Looking forward to the last results from Cassini's swansong/dive. It has already yielded a wealth of data and magnificent images. Kudos to all those who made it possible!
Posts by Michael H.F. Wilkinson
4248 publicly visible posts • joined 24 Apr 2007
Page:
Goodbye, cruel world! NASA's Cassini preps for kamikaze Saturn dive
Head of US military kit-testing slams F-35, says it's scarcely fit to fly
BOFH: The Boss, the floppy and the work 'experience'
Like!!
Nice episode, but Simon could have gone for the 8" floppy option. I still have one lying around somewhere (original CP/M 2.0 disk, entire OS on just 128 kB), but I doubt your average intern could find one. You could certainly defend that as a valid way of ensuring security through obscurity.
Forget robot overlords, humankind will get finished off by IoT
SpaceX wows world with a ho-hum launch of a reused rocket, landing it on a tiny boring barge
Re: Uncanny
"Perhaps the Law of Conservation of Stupidity will soon mathematically establish that every time someone does something wonderful, clever, beautiful then a counter-balancing amount of cretinous mendacity is unleashed?"
To true I am afraid.
Hopefully all the cretinous mendacity released by the current US administration will be counterbalanced soon by something breathtakingly clever and beautiful.
One can but hope
Re: Amazing
Wonderful, wonderful wonderful stuff. Takes me right back to the excitement of the Apollo era of my youth. Astronauts and cosmonauts simply trump ALL other celebrities for sheer cool!
Big thumbs up to all rocket scientists, engineers and all other staff at Space-X for making this happen. I will certainly raise a glass (or two, it's Friday after all) this evening
Boffins crowdsource hunt for 'Planet 9'
Interesting project, and it is great getting the public involved, but I do wonder: if good old image processing can already highlight moving stuff, how hard should it be to generate candidates automatically? It is fairly easy to compare the position of moving stuff to known objects. No AI needed for that: if the epoch of the observation is known, then the position of known objects can be computed from their orbital parameters. Sometimes these parameters might not allow very precise positions, but we could compensate for that by plotting the uncertainty of the position on the night sky. If an object falls into the region that might be occupied by a known target, and there is no other clear candidate for the known object, provisionally rule it out.
We also know roughly how fast it should be moving (slow(!) it is far away from the sun) so we can rule out a lot of the unknown objects too. Finally, we have some idea of the brightness (because it is most likely an ice giant with a mass ten times that of earth, and we have a rough guess of the distance). This too will rule out many objects. I would love to write some code to scour these data automatically. Might be a nice student project
Our Sun's been using facial scrub: No spots for two weeks
I have been keeping tabs on sunspots for quite a while, and the last minimum was way deeper than I have ever seen in almost 40 years of observing. The current minimum looks set to get very deep as well. I know several amateur astronomers who wonder if it currently makes sense to invest in (expensive) H-alpha or Ca-K filters at this point in time. I look at it this way: If a new Maunder minimum is imminent, we will be the first to be able to observe that in great detail. If not, we can carry on observing interesting detail.
BTW, one small sunspot group was visible in my shots from yesterday in white light and Ca-K (393nm). Still very quiet though
TRAPPIST-1's planets are quiet. Quiet as the grave, in fact
BOFH: Don't back up in anger
Great episode
I had almost expected some questions on how you could "accidentally" secure erase a file. You could even envisage some conversation on "secure erase protocol requiring mandatory erasure of all back-ups to ensure ISO-9001 standards compliance following IEEE-235478523676 standard on complete data security, in line with EU directive 72563762357-2016 clause 42-B"
Come to think of it, I might try that line of what can pass for reasoning in low light
Dark matter drought hits older galaxies: Boffins are, rightly, baffled
Re: re: Many, many phenomena in physics were predicted long before they could be detected
It does look like a bodge though. Calling it "amending the model to fit the data" sounds better as long as you don't think about it for too long.
Better than amending the data to fit the model, isn't it? The cyclic process of observation/experiment -> derive theories that explain the observation/experiment -> design observations/experiments that can falsify the theory -> back to step one, means that as we develop and refine theories, they evolve into forms that explain more and more. For example, Newtonian gravity explained far more than Plato's or Aristotle's idea of gravity, which was essentially separate from their model of planetary motion. Newton unified it, and explained the elliptical, rather than circular orbits put forward by Kepler. What it couldn't explain was the precession of the perihelion of Mercury. Einstein's general relativity put forward a single framework that explained both Mercury's odd, and the other more regular orbits, plus a whole lot more.
Replacing general relativity may well be called for to rid ourselves of dark matter/energy, but at the same time the replacement must explain all previous stuff as well.
Oxford Uni boffins say internet filters probably won't protect teens
Re: Anon for reasons - Basically to avoid the SJW'ers
It really is as simple as teaching your kids that's ok to come and speak to you if they are unhappy with the content.
Absolutely right. And anyway, what I find hypocritical (if not downright disturbing) is the focus on blocking sexual content. I find violent content far more upsetting than sexual. After all, sex is a natural experience which aims to be pleasurable to all sides involved (if it doesn't we are talking violence again), whereas violence might be natural, but certainly doesn't aim at being pleasurable (unless some seriously deranged people are involved). I still would prefer talking about violent content (which could simply be news footage from war zones) with my kids than blocking it.
