Slashdot Mirror

Two simple reasons.

• IE5 cannot handle 100% html compliant gzipped postscript files. See this page from xxx.lanl.gov, which is the premier repository for preprints in mathematics and physics.
• ie5 isn't available for enough operating systems.

Yes, for browsing web pages in any of the Windows operating systems, ie5 is much better than Netscape 4.x. You'd be crazy to say otherwise. But, the way it is now, I have to close ie5 and open Netscape in order to view research papers. I know I can use ftp and gzip and a postscript viewer to do it, but it's easier with a browser, now that the preprint servers have nice front ends.

I know people say Mozilla is open source and therefore better. I agree, abstractly, but most people ain't gonna read it anyway. It does give a kind of confidence that nothing sneaky is going on (like personal information being tranmitted to web sites) and that is important to me. But when it comes to performance, the inability to handle gzipped postscript is absolutely unacceptable. I have heard of workarounds, but haven't been able to carry them out.

That's my honest opinion.

Two simple reasons.

• IE5 cannot handle 100% html compliant gzipped postscript files. See this page from xxx.lanl.gov, which is the premier repository for preprints in mathematics and physics.
• ie5 isn't available for enough operating systems.

Yes, for browsing web pages in any of the Windows operating systems, ie5 is much better than Netscape 4.x. You'd be crazy to say otherwise. But, the way it is now, I have to close ie5 and open Netscape in order to view research papers. I know I can use ftp and gzip and a postscript viewer to do it, but it's easier with a browser, now that the preprint servers have nice front ends.

I know people say Mozilla is open source and therefore better. I agree, abstractly, but most people ain't gonna read it anyway. It does give a kind of confidence that nothing sneaky is going on (like personal information being tranmitted to web sites) and that is important to me. But when it comes to performance, the inability to handle gzipped postscript is absolutely unacceptable. I have heard of workarounds, but haven't been able to carry them out.

That's my honest opinion.

What a great article! There's a lot to get through there, so I thought I'd add a bit of extra info about one bit I know well.

Tim mentions how much more clear it is to use direct assignment (A=B+C) for all types (including lists) than having to worry about the details of the specific classes. This issue is part of the study of generic programming. This is generally quite easy to do with functional languages (e.g. have a look at some Mathematica programs to see what I mean).

A lot of effort is going into making this available in C++, with a particularly success being achieved by the STL. The STL gets over the problem of speed through using templates, which (as Tim mentions) are dealt with at compile time rather than run time. Although Tim apparently has no time for templates, they certainly achieve good results in the STL.

Those of you with their thinking hats on will be wondering "but what happens if I write A=B+C+D?". The STL would create two loops and two memory copies, since it thinks of it as ((B+C)+D). Some very interesting work to avoid this problem is being implemented in class libraries such as POOMA and Blitz++ which implement the fiendishly clever concept of expression templates to actually create and traverse a parse tree of the expression at compile time. Of course, this can lead to very long compilation times. An amazing side effect of this approach is that if you never actually use a value in A (having assigned B+C to it), then it will never get evaluated (e.g. the classes use 'lazy evaluation').

If you're interested in trying out some of these ideas in a strongly typed language, it's worth looking at C++. You'll need a good compiler though--if you're using GCC make sure it's the latest version, or for better results still try KAI C++. If you're a complete newcomer try out a functional language--I particularly recommend Mathematica, although you can use the functional concepts in Python (e.g. Map()) and Perl to make a start.

What a great article! There's a lot to get through there, so I thought I'd add a bit of extra info about one bit I know well.

Tim mentions how much more clear it is to use direct assignment (A=B+C) for all types (including lists) than having to worry about the details of the specific classes. This issue is part of the study of generic programming. This is generally quite easy to do with functional languages (e.g. have a look at some Mathematica programs to see what I mean).

A lot of effort is going into making this available in C++, with a particularly success being achieved by the STL. The STL gets over the problem of speed through using templates, which (as Tim mentions) are dealt with at compile time rather than run time. Although Tim apparently has no time for templates, they certainly achieve good results in the STL.

