28 posts • joined Wednesday 22nd July 2009 15:03 GMT
Models and truth
"People underestimate the power of models. Observational evidence is not very useful"
WRONG! The originator of this quote is not a scientist whatever they claim or their job title is. This directly violates Karl Popper's expression of scientific philosophy that, to be considered scientific, any hypothesis (of which a model is one form) must be falsifiable by observations. Observational evidence or measurements is the closest we can get to ground truth. If models don't agree with observations we have a term for them (watch out for technical jargon here): "wrong".
Models are worthless unless they correctly predict observations.
My school had mandatory General Studies too (nearly 20 years back). Everyone had to endure the lessons except the Further Maths group as the two clashed in the timetable. The idea of exposing science kids to arts and arts kids to science is a nice idea in theory but we couldn't help but notice that all the Further Maths group got A in General Studies and very few of the rest did.
Wrong way round
I disagree entirely with this one. What we need is tuition fees that reflect the costs of the training and potential rewards for completing it. Science and engineering training is vastly more expensive than English Literature or other soft subjects. Labs stuffed full of expensive, precision equipment, high cost consumables (chemistry and biology), computers used for actual computational work rather than email and Facebook and the qualified staff to run the whole thing properly. Compare that to the requirements of English Lit: a fiction library.
We have arrived at the current dysfunctional arrangement precisely because of homogeneity of costs to the student. If it costs the same to get a degree by doing a hard course as as it does by doing an easy one, many will pick the easy one irrespective of value or quality. If different courses cost different amounts then students will start asking very relevant questions like "Why is this course more expensive?" and "Is it worth me paying extra to get on a better course?" Questions like these lead to rational and informed decision-making, exactly the kind of thinking we want to encourage in general and most definitely in science and engineering students.
Dropping fees for science and engineering courses will produce more dropouts than new scientists. The reason is simple, more students will take the cheap option without having the skills required to study hard subjects, fail and quit.
If you want more scientists and engineers to go into teaching, you need to make it worth their while. Why take a teacher's low salary when you could get paid more for your hard to learn and expensively acquired skills elsewhere? Reward that hard work...
Not long enough for a signing key
The twitpic link contains 20 hex pairs. In other words that's 20 bytes or 160 bits. That's nowhere near long enough to be a signing key. Asymmetric RSA/DSA keys used for signing certificates and the like are 1024 bits minimum. 160 bits is more likely to be the SHA1 hash of the key.
Modelling is just that...
A model is just that, a model, a system for approximating reality (i.e. measured data) and making testable predictions.
Those of us in the science community who remember what a model really is and why it can never take primacy over measured data have a special word for a model whose predictions do not match measured reality. Such a model is, technical jargon here, "wrong". If a model is "wrong" then any decisions based on its predictions are also "wrong".
All of the models used by the IPCC fall into this category of "wrong" and thus must be discarded. Any decisions based on these models are "wrong" are thus baseless and have no value.
Let's feed the data gathered so far (without the data falsifying edits by the Hockey Team) until, say, 1970 into the new model. If the predictions it makes do not match the actual data for 1970 to the present then this new model is also "wrong"
AFAIK there is not one single model of climate that passes the above test.
Thanks Ralph, this is the only thing I wanted to know about iOS4.2 for my iPhone: "Fixed the embarrassing repeating alarm problem too, thank goodness."
None of the supposed fixes worked on my phone so this makes 4.2 must-upgrade-today for me. After the kerfuffle about this bug at the end of last month, I'm puzzled why this didn't make the main copy.
"Same fallacious point usually wheeled out by pro-speeders. The speed affects the *outcome* of the accident."
You seem to be deliberately missing the point and trying to divert attention with a statement of basic physics. Since only 5% of crashes are caused by excessive speed, that means that 95% are caused by something else. One of the major objections to speed cameras is that they target the cause of only 1 in 20 crashes.
If the intention was genuinely to improve road safety it would be far better to determine the causes of the other 19 in 20 and tackle those instead. A short list from my daily experience: stupidity, incompetence, recklessness. Sadly to actually deal with those problems would be politicial suicide as all but the 17 year old drivers can vote. The following measures would be unpopular but effectively cut the crash rate:
* Make the driving test tough enough to be meaningful. Include motorway driving and skid pan as mandatory parts of training. A driving license is a privilege for those competent enough to handle the responsibility, it is not a right.
* Mandatory 3 yearly retests for the over 70's. It doesn't matter that you've had X decades of driving, ability is proven to diminish with age and your ability now is what matters not your ability X decades ago. Your desire to dawdle to the shops does not trump everyone else's right not to be endangered by your lack of ability.
* Mandatory re-qualification for drivers causing a crash resulting in injury. Mandatory training "refresher" courses for causing lesser crashes.
* Limitation of power for new drivers like with motorcycles.
* Limit who new drivers can carry as passengers to close family members. Prevents peer pressure causing recklessness but allows acting as family driver to gain experience.
I'm not a "pro-speeder", I'm just against arbitrary, unproven limits and the failure of critical thinking that usually accompanies them, just to target 1 crash in 20.
