"e=mc^2: 103 years later, Einstein's proven right"

[jinydu]

http://news.yahoo.com/s/afp/20081120...XcH3m9Fz9xieAA

Needless to say, the content of that article is heavily watered down. Any ideas for where I can find a not-so-watered down article? A search through the website of France's Centre of National Research turned up nothing.

[Mini-Geek]

/. linked to a more technical look at it: (I haven't read either link in detail so I don't know the difference)
http://www.newscientist.com/article/...ctuations.html

[jinydu]

Thanks, that's much better. I don't think New Scientist is technical at all; it has no equations in it.

Obviously, the AP's claim that E= mc^2 has never been verified before was way overblown. As far as I can see from New Scientist, the relevant research was about testing one of the predictions of quantum chromodynamics.

[cheesehead]

One commenter wrote:

Really?

Fri Nov 21 10:25:59 GMT 2008 by Skivehard



What really gets confirmed here is the effectiveness of lattice QCD (a way of doing ridiculously hard QCD calculations). The stuff about most (but not all) of the mass of a proton coming not from the valence quarks but from the virtual cloud that surrounds them is kind of old news. Still awesome, but its not really the story here.

[fivemack]

http://www.fz-juelich.de/portal/inde...=650&index=163 is the news report from the place whose supercomputer they used; I get the impression this was at least in part a matter of throwing computation (from a 180TFlop Blue Gene) at the problem. People have built custom supercomputers for QCD in the past, the fastest of which is QCDOC (http://www.bnl.gov/lqcd/comp/) which is basically a custom Blue Gene, but funding means that it's easier to use a pre-existing larger but less custom machine.

Juelich is the big academic supercomputer centre in Europe, and next year they're putting in a machine with 16000 Nehalems and aiming to expand that to a petaflop system in 2009, presumably by adding GPUs since another 64000 Nehalems would be a little expensive.

[davieddy]

I find it depressing that Physics gets less and less tractable.
With maths it's understanble: the low hanging fruit has been taken.

David

[jinydu]

In summary, a much more accurate headline would be:

Supercomputer Calculation Yields More Accurate Estimate of Proton's Mass

[geoff]

Originally Posted by jinydu

In summary, a much more accurate headline would be:

Supercomputer Calculation Yields More Accurate Estimate of Proton's Mass

That is all any science can hope for, answers that are a little less wrong.

[xilman]

Originally Posted by geoff

That is all any science can hope for, answers that are a little less wrong.

Au contraire, if something turns up answers that are radically wrong, that is very exciting, much more exciting than calculating another decimal place and finding it agrees with what you expect.

Just one example will suffice: "as if you fired a 15-inch shell at a piece of tissue paper and it came back and hit you."

Paul

[ewmayer]

So would it be too much to hope that sometime in the *next* 103 years, string theory will provide physics with at least one nontrivial, not-true-by-construction, falsifiable claim about the physical world? How many "Nth string revolutions" do all those incredibly clever folks need to actually get around to some real physics?

[xilman]

Originally Posted by ewmayer

So would it be too much to hope that sometime in the *next* 103 years, string theory will provide physics with at least one nontrivial, not-true-by-construction, falsifiable claim about the physical world? How many "Nth string revolutions" do all those incredibly clever folks need to actually get around to some real physics?

Who knows? It took Einstein the best part of a decade to come up with some testable predictions of GR. It took the best part of a century to find significantly more than the paltry three that Einstein found.

Paul