Those of you with their thinking hats on will be wondering "but what happens if I write A=B+C+D?". The STL would create two loops and two memory copies, since it thinks of it as ((B+C)+D). Some very interesting work to avoid this problem is being implemented in class libraries such as POOMA and Blitz++ which implement the fiendishly clever concept of expression templates to actually create and traverse a parse tree of the expression at compile time. Of course, this can lead to very long compilation times. An amazing side effect of this approach is that if you never actually use a value in A (having assigned B+C to it), then it will never get evaluated (e.g. the classes use 'lazy evaluation').

If you're interested in trying out some of these ideas in a strongly typed language, it's worth looking at C++. You'll need a good compiler though--if you're using GCC make sure it's the latest version, or for better results still try KAI C++. If you're a complete newcomer try out a functional language--I particularly recommend Mathematica, although you can use the functional concepts in Python (e.g. Map()) and Perl to make a start.

[From the Software Carpentry project coordinator]

Thank you all for your postings regarding the Software Carpentry project. To answer some of the points that have come up several times:

This is a design competition, rather than a programming competition. Good entries should be relatively language-neutral --- we believe that at the 5000-word level, the similarities between modern object-oriented languages (C++, Java, Python, etc.) are more important than their differences.

Designs based on existing tools are very welcome. If, for example, you think the only way to meet the criteria for the "build" category is to extend the syntax of standard Makefiles, then please submit that as a design. (However, for the reasons discussed in the FAQ, if your plan for an implementation is simply to provide a Python scripting interface to GNU Make, you'll have to convince the judges that there's no "pure Python" way to achieve the same ends.)

No, Software Carpentry is not a company looking for some publicity. The project is being funded by Los Alamos National Laboratory, who believ that computational scientists and engineers need easier-to-use software engineering tools, and administered by CodeSourcery, LLC, who believe that those tools would be of use to the whole Open Source community. The FAQ talks about LANL's reasons for funding the project, as does this article from Doctor Dobb's Journal.

Yes, one of the project's goals is to give up-and-coming software designers a chance to get some attention, just as architects and classical musicians do.

Yes, the competition is open to submissions from any country.

No, this is not part of some perfidious Pythonesque plot for world domination :-). We thought very seriously about using Perl for the implementations, but after teaching classes in both Perl and Python at Los Alamos National Laboratory, came to the conclusion that the latter had a gentler learning curve. (This is not meant as disparagement of Perl as a tool for full-time professional programmers, it is simply an empirical observation of computational scientists and engineers.) Neither Guido nor any other member of the Python development team had any part in setting up the project, choosing Python, or choosing the competition categories.

Thanks for your posting regarding the Software Carpentry design competition. To answer your first point, S.C. is not a company --- it is a project funded by Los Alamos National Laboratory (whose scientists and engineers need software tools that are easier to use), and administered by CodeSourcery, LLC (which believes that the project will further the aims of the Open Source movement).

Your points about encouraging secrecy and non-cooperation are addressed in the FAQ. As for preventing people from fixing the things that are perceived to be broken in existing tools, that is certainly allowed --- we welcome designs based on existing software, as long as those designs include a scriptable Python interface. (I personally believe that a re-design that takes good ideas from several existing tools is more likely to win, but I'm not one of the judges...)

Thanks again for your posting,
Greg Wilson
http://www.software-carpentry.com

Thanks for your posting. In answer to your first comment, no, this isn't a company --- as the project home page and FAQ explain, this is being funded by Los Alamos National Laboratory (because scientists and engineers need better software tools), and being administered by CodeSourcery, LLC (who believe that it will further the aims of the Open Source community).

As for existing tools being "good enough", that's addressed in the FAQ as well (but yes, we did forget about sendmail, and will correct that).

Finally, as for this being part of Guido's grand plan, he was not involved in the project in any way until after it had been launched --- in fact, he was the last-but-one addition to the list of judges.

Thanks again for your posting,
Greg Wilson
http://www.software-carpentry.com

Here's a paper showing that the brain can be thought of as a classical mechanical system, not a quantum one. This seems to be relevant criticism of Penrose's theory.

Depending on what you need the cluster for you should adapt network at your clustering technology.

KAOS
PAPERS

Virtually all physics papers are published online before they appear in printed journals. Physicists invented the web, remember?

http://xxx.lanl.gov/abs/cond-mat/9911101

I don't understand the physics ... but I find the idea that local dark matter (yes, within our own galaxy) may be remnants of stars from a far-future universe very unlikely.