"Regular drivers accept the fact they may be wiped out by speeders."
Really? Of all the dangerous behaviour on the roads I see every day, virtually none of it is speed related. Most of it is simple incompetence or rank stupidity from people who simply shouldn't be on the road.
I accept that I face a substantial risk that I'll get wiped out by another driver but it'll be some clown rounding a blind bend in the middle of the road, joining a motorway well below traffic speed, failing to see my not inconspicuous family saloon and pulling out of a side road in front of me or simply not using mirrors and sideswiping me on a dual carriageway or roundabout.
No cameras, speed bumps or anything else will get these people off the road. We need a return to properly trained traffic police with discretion who can tell the difference between those driving dangerously and those driving quickly.
SNR probably lousy
In between the dross about necro-bestiality and the rest, there was at least one very sensible suggestion for a law to repeal: European Communities Act 1972. Repeal that and all the EU interference and stupid laws can be tossed out in one fell swoop. Save us 6 billion quid a year to start with too. Somehow I can't see starry-eyed EU-fancier Clegg trying to actually get that repealed if it even gets noticed as a lone voice of sanity in all the clamouring :(
I'm no IE fan but look how much better IE8 is than IE6. Why is that? Because the market offered something much better so Microsoft had to improve. Competition is good. That's why I'm very keen on HTML5 because it will bring competition to web video, especially if VP8 and WebM take off.
Apparently a default or empty title is beyond El Reg's web coders FFS...
I think you'll be needing a new dictionary, old chap. Independent does not mean "stuffed with old cronies and ex-colleagues who'll never accept any failing on the part of their personal special interest group"...
More than just steel...
Quite so, there's a number of reasons MRI units in hospitals have carefully controlled access and patients have to be interviewed before being cleared to enter the magnet. Of greatest concern are things like pacemakers, surgical clips and metal particles in the eye sockets.
Another thing that MRI machines have are large RF shields around them. NMR signals are extremely small and even small amounts of outside RF noise can swamp them. There's also the small matter of a clinical system requiring a three phase electrical supply for the RF and gradient coil amplifiers.
Unfortunate reality nerdery aside, one possibly entertaining side effect would be down to the fact that humans are essentially large bags of conductive liquid. Move a powerful magnet next to a conductor and you get large eddy currents. If that conductor happens to be someone's head, the eddy currents will flow through the brain and induce disorientation, nausea and vertigo. Wave one of these quickly past a patient's head and watch them fall over...
Tethering in UK?
O2 charges extra for tethering. How can that be justified? If you have a phone plan that allows unlimited data to the phone, what possible difference is there between data going to the phone and data passing through the phone to a computer? It's all just data flowing over the network... Not tethering my phone until the network gets rid of this artificial compartmentalisation of function.
Tensor Analysis Necessary
If by "predicted in the 18th century" you are referring to the Michell-Laplace dark body, no they weren't. This is a misunderstanding of the definition of escape velocity. This and other errors in black hole theory are explained in accessible form here: http://www.ptep-online.com/index_files/2006/PP-05-10.PDF
Schwarzchild's paper is here: http://www.sjcrothers.plasmaresources.com/schwarzschild.pdf
Hilbert's (commonly but incorrectly referred to as Schwarzchild's) is here: http://www.sjcrothers.plasmaresources.com/hilbert.pdf
The black hole theory and its history are a neat example of how failure to examine the underlying assumptions of a theory leads to false conclusions. Once the erroneous basis is commonly known, academic reputations based on the false conclusions will be wiped out.
To the person who wanted to know what was at the centre of the Milky Way, it is thought to be a plasma focus in an electrical discharge. This theory is taken seriously by people such as the IEEE and their journal IEEE Transactions on Plasma Science.
Tensor Analysis *is* necessary to understand Hilbert's error and why no fewer than four derivations of the solution to Einstein's field equation are mutually consistent and contradict Hilbert.
Given that black holes are a theoretical construct that derive from a mathematical error made by Hilbert in the early 1900's and thus cannot exist, I very much doubt they will destroy anything but academic reputations...
I need no title
I'd stopped buying games with intrusive DRM before even Bioshock and its rootkit came out. However, I have many games on Steam. Why? Because Steam gives me, the paying customer some definite advantages. They have taken, what is effectively a password-based DRM scheme and made it beneficial to me.
1) Any game I have bought through Steam, I can play on any PC I install it on, no need for disks.
2) Steam auto-updates all my games.
3) Steam plays offline but syncs data with the cloud so my controls stay the same etc.
4) Steam allows me to play games I have lost or damaged the disks for if I'm prepared to wait while it downloads.
There's other conveniences too like early access to some games, being able to buy without braving the great unwashed on the high street, finding multiplayer games to join, easy access to friend's game lobbies.
Steam may be DRM at heart but it is a consumer-focused product that gives me real advantages. If Steam can make it work this well, there's no game possible that could be good enough for me to use UBI's type of DRM.
...is the one thing I could never reconcile myself to with compacts which is why I bought an SLR a couple of years ago and never looked back. So many missed shots with compacts and their shutter lag :( I got used to the bulk and weight of an SLR very quickly so when there's a compact with SLR-like focus speed and shutter delay, that'll be news. Until then, this is just another compact with the failings thereof. RAW, face recognition, low-noise at high ISO are meaningless unless you get the shot.
