Weren't www.beautifulpeople.com previously listed as www.narcissism.com ?
Posts by Michael H.F. Wilkinson
4257 publicly visible posts • joined 24 Apr 2007
Page:
The Reg's best-looking reader reveals list of jobs for the beautiful
Leaked Obama brief reveals US cyber defense, offense policy
Monday 10th June 2013 07:48 GMT 
Having plans in place is not a problem
Running around like a headless chicken when the proverbial shit hits the fan is. So yes, it makes sense to have plans in case war breaks out.
Thus, I do not find it at all surprising that there are plans in place for both defensive and offensive action. What is a problem is their constant data gathering on everything. Adding more data to a pool indiscriminately is like wall-of-death fishing: you catch and kill all sorts of stuff you did not intend to. I also do not buy the argument that they have "advanced data-mining algorithms". Data mining is a fancy word for pattern recognition in big data. To stay in the parlance of mining, let data be the ore, and information be the metal you want to extract. As the percentage of metal in the ore becomes lower, it makes less and less sense to extract the ore from the ground in the first place, unless the metal is very very precious. Even then, it is possible to be sidetracked by things that look like the metal, but aren't. Fool's gold (iron pyrite) springs to mind.
Likewise, if you gather data about everybody, looking for a minute percentage of people that actually hate the US sufficiently to do it real harm, the chances are false positive rates will skyrocket. This is a very real danger. It is much better to gather information in a far more directed manner, instead of implicitly suspecting everybody of ill will (which in itself seems to be a self-fulfilling mode of action).
As I have said over and over again: adding hay does not make finding needles easier. In mining parlance, there may be gold in them there hills, but it is better to drive a well-chosen shaft than indiscriminately removing the whole range of hills from the face of the earth, and sifting through tons of worthless rubble.
Given that the earlier leaks were about the far more dangerous data snooping, a cynic (who? me?!), might think this leak is there to distract from the more dangerous issues.
Enter the Dragon: The Chinese superputer set to win the Top500 crown
Help! I’m trapped inside the Chamber of Hollers
Friday 7th June 2013 13:22 GMT
Somehow this reminds me of my parents in law, who (after a career teaching mentally less agile people, to put it mildly) were used to settle every argument by the application of equal parts of authoritarianism and volume (they still go at it hammer and tongs at each other, producing Pythonesque near infinite loops of "Yes it is! No it isn't!") .
They found out the hard way that this did not work on a scientist with a voice that some have compared to Brian Blessed in full spate (but only when I talk enthusiastically about anything or am just plain angry).
Curiosity prepares for year-long road trip to Mount Sharp
Raspberry Pi DUMBS DOWN to target world+dog
Jack Vance: Science fiction’s master of magic, mischief and sex
Women shun careers in IT security
Germans purge selves of indigestible 63-letter word
Wednesday 5th June 2013 08:34 GMT
The Dutch have their own way of creating massive words:
Hottentottensoldatententententoonstellingspersonenautoparkeerterrein
is one construct, but their tendency to create words with an uncomfortable number of consecutive consonants really stands out:
angstschreeuw
with 8 was long held to be the champion, but
slechtstschrijvend
might top that with 9 (jury still out on whether this word is OK. Trust the Welsh to come up with place names that (appear) to consist only of consonants:
Cwmtwrch, Bwlch and Mwnt were all places I have passed on my meanderings in Wales.
In these cases the 'w' is pronounced more or less like an "oo" as in food
Pen+tablet bandwagon finally rolling, Nvidia leaps aboard
Hitchhikers' Guide was WRONG, Earth is not in a galactic backwater
LOHAN team regroups for second pop at SPEARS
Intel's plan for Haswell, Silvermont, Bay Trail: WORLD DOMINATION
Quantum boffins send data ACROSS TIME AND SPACE
New EXPLICIT pics support notion of moist, welcoming past for Mars
Friday 31st May 2013 12:11 GMT
Dead right. Imagine the excitement if the fluid turned out to be alcohol (how easy would it be to get volunteers for a trip to Mars then?).
Alternatively, it was octane, and some Martian equivalent of the Humvee became so popular that they used every last drop (hence the carbon-dioxide atmosphere).
Sacred islet Rockall repels Brit adventurer's first assault
Google nuke thyself: Mountain View's H.264 righteous flame-out
Friday 31st May 2013 11:20 GMT
Re: Codec patents
The difference is actually made in patent law in most countries.