User lubed PC with butter, because pressing a button didn't work
AMD does an Italian job on Intel, unveils 32-core, 64-thread 'Naples' CPU
Very interesting
Currently testing a new MPI algorithm for seriously big images (max 38.6 Gpixel and growing) on our cluster, and getting some pretty good speed-up up to 64 and even 128 and 256 processes (reaching rates of >300 Mpixel/s for a complex image processing task). A cluster of beasts like this would be very nice to test this on further.
Very nice indeed
IBM has cloud access to quantum computer 400 times smaller than D-Wave system
BONG! Lasers crack Big Ben frequency riddle BONG! No idea what to do with this info BONG!
Clouds can compete with HPC, say boffins
Interesting, but ...
1: "The researchers believe the biggest determinant of scalability is the interconnect in place"
Well, well, well. Who would have thought? I finally understand why a distributed-memory set-up with Infiniband interconnect is better than an old Beowulf-style cheapo Ethernet interconnect. </sarcasm>
2: Speed-up isn't always the best measure: I might achieve near linear speed-up, but without knowing the performance on a single machine, I still know little about actual performance.
3: Linpack is ubiquitous, but it doesn't necessarily reflect real-world applications (although much code is written on the back of Linpack. I am sure some of our code would not necessarily run well on the cloud (we are testing it)
4: 32 nodes is not huge by HPC standards, unless each node is really massive in terms of numbers of cores. The Edison system has 5586 nodes. Scaling computation over those numbers is a very different ball game.
The Register's guide to protecting your data when visiting the US
Re: Timely advice
I am going to see the eclipse, as I had booked tickets already. I am wondering what they will make of all the kit I will be carrying to record the event, but in my experience the response to carrying telescopes and the like is generally that you get put into the category "Harmless nerd". I'm OK with that, although "Mostly harmless" may be more accurate
The advice on remaining polite is important. I have only once had a really grumpy customs officer in the USA, and remaining polite got me through that (unpleasant) experience without much trouble. All the other times I was treated with courtesy.
University DDoS'd by its own seafood-curious malware-infected vending machines
Pulsating white dwarf described as a 'dynamo' found, no, not in the back pages, 380 LY away
Last Concorde completes last journey, at maybe Mach 0.02
A non-Standards Soviet approved measure of weight? Sod off, BBC!
2016: Snapchat loses $515m... 2017: Snapchat rips veil off $3bn IPO
Apparently, profit is so twentieth century
OK, I am in no way an economist or financier, but I am good at mathematics. The only "rationale" I can see to buy shares that do not promise to yield any dividend (due to lack of profits) is to gamble that the share price will go up. This assumes that there are more people that will prepared to make that gamble, because that is the only rationale to buy these shares. The moment people no longer assume the share price will go up the value will automatically come crashing down. This suggests that like with a pyramid scheme, only the initial investors have any chance of making money. Of course, it is not a pyramid scheme per se, but it is getting close, and I wouldn't touch these shares with a 20 foot pole
'Mafia' of ageing scientists, academics and politicos suck at picking tech 'winners'
Re: Industrial policy versus science funding
@Shultz
Very good points. We see a similar trend in the Netherlands: science funding of big projects, preferably with industry. This can and does produce some good science, and the odd bit of new product development. However, there is less and less funding for fundamental science, and in particular the smaller "what if?" projects for just one PhD student who can work on some weird idea of their own or their supervisor. In such projects you often end up solving totally different problems than we set out to do along the way. They rarely fail in their entirety, for even if the original goal isn't reached, the serendipitous discoveries along the way, and the insights gained on why something didn't work are still valuable in their own right. Not necessarily valuable in terms of money (what price do we put on the Schroedinger equation?), but valuable for other scientists to build new ideas on, and not least in having trained another intelligent person with an inquiring mind. Money spent of research projects is very much money spent on education, after all. Besides, many scientific breakthroughs only bear fruit much later, and often in unexpected areas. If we hadn't gone into the weird world of quantum mechanics, we would probably not have developed the transistor, which would be a pity, as I am quite a fan of the transistor.
This is not to say collaborations with industry should not be encouraged. I get a lot of ideas from problems facing such industrial partners as I have, and in computer science in particular, the distance between theoretical concept and practical application is small. I might develop the theory of some new image processing tool in the morning, develop an algorithm in the afternoon, and have a working prototype the next day. In e.g. material science, this is unthinkable.
Science is an evolutionary process, with ideas mutating and recombining in new ways, and being submitted to selection by experimental validation (or more properly, ideas survive as long as we cannot falsify them), and by allocation of funding. In any evolutionary process, we can easily get premature convergence on suboptimal solutions by applying too much selection pressure. By funding a few, big projects in narrow fields, we run precisely that risk.