Working quantum computers with a reasonable number of qubits will render all current public-key encryption techniques useless, regardless of the key length used. Peter Shor has exhibited explicit algorithms for solving the factorization and discrete logarithm problems in polynomial time using a quantum computer. Since all major current public-key systems (El Gamal, RSA, Diffie-Hellman) are based on one of these two problems (sometimes a generalization of dicrete log to elliptic curves) they are all then breakable in polynomial time. Which means breaking them is a very easy problem given fast quantum computers, not much more difficult than encrypting/decrypting a message using them. The algorithms involved use some nifty randomization techniques, quantum Fourier transforms, and basic group theory; see the paper on quant-ph if you're interested.

As far as symmetric ciphers go, I don't know of any algorithms to break them using a quantum computer. It doesn't seem to be an especially difficult problem, however, and I wouldn't be surprised if the NSA or someone had already written up a few to crack 3DES, Blowish, etc.

Quantum encryption will possibly provide a secure alternative; basically the idea here is that the two parties involved use a shared quantum source to generate identical one-time pads. Artur Ekert showed that the parties involved can use the Bell inequality to detect any interference in the shared quantum data; it appears this is generally true for any sort of interference in the commonly generated data. Now quantum encryption is provably impossible to break as it provides a completely random one-time pad; and it seems likely that if quantum computers become a reality we will have stable enough quantum systems to make quantum encryption a reality.

So the net result of the whole quantum thing will probably be the destruction of all previous cryptographic schemes and the use of quantum cryptography. That's assuming we ever build a useful quantum computer, which is by no means assured, given the large difficulties currently faced in the field.

Also, it's unlikely that quantum computers will ever become a device for common use; simply because they're not a replacement for the classical computer. A quantum computer would require a ridiculous amount of classical control and error-correction hardware to run at all. Also, you wouldn't be able to just program it. Quantum computation requires special algorithms that can be parallelized correctly to work at all; it takes people who know quantum very well to come up with these algorithms.... you can't just plug a program into a quantum computer and expect it to run quickly.

So most likely, even if they become widespread, quantum computers will just be used as an "add-on" sort of module to do massively parallelizable computationally intensive things, like Fourier transforms, prime factorization, etc.

A great deal of the usefulness of quantum computers depends on whether they can solve NP-complete problems. This is as yet unknown, but some preliminary work indicates that using special nonlinear aspects of quantum behavior, systems may be developed which can do this. There is evidence indicating standard quantum computers will not be able to generally solve NP-complete problems.

Ali Soleimani

alis@caltech.edu

Caltech Phys/Math

The richest ressource on quantum computing is the LANL e-print archive on quantum physics. A more structued site dedicated only to QC is the Oxford Centre for Quantum Computation. My own stuff can be found here.

Shor's algorithm allows factorisation with cubic effort on a quantum computer, so RSA would be vulnerable. A simmilar algorithm can be used to break certification schemes based on discrete logarithms.

As for symmetrical encrytion schemes, Grovers database search algorithm allows to serach N items in O(sqrt(N)) steps, which would require the key-lengths to double to prevent a brute-force known plaintext attack. This should apply to all block encryption schemes, but as the effort remains exponential in key-size the consequences aren't as severe as with factorisation-based public-key methods.

Keep in mind, however, that quantum computing is a relatively new discipline which up to now has mainly been carried out by physicist and not by crypto experts, so the actual potential for quantum crypto analysis cannot even be estimated right now.

Shor's algorithm allows factorisation with cubic effort on a quantum computer, so RSA would be vulnerable. A simmilar algorithm can be used to break certification schemes based on discrete logarithms.

As for symmetrical encrytion schemes, Grovers database search algorithm allows to serach N items in O(sqrt(N)) steps, which would require the key-lengths to double to prevent a brute-force known plaintext attack. This should apply to all block encryption schemes, but as the effort remains exponential in key-size the consequences aren't as severe as with factorisation-based public-key methods.

Keep in mind, however, that quantum computing is a relatively new discipline which up to now has mainly been carried out by physicist and not by crypto experts, so the actual potential for quantum crypto analysis cannot even be estimated right now.

http://xxx.lanl.gov/ps/astro-ph/9911314

The URL above gives the abstract in ASCII, but the article itself is in Postscript. If you look at the main article, pay special attention to the figure which shows the REAL "signal" of reflected light from the planet, and the signal from a fake planet inserted into the data. The fake signal is clearly there, but I'm not confident that the claimed signal is.