Most patent laws explicitly exclude patenting mathematical equations (though not necessarily their application). Because any algorithm can be expressed in lambda calculus, it could be (and has been) argued that algorithms should not be patentable. Their application (as e.g. as codec) could be under various laws.
So, as Mr. Slant would say: "This is an interesting case."
With degree of interest being defined as proportional to the amount of money it brings to lawyers.
Fedora's Schrödinger's Cat Linux gives coders claws for thought
BOFH: Go on, beancounter, type DROP TABLE asset;
Curse you, old person, for inventing computers!
Friday 24th May 2013 11:57 GMT
Old git 1: When I was a lad, we never had these newfangled integrated bleeding development environments!
Old git 2: Aye, we had emacs, and we were glad of it!
Old git 3: Emacs! we would have loved to have had emacs, we were stuck with vi, but we were happy!
Old git 4: We did not have editors at all, we had piles of punched cards! Hundreds of lines of FORTAN, one to
each card
Old git 1: Luxury! We had no programming languages at all! Everything was machine language
Old git 2: Machine language!? We would have died for machine language. We had to hand wire connections
in our computer to program it!
Old git 3: Right!
When I was young we had to compute entire navigation table with nought more than pencil and
paper and a hand-cranked mechanical computing apparatus!
Old git 4: And the problem with kids today is that if you tell them they don't believe a word you are saying!
The BOFH is BACK: And it's cloudy with a 90% chance of beatings
Smartwatch face off: Pebble, MetaWatch and new hi-tech timepieces
Thursday 23rd May 2013 11:52 GMT
They seem to be getting better very slowly. I have yet to see a real killer app for these, but maybe one will come along.
I will just stick with my current watch. I only ever buy a watch when the previous one breaks in such a way that repair is no longer economical. Given that my current (only my third) watch is a year old, and on average, my watches last about 19 years (OK, N=2 statistics), there could be some really enticing options available when I am up for my fourth watch.
WW II U-boat attacks prompt new US response
Soylent Corporation prepares to DEFEAT FOOD
Wednesday 22nd May 2013 07:37 GMT 
Thanks, but no thanks
Tonight I think I might be serving some portions of monkfish fillet, lightly drizzled in olive oil and some lemon juice, sprinkled with pepper and some fresh sage from the garden, wrapped in lean smoky bacon, baked in the oven at 220 C for just 15 minutes served with pasta and pesto alla genovese, and spinach.
Alternatively, I might just have some pizza. Some beer or wine would go down a treat as well
HTC woes prompts 'leave now' tweet from former staffer
Wednesday 22nd May 2013 06:53 GMT
Pity really, I liked my HTC Desire a lot. I now have a Desire X (in part because the interface was familiar, in part because I got a good deal), but when that needs replacement I may well have to look elsewhere, unless by then telcos are dumping their HTC phones as they too bail out (something tells me living in the Netherlands has influenced me ;-) ).
Computer use irrelevant to education outcomes, says US study
Tuesday 21st May 2013 14:42 GMT
Re: The key to teaching is the teacher
One thing that the government here in the Netherlands might try (maybe the same elsewhere), is to let teachers teach, not administrate. I gave a few lessons of computer science at schools nearby, to show kids that CS is certainly not learning to work in MS-Office. I was appalled at how much time teachers have to waste on administration during each lesson. And then there are the endless assessments and test that are foisted upon school children here from the age of 4 upwards. Endless assessments to measure outcomes. As one teacher pointedly stated: "A pig doesn't get fatter by weighing it more often". This tendency to measure everything to death has even reached preschool creches we have. I have heard people agonize about the fact that a 3 or 4 year old kid only recognized 4 letters whereas they should know 5 at that age. As if that difference is at all meaningful!.
When teachers are assessed by the teaching outcomes of their pupils, they will train them to perform well at the test, which is not the same as giving them understanding of the subject matter (let alone stimulating them).
</rant>
Curiosity plunges its drill into Mars AGAIN, seeks life-giving sample
New 4TB drive spaffs half a telly season into your eyes AT ONCE
German robots sent to Oz to make GPS millimetre-perfect
The iWatch is coming! The iWatch is coming!