Felted! AI poker bot Libratus cleans out pros in grueling tournament, smugly trousers $1.8m
Cassini sends back best ring-shots yet en route to self-destruct dive
God save the Queen... from Donald Trump. So say 1 million Britons
I actually quite enjoy shouting at the telly from time to time. A bit of a variant of punching a sack of potatoes like the Silastic Armourfiends of Striterax, as an alternative to the healthy and natural channeling of aggressive instincts in deeds of senseless violence.
Doffs hat (roo leather Barmah today) to the late, great Douglas Adams
Devonians try to drive Dartmoor whisky plan onto rocks
Naughty sysadmins use dark magic to fix PCs for clueless users
Re: "Mechanical Sympathy" and magic
This is just one specialised example of Murphy's Law, called the Inverse Demo Effect (IDE). The Regular Demo Effect (RDE) states that the chances of a program being demoed crashing is a steeply increasing function of the number of people watching, potentially weighted by the Embarrassment Factor (EF) which increases the chance of a crash with the importance of the event, or pay check of those watching. Conversely, the IDE states that the chances of a bug or crash occurring is inversely proportional to the number of sys-admins or developers watching
Chinese bloke cycles 500km to get home... in the wrong direction
President Trump tweets from insecure Android, security boffins roll eyes
Douglas Adams nailed it
Those who most want to rule people are ipso-facto those least suited for the job. If anyone is vain enough or stupid enough to get himself elected president, he must at all costs be prevented from wielding any real power. Thus, the president's job is to draw attention away from power.
We clearly urgently need to find some harmless recluse with a cat he calls "The Lord"
Congratulations – you're looking better than ever this morning!
Interesting stuff
Might this new technology be able to spot a load of surplus hot air around the Washington DC area?
On a more serious note: great work by the engineers. The high speed and resolution open up interesting possibilities to do high-spatial resolution time series analysis at decent temporal resolution.
'It will go wrong. There's no question of time... on safety or security side'
Formal proofs have their limits too. I have used formal methods to prove algorithms correct, but the correctness proof very often (if not always) has a set of preconditions. If the actual input means the preconditions are violated bets are off. Besides, even if my algorithm is correct, I must then show that my implementation is correct, and that my compiler is correct, and that the CPU is correct (remember the old Pentium bug?). I found (ages ago) that in MS Pascal the statements
current := current^.next^.next;
and the code snippet
current := current^.next;
current := current^.next;
had a very different outcome, even when used (correctly) in a linked list with an even number of nodes. The first version caused a crash of the program, the latter worked flawlessly. Both are formally correct, but the compiler apparently didn't handle the double indirection correctly.
This is not to slag off formal proofs, just to say they are not the full answer
NASA fires first shot in plan to bring a chunk of asteroid down to Earth
Blockchain: A digital 'golden section' that's the 'gestalt of its pieces'
Euro space agency's Galileo satellites stricken by mystery clock failures
Japan's terrifying techno-toilets will be made foreigner friendly, vow makers
Now that's a Blue Screen of Death: Windows 10 told me to jump off a cliff
Terry Pratchett's self-written documentary to be broadcast in 2017
Re: A man is not dead
And so Sir Terry lives on in the overhead, bouncing from one end of the Grand Trunk to the other,
and in also in L-space of course, or wherever else orangutans say "Ook",
and in the hearts of all his wit, and deep humanity touched.
...
And now will have to go and get a new handkerchief. I'll raise a glass once more later
Peace-sign selfie fools menaced by fingerprint-harvesting tech
Re: Iris scan
It can in principle, and identifying people in photographs using their iris patterns has been done (the famous Afghan refugee girl featured on the National Geographic magazine is a well-known example). To gain access using a photo could be done, but a good iris scanner can (or rather should) check it is a real iris. Any digital print usually shows a very fine regular pattern that stands out hugely in the Fourier spectrum of the image, besides, changing illumination levels causes a real iris to contract. This can be detected easily. Note that iris scans are preferably done in infra-red, so when printing the captured iris, you may also need to get the right reflection in that band as well (certainly possible, not perhaps trivial in your regular ink-jet printer), and indeed capture may need to be done in IR (some DSLRs allow that).
Oh ALIS, don't keep us waiting: F-35 jet's software 'delayed'
Anti-smut law dubs PCs, phones 'pornographic vendor machines', demands internet filters
How would such a filter even work?
As a person working in image analysis and computer vision research, I have to wonder how we are to identify "obscene images" from artistic nudes (although an algorithm to detect urns could help, according sgt. Colon), or even medical images.
Or do the legislators want to vet all the smut sites personally.
New Windows 10 privacy controls: Just a little snooping – or the max
Hmm
"Engineers, with permission from Microsoft’s privacy governance team, can obtain users' documents that trigger crashes in applications, so they can work out what's going wrong, from people's machines running in "full" mode."
So what exactly is in place to stop an engineer from putting in a forged request to access a users files, and stealing important stuff? Microsoft’s privacy governance team? How do they set about checking the files are really necessary? How do they ensure the data are discarded after the problem has been sorted?
Why not ask the user for a file that triggered a crash of a program? Is that not actually far less work than going through some internal privacy governance team? It would also look MUCH more customer friendly. And of course the user has the chance to say that the data is confidential and they can go suck a neutron star.
Somehow I think "full" mode is out of the question for many, if not most professional users.