Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror

Comment: Re:Particles are baloney (Score 1) 20

Taking a look at the paper, it doesn't appear that these results are highly dependent on the striped morphology, which, I agree, is dubious. These guys aren't using that, they're claiming a "checkerboard" shape.

I haven't read through the whole literature on this - are *all* of the patterns crap? Are there checkerboards, or just noise + STM artifacts?

Comment: PNAS contributed paper (Score 4, Informative) 48

by Badge 17 (#45161223) Attached to: Grand Unifying Theory of High-Temp Superconducting Materials Proposed

Disclaimer: I am not qualified to evaluate the science presented here. However, I always wince when I see something with such big claims as a PNAS contributed paper. PNAS allows National Academy members to "contribute" a paper, i.e. they act as the editor, selecting referees for the paper. This allows well-established scientists to get controversial ideas published without a big fuss - but it also means that sometimes goofy and incorrect stuff can slip through.

Of course, if the theory works out, it will be a huge, huge result. Just add a slightly larger grain of salt than you usually do, because the paper came out of a different peer review process.

Comment: Re:A comedic work? (Score 1) 112

by Badge 17 (#44742227) Attached to: John Scalzi's <em>Redshirts</em> Wins Hugo Award for Best Novel

Quick, name me another prominent comedic science fiction author. Are there really enough that this is an issue? (I actually couldn't think of another one off-hand, if you don't count Vonnegut.)

There is a lot of great science fiction out there. Questions of identity, memory, and continuity? Try Brin's Kiln People. Reaction of societies to profound changes? Robert Charles Wilson's Spin. The Fermi Paradox? Try Revelation Space or Brin's Existence. Also, for "can't find a category for it," try Mieville's The City and the City. If you aren't reading anything past Asimov, you are missing out for no good reason.

Comment: Re:I have the book but haven't read it yet. (Score 2) 112

by Badge 17 (#44742185) Attached to: John Scalzi's <em>Redshirts</em> Wins Hugo Award for Best Novel

I have either read or given up on all of the nominees. I am not convinced that Redshirts was the best novel, but it was probably the intersection of "mainstream / well-known" and "not so bad." Remember, Hugos are determined by a vote of science fiction fans at the convention (or who bought "supporting memberships"), and there's no requirement that they read all of the books.

The nominees were (in order of placement)
5. Blackout - the third in Mira Grant's Newsflesh zombie series. These books are entertaining, and the setting is fairly clever - where zombies are just a fact of life, and an ongoing danger - but this wasn't the best of the trilogy, and the previous two were nominated but didn't win.
4. Throne of the Crescent Moon - an entertaining fantasy novel with a lot of Arabic mythology as an influence. I enjoyed it thoroughly, but it was the author's first novel, and it's not as well-known as the others on the list.
3. 2312 - Another Kim Stanley Robinson book - a little heavy on geology, and a little meandering for my taste, but interesting in bits. KSR has won Hugos before, and is pretty well-known.
2. Captain Vorpatil's Alliance - part of the whole Vorkosigan / Miles Saga from Lois McMaster Bujold. I usually like her books, but something about the voice this one was in just bothered the crap out of me, so I didn't finish it. Once again, Bujold has a lot of past Hugos, and this is like book 15 in the series, so it definitely has a constituency.
1. Redshirts - has Scalzi's usual flaws with fairly bland characters and a little bit of generic plotting. But it was funny, and ended up being at least a little touching and thought-provoking.

Essentially, the fan vote gives a strong advantage to well-known authors. Of the books that apparently had a chance (the top 3), Redshirts might have been the best. (Full stats at http://www.lonestarcon3.org/)

Comment: Other aspects of the paper - health data (Score 3, Interesting) 245

I can't really comment on the slashdot summary, but take a look at the actual abstract: http://eprint.iacr.org/2013/451

"In functional encryption, ciphertexts encrypt inputs x and keys are issued for circuits C. Using the key SK_C to decrypt a ciphertext CT_x = Enc(x), yields the value C(x) but does not reveal anything else about x. Furthermore, no collusion of secret key holders should be able to learn anything more than the union of what they can each learn individually."

In other words, it seems that their technique allows you to encrypt some secret data, then (provably) only release the result of some arbitrary function of that data. It sounds like this means you could (in principle) release health data in encrypted form, then allow researchers to study some ensemble properties of it by giving out the appropriate keys. This aspect of it certainly seems pretty cool.