Tuesday 21st May 2013 07:55 GMT
Why not a BOFH watch?
which runs linux, allows you to log into your servers. A full keyboard would perhaps be a bit much, bt a small set of function keys with the following functions could be enough:
F1: rm -rf *
F2: kill -9
F3: shutdown -h now
F4: reboot
like this one
Suggestions for other keys welcome
MYSTERY Nokia Lumia with gazillion-pixel camera 'spotted'
Tuesday 21st May 2013 08:58 GMT 
Let's talk photon counts and well sizes
Cramming 41Mpixel into such a small format is apparently possible, but only by making the photosites tiny. This means the electron well can only hold comparatively few electrons before saturating. This in turn means that the number of photons detected before saturation is low, which means increased noise and lower dynamic range. By binning multiple photosites together into a much smaller number of pixels, you can counteract this, and no doubt produce some very nifty 8MP images (especially with good processing techniques).
However, this will never match the quality of a DLSR, simply because the much bigger lens of a DLSR gathers more photons, which means better signal to noise. The reason is that quantum efficiencies (percentage of photons detected) of photosites are very high indeed (40% and above) so little can be gained there (and every gain at the phone end can be made equally at the DSLR end). A DSLR lens easily has 4x the diameter of that of a camera phone, meaning it gets 16x the number of photons. A phone sensor would need to collect 800% of all photons (which is clearly impossible) to match a DSLR sensor with 50% quantum efficiency.
So, will these cameras produce nice images, good enough for most purposes? Even when deducting marketing speak: yes. Will they match professional DSLR kit: No; this little thing called physics gets in the way. On the other hand, do I carry a phone with me all day? Certainly. Do I carry several kilogrammes of DSLR kit around all the time? No way!
Crack Army pilot to be first PROPER British astronaut IN SPAAAACE
Nine-year-old Opportunity Mars rover sets NASA distance record
Acorn founder: SIXTH WAVE of tech will wash away Apple, Intel
Larry Page acknowledges creeping vocal paralysis
Bing uncloaks Klingon translator
Drone to deliver beer-as-a-service
Astronaut Chris Hadfield's Space Oddity ends in Kazakhstan
Tuesday 14th May 2013 11:12 GMT
Cheers to him
Every time I see the ISS fly over (happens quite often whilst stargazing, I could even spot its overall shape with big binoculars), I just have to give quiet praise to the people up there, and the people who helped put them there. However we bemoan not doing enough space exploration, some people are dedicating and even risking their lives in the name of space exploration.
People like Cmdr Hadfield set a shining example in my book. Cheers to him and all like him
ScaleMP: Use RAM plus vSMP, not flash, to boost server performance
Tuesday 14th May 2013 09:46 GMT
Re: Latency, it's all about latency
CSP is powerful. The main problem I find is in keeping communication down, especially in terms of how often processes need to communicate. It is much cheaper to have a few large chunks shuttled from on process to another, than it is to have a whole lot of little messages. It is not just the latency in that case (it also plays a role), but it is also the synchronization that costs time (barriers are particularly costly).
Monday 13th May 2013 09:40 GMT
Latency, it's all about latency
I have some rather memory-intensive code, that I did once run (or rather walk) on a ScaleMP machine (8 boards with 8 cores each (older incarnation)). Performance was dismal. Why? Each core thread may need access to each part of memory, because the outcome for each pixel in these huge images may depend on any pixel in the image, and you do not know beforehand which ones matter. Everything works hunky-dory as long as each processor only accesses the memory on its board, but the moment it needs large amounts of data from another board, latency kills performance. Getting a speed-up of 0.5 at 2 threads (if they weren't explicitly pinned to cores on the same board) is rather discouraging.
What they are doing is putting a software layer over a distributed memory or NUMA machine, so as to hide the complexities and allow shared memory algorithms to run (or rather walk) without the need to rewrite the code. ScaleMP does hide the actual NUMA architecture very well. Curiously, this leads to problems when optimizing the parallel code for that particular hardware. Because details are too well hidden. You really need to understand the memory architecture and the latencies of the machine to design the appropriate algorithm. Parallel programming on shared-memory machines and distributed-memory/NUMA machines are two very different ball-games, often requiring a careful rethink of the algorithms, in order to get the processors spend their time working, not talking (just like an old-fashioned classroom), or waiting for data.
A ScaleMP-like approach could work if the latencies are kept very low (like the QPI approach). On run-of-the-mill network connections (or even Infiniband), you need to rethink shared memory code, not so much because of bandwidth, but because of latency. For Cell/GPU type systems similar rethinks are needed, for much the same reason
Biting the hand that feeds IT
