree in Earth Sciences, poor girl, so you can't expect her to know about these simple things.
Just sounds like Java to me
Well, not exactly like Java, obviously. But if the portable version was implemented its goals seem pretty close to what Java* was originally trying to achieve rather a lot of years ago now. That's especially so if you recall that Java turns a lot of its critical byte code into native code at run time for faster execution.
So if this is to have more success than Java it needs to have something extra to offer and I just can't see what it is.
*Not Javascript - thats something different.
AAAAhhhhhh!
I too used MTS in Durham from 1976. It still staggers me that we could do interactive image processing on it while there were a zillion other users also hammering away. Happy days.
Languages all have their own tricks
An important factor is whether the language allows (or entices) you to use constructs that defeat compiler optimisation. For example, much of the speed of old-fashioned Fortran came from the absence of anything like pointers - so the compiler could more accurately assess the scope within which a variable might be referenced.
C's use of pointers is probably the main thing that makes it slower than old fashioned Fortran, but if you code carefully and don't use pointers, that gap will close. With C++ you're one more step removed from knowing whether the compiler will be able to optimise what you write, so more frequently it can't.
To give another example from Java, garbage collection can be a real problem but can often be mitigated by avoiding unnecessary object creation/destruction. Unfortunately, this is again something that a compiler is unlikely to manage on its own as it's part of the logical design of the software and at a higher level than compilers work at.
So while compiler optimisation is always a good thing, good software design also lies at the heart of run-time efficiency. Those who say you can't beat the compiler at optimisation may be right at the level of loops and method calls, but at a higher level it's easy to design something so it runs slowly in any language you like. Knowing what the language does fast and what it does slowly is where the solution lies. So there's no real substitute for experience with the particular language in question.
Of course, this does make C++ programmers superior beings, as few are able to gain much experience with this language without shooting off both their feet at some point.
But how serious are the health issues?
I can't see why you'd think the Ars piece to be better. A lot of the content says much the same thing as Lewis says, but there's no analysis of the health risks beyond saying that the reactor workers are "putting their health at risk".
To my mind, quantifying the health risks is the crux of the matter. Health is the only thing we're really worried about here and if the risk is negligible then it's important to know that as it changes the entire story.
I think the fact that the reactor workers may never suffer any health problems at all because of what they are doing (never mind the general public), if correct, puts the whole thing into perspective. I haven't seen that discussed anywhere else. (Unless you count the Daily Mail's radioactive poison cloud that's currently swallowing up the US of A.)
Brilliant!
Haven't laughed so much in ages.
A title
Well, Lewis does like to be a bit contrarian, so it's interesting to see how much contrarian spin can actually be applied and still be (mostly) consistent with the facts. This is no worse (or better) than most of the media who are hell-bent on spinning the facts in the opposite direction. I don't mind reading either point of view, so long as it's intelligently presented.
My own view is that the reactor engineering has proved itself, in that it would almost certainly have breezed through an earthquake of the size it was designed to withstand. But this one was much bigger.
If there has been a failure - and I think there has actually been a HUGE failure - it is that an earthquake and tsunami of this size wasn't considered possible. I know it's Japan's biggest, but having an earthquake bigger than you've had before isn't something you should discount when designing a nuclear power plant, especially in an active earthquake zone. And it's no good saying that these were old reactors. If that meant they were no longer up to the job, they should have been shut down.
As things stand, Japan may well prove lucky and escape without dangerous radiation release, but the measures being taken are very much seat-of-the-pants stuff. Even if they succeed, it should never have come to this, and it wouldn't have done so if the reactor design had been up to the risks posed by their location.
So this doesn't put me off nuclear power, but it does make me worry a good deal about the risk assessments and the regulation of nuclear power plants.
Hang on a mo...
I think there's a fundamental point being missed here.
I can see that on the face of it charging a male more for car insurance than you charge a female may look like sex discrimination in favour of females. But you've only looked at half of the contract.
The other half of the contract says that if you have an accident, the insurance company will pay you a big wad of cash. Being male (apparently) makes it more likely that you'll receive this cash, so males actually stand to benefit more than females from insurance payouts. So on this measure, the insurance contract discriminates in favour of males.
So if the insurance company has its statistics right, gender-based risk assessment is, on average, not discriminatory. Both sexes will have the same expectation of net loss due to car accidents.
Of course, males will still get a raw deal because they can expect to spend more time in hospital and are more likely to be killed, but that's apparently part of being male and not something you can legislate away.
So what this ruling is actually doing is "positively discriminating" in favour of men to compensate them for this failing.