Comment: Re:He's right (Score 1) 276

by Badge 17 (#43513475) Attached to: Terrible Advice From a Great Scientist

Um... well it's possible that the people I know are exceptions. But the MD/PhDs I know are pretty frigging impressive. One in particular is getting his PhD in experimental physics, and he can argue with the best of them on fairly hard stochastic differential equations.

For the people I know, the MD/PhD is a research degree with added training so that they can effectively and ethically treat people in a research setting. (But I can only speak to the ones I know as a research scientist!)

Comment: Re:He's right (Score 1) 276

by Badge 17 (#43512903) Attached to: Terrible Advice From a Great Scientist

Can't speak too much to your complaints about inconsistency, since they aren't very specific. Certainly MATLAB is fairly idiosyncratic. But your complaints about performance are a little unfair. "People going from naive matlab to naive c++ can get x1000 speed-ups." This is kind of true (though 100x is more my experience). But only if you define naive = someone in their first week of MATLAB coding. A relatively small amount of experience with MATLAB will teach you vectorization, which avoids this problem. If your problem can be vectorized (true of many scientific computing issues) the comparison between "MATLAB by someone who is not a total idiot" and "C++ by someone who is not a total idiot' is much closer (maybe a factor of 2), and probably depends on some of the details of the problem being solved (if the rate-limiting step is calculating a giant FFT, only your FFT library matters).

Comment: Re:He's right (Score 1) 276

by Badge 17 (#43512707) Attached to: Terrible Advice From a Great Scientist

Yes, it does a great deal of that. Since about 2009 or so, a large chunk of basic MATLAB libraries (most matrix operations, including FFTs) automatically use multiple cores with zero added effort.

Here's why I use MATLAB: I'm a theoretical physicist. A lot of my time is spent coming up with models, and then I want to - as quickly as possible - see if these models are practical and sane, and work out the quantitative consequences. I don't need the best efficiency possible, but I do need a giant set of linear algebra libraries, and having built-in ODE and PDE solvers saves a great deal of time. Even for the largest-scale simulations I do, the vast bulk of the computational needs are doing FFTs, generating random numbers, and solving linear systems. These operations are just as fast on MATLAB as in anything else - since it's some highly-optimized library at the core.

There are a few things that MATLAB is just terrible about - performance with for loop-based code, obviously. I have definitely had cases where you can get a factor of 100 from implementing an inner loop in C. You have to use the right tool for the right job.

As for scaling and GPU operations, does anyone know anything that will run on a GPU "without any changes"?

Comment: Re:Astonishing amount to win. He'd better run n hi (Score 0) 308

by Badge 17 (#43199529) Attached to: CCTV Hack Takes Casino For $33 Million

The Reg article and the summary are incorrect. There is no evidence that this was a poker game. In fact, since the Crown casino talks about hoping to recover the money, and not recover it for the players, it might not be. However, I'm not sure what game is played against the house where players' cards are concealed. Maybe they mean looking at the dealer's down card in blackjack? Is Pai Gow Poker dealt with other hands hidden?

Also note, the cheating did not occur solely over eight hands as the Reg claims. According to the Herald Sun article (the original), "cheating was exposed over eight hands of cards played in a short space of time." The cheater got too greedy, and played too obviously.

Comment: Re:Headline is wrong. (Score 1) 308

by Badge 17 (#43199341) Attached to: CCTV Hack Takes Casino For $33 Million

Actually, it could be correct, but there isn't enough information to tell. The Register article claims that the game was poker, but that isn't supported by the original Herald Sun article, which just says "hands of cards." Note how completely devoid of details the Herald Sun article is - they won't say much about the details, the technique, or even what game was being played. This might be because it would identify the players involved.

Comment: Re:"The Tool" (Score 2) 128

by Badge 17 (#42953193) Attached to: Oxford Temporarily Blocks Google Docs To Fight Phishing

It's even worse than this. Occasionally, our University's IT actually does send out emails that sound like a phishing attack. The only difference is that they link to a legitimate website. However, because of the general mess of different sign-ons (e.g. billing, payroll, course schedule, parking, etc...) it takes me a while to remember if this is a real service or a fake one.

I think, somewhat optimistically, that people can be trained to not send username/password over email. However, far too many things reinforce the "go to website linked in email, put in password" message for this to not work some percentage of the time. Maybe we need to normalize "exchange of information" type logins, where you won't input your password until the website provides a signal / response to a challenge?