Free versions of content do not require unblockable ads to be inserted at the OS level, only at the application level. The first time I get this on my Apple gear will be the last time it boots OS X and the end of my time as an Apple customer.
@Ignorance not stupidity
By your own definition:
"Ignorance is simply not having learned something, stupidity is not wishing or caring to learn, or acting on that knowledge."
most computer users are indeed stupid. The vast majority of users do NOT care as long as they can surf the web, read their email and download cutesy screensavers and wotnot. They have no more interest in how their computer works or how it affects other people than they do in the functioning of a rotary Wankel engine. Learning might involve a bit of effort and that might tax their brain cell a bit too much or take up mindspace that they'd rather devote to football/Eastenders/X-factor.
I don't recall the source of the quote but: "Your right to swing your fist ends at the tip of my nose." In other words, you are free to do what you will until you have an impact on someone else. This case is completely clear cut. She is causing a noise equivalent to a normal conversation in *someone else's bedroom* in the middle of the night thus impacting that person's ability to live normally. She must bear responsibility for her actions, whether it be keeping it down or installing soundproofing.
I hope she gets laughed out of court and saddled with a fat bill for her stupidity.
Crypto + RIPA == FAIL
Right, so let's get this straight...
Government fucktards can lock us up for keeping our private data private with proper encryption and refusing to give the key but they can't be locked up for disseminating private data because they can't be arsed to encrypt it?
I'd wish the entire cabinets' personal data was publically disseminated to show them why we don't want our perfectly innocent personal lives open to scrutiny but in the current culture of zero accountability the only people who'd benefit would be headline writers.
Law is an arse
Just because something is enshrined in UK law does not confer upon it the properties of being just or legitimate. RIPA and other supposedly anti-terror laws spring to mind. Copyright laws need significant change to reflect the fact that the marketplace has been irrevocably changed by widespread access to digital media. In the past, every time a new and disruptive technology has arisen (paper fed pianos, vinyl, radio, etc) and changed the marketplace the copyright laws have changed. This time should be no different.
I think if you read the bit you quoted again and emphasise the crucial part ("drive as fast as is safely possible within that limit") then you'll see that it allows for common sense behaviour like slowing to below the 60 limit on steep winding roads. While you may have the awareness that you are holding people up and good courtesy to let them past where possible, a large number of drivers who dawdle along roads at 2/3 of the speed limit do not. This leads to frustration which regrettably can lead to unwise overtaking maneuvres. The fault is bilateral here: the person driving too slowly is at fault for failing to make good progress and holding up others as well as the person who overtook inapproriately. Scotland has many main roads which are difficult to overtake on and has signs reminding people that driving slowly causes frustration and crashes.
The most important place for a minimum speed limit is motorway slip roads when joining the motorway. Forcing a truck doing 70mph (along with the rest of the motorway) to emergency brake by pulling out in front of it while doing 40mph is extremely dangerous. Doing it to other vehicles is still damn dangerous and fortunately most of the (frequent) occasions I see this kind of stupidity it doesn't result in a crash as the drivers already on the motorway see the moron coming and can maneuvre away.
I certainly don't live on Mars, I live within the M25. The terrain here is a bit steep but that affects FM reception not one tiny bit. My DAB set has been switched off since we moved here because it is completely unusable. I cannot find a single multiplex that gives anything but the bubbling mud. With such appalling performance from a decent quality non-portable set, why would I ever contemplate replacing the perfectly functional built-into-the-dashboard FM radio in my car?
Audio quality, critical to an audio only medium, has been covered very well by others.
The fool who thinks that reception indoors isn't so much of a problem needs to realise that there are, in the main, two places where radio is listened to. On transport (car, bus, train) and in buildings. There are some crazy folk who commute 2+ hours each day but that still leaves the rest of us spending the majority of our listening time *indoors*. No indoor reception means it will never be used. Internet radio isn't an option for those of us who are behind draconian content filters.
So let's get this straight... XP mode is a full VM and you need to patch it as a separate machine, give it anti-vx as a separate machine, firewall it as a separate machine, in fact do everything as a separate machine? Quite apart from the double performance hit of doing everything twice, how many end-users are going to realise they need to do this?
Only reason I'd buy this is to get a blu-ray player. Looks like they've retained the piss-poor PS2 controller. The Xbox360 controller fits so much better in my hands... No, I'm not an M$ fanboi but I acknowledge hardware that fits my comfort and needs. Shit, I use an Xbox360 controller for some PC games cos it just works better with the analogue (yes, I'm English, the ue matters :P) triggers.
PS3 still only a cut-price BR player...
@AC - More CCTV Please!
Speaking as someone who has been mugged at knife point in an area liberally covered by CCTV in broad daylight by people not wearing anything to disguise their identity, I say tear all the cameras down and get coppers doing a proper job of patrolling. All that footage and no prosecution was brought, not even a suspect was identified, CCTV doesn't fucking work - the image quality is piss poor...