Comment: Re:For fuck's sake, not string theory! (Score 1) 192

by Badge 17 (#40838353) Attached to: Internet Billionaire Creates Huge Physics Prize

While I generally agree with you about the esoteric nature of string theory, I should correct the record on supersymmetry and inflation (I know you didn't complain about inflation, but it's there further up the thread).

Supersymmetry is an idea with some fairly strong motivations that has driven the last several decades of experimental work in particle physics - there is not solid evidence for it yet, but it is not ruled out. Some of the simplest variants have, however, been ruled out. Here's someone more representative of the experimental particle physics consensus at the moment: http://profmattstrassler.com/articles-and-posts/some-speculative-theoretical-ideas-for-the-lhc/supersymmetry/where-stands-supersymmetry-as-of-42012/
(Should we be giving awards for theory that is not yet proved, but has motivated and clarified our understanding of particle physics? Maybe not, but supersymmetry is definitely not ruled out.)

As for inflation, there is a reasonable amount of experimental evidence to support it; I know fewer astro people, but I would not describe inflation as unsupported, or even necessarily that controversial. http://en.wikipedia.org/wiki/Inflation_(cosmology)#Observational_status

Comment: Re:My immediate response was (Score 5, Informative) 192

by Badge 17 (#40835073) Attached to: Internet Billionaire Creates Huge Physics Prize

Unfortunately, no. Many intelligent young students are already going into high-energy theory and string theory (the primary recipients of this prize). In fact, there are far more students than jobs. I'm a recent PhD from a top physics (and particularly string theory) school. My classmates in string and high energy theory who recently applied for postdocs applied to 100 in order to receive 1 job offer; none of their jobs were in the U.S. These are not permanent jobs; they are usually 2 or 3 year positions, paying $40,000 or so. At the end of this time, you may then enter the lottery for the (literally) one string theory faculty job per year (see http://www.math.columbia.edu/~woit/wordpress/?p=4701 for job statistics). This is what causes students to leave to go to Wall Street, and piping in more money to the already-established best of the best of the field will not change this.

The purpose of this award seems to be to raise the profile of so-called "fundamental" theoretical physics; perhaps it will cause more funding to be directed in that direction, which might be good. More likely, it will simply encourage more optimistic, talented students to step into the meat grinder of a particularly depressed job market, making it even worse, and eventually redirecting another generation's best minds into Wall Street.

I'm not saying don't celebrate physics (I love physics, and am continuing in the field, though on a much more applied topic, where there is more funding) - but there is already enough hype for string theory, and it burns out enough students already.

Comment: Re:weird inference (Score 2) 55

by Badge 17 (#40792591) Attached to: Cray XK6 Supercomputer Used To Simulate Ice Cream

While ice cream makers have probably done lots of experimentation, simulations like this (in addition to being cool physics for other reasons) can give us new directions to look in - e.g how should we change the emulsifier used. More fundamentally, what are the controlling factors for the failure of ice cream? Since TFA is a press release, you can get a better taste for the work the Edinburgh group does from their papers - a related one is at http://arxiv.org/abs/1203.0410

I should also pitch this group's work more generally - Michael Cates' group does a lot of cool simulation work on the dynamics of liquid crystals, as well as active fluids, such as swarming tanks of bacteria. Most of it is on the arxiv, http://arxiv.org/find/cond-mat/1/au:+Stratford_K/0/1/0/all/0/1 and http://arxiv.org/find/cond-mat/1/au:+Cates_M/0/1/0/all/0/1

Some of the behaviors of fluids with active objects (bacteria or molecular motors) are quite unintuitive - if you try to push a sphere through one of these active fluids, it will sometimes go in the opposite direction of the force you apply to it! http://arxiv.org/abs/1204.1279

(I am a soft matter physicist, but not from this group.)

Comment: Re:Working at CERN, wouldn't even want to change (Score 1) 374

by Badge 17 (#40599145) Attached to: To get me to switch jobs, it'd take ...

So what does it take to get a permanent position doing particle physics these days? I'm a newly-minted physics PhD (different subfield, though), and the job market seems fairly moribund.

That said, my classmate who's going into the financial industry is probably going to be paid 300% what I do as a postdoc, and I wouldn't even consider it.

"Bureaucracy is the enemy of innovation." -- Mark Shepherd, former President and CEO of Texas Instruments

Working...