How did they do it? The basic idea is: they took spectra of the star with very high resolving power, which spread out light into different wavelengths; or, equivalently, spread out light at different velocities. They knew when the planet ought to be coming towards us, and going away from us, in its orbit around tau Bootis, and with what speed. That meant that they could examine _very carefully_ spectra to find hints of spectral features which were shifted -- ever so slightly -- from their ordinary positions. They found weak evidence that some spectral features were shifted to the blue when the planet was coming towards us, and to the red when the planet was going away from us.

But, if you ask me, it's a two-sigma result. Needs more observation.

As for the claims of specific elements ... well, my guess is that the authors saw the weak, time-varying features in spectral lines produced by some element, and the press misinterpreted this to mean that that element was present in/on the planet. If the planet is simply reflecting light like a perfect mirror, it will produce these weak, time-varying features for ALL lines of ALL elements. Given the signal-to-noise ratio in the spectra I saw in their paper, I don't think that the authors are really advocating the presence of any specific elements in/on the planet.

And if they _were_ (which I emphasize is unlikely), it would be elemental oxygen, not atomic oxygen.

A good idea, certainly, but _really_ tough technically. I think that the press is running away with the story.

Michael

Whoops... the original LED scenario was Arkani-Hamed, Dimopoulos and Dvali. Their original paper is here. Sorry!

Yonatan

There's actually been a lot of fuss about what's called "large extra dimensions" recently. The original problem was that the energy scale associated with gravity is about 10^19 GeV (1GeV = the energy an electron would get going through a potential gap of 10^9 V = approximately the mass of a proton) while the energy scale associated with all the other forces of nature is only 10^3 GeV. This is really bad because it means that (for instance) particles would get gravitational fields surrounding them that give them masses on the order of 10^19 GeV, which would turn everything in sight into a black hole.

This problem can be solved in a number of ways - notably supersymmetry, which causes those giant gravitational fields to cancel out. But there's one other odd problem to deal with, which are "extra dimensions." Basically string theory requires that the universe is actually 10-dimensional, and the other 6 dimensions are simply wrapped up very tightly. (Mental picture: If you wrap up a sheet of paper (which is 2-dimensional) into a very tight tube and look at it from far away, it looks 1-dimensional. Unless you're scanning it on distance scales comparable to the radius of the tube.) The problem is that you have to somehow wrap up these 6 dimensions on a really small distance scale (the length scale of gravity, about 10^-42 cm) and keep the other 4 really big. (the size of the universe) This again happens because the energy scale of gravity is big.

So about a year ago, Nima Arkani-Hamed, Savas Dimopoulos, Gia Dvali and John March-Russell had an interesting thought: We don't *know* that gravity really behaves like anything in particular at length scales below about a millimeter. (The current limit of experiment is about 0.8mm) So they noticed that the following setup gives the right answers too:

* We live in a universe with however many "extra" (small, rolled-up) dimensions, but these are rolled up with radii on the order of somewhere between 1fm (10^-15m, the size of a nucleus) to 0.1mm. (The range of sizes is because there are several different models)

* In this loosely rolled-up world, there are these 4-dimensional objects called "branes" floating around.

Then several amazing things happen. First of all, all matter particles (electrons, quarks, people) are bound to the surface of the brane and can't leave it. So are all the non-gravity force particles. (Photons, gluons, etc.) This just follows from the physical properties of branes in string theory, and it means that as far as anything but gravity is concerned, the universe is 4-dimensional and we won't see the extra dimensions.

Second, gravity completely ignores the brane (except insofar as there's matter, and therefore sources of gravity, there) and flies around freely in all of the dimensions. But because some of them are rolled up, what happens is that at long distances (bigger than the radius) all the gravity gets "squeezed" along the extra dimensions and gravity behaves like ordinary 4-dimensional gravity. At short distances, this changes -- for instance, the 1/r^2 force of gravity becomes something like 1/r^4.

But the real magic is, if the fundamental energy scale of gravity was 10^3 GeV, (the same as the scale for everything else) the distortion of gravity by the rolling up of space would make it seem like the scale was 10^19 GeV to any observer looking at distance scales bigger than the radius!

So the bonus of the Large Extra Dimensions (LED) scenario is, everything has the same energy scale, and it only seems that gravity has this high energy scale because we're looking at too long a distance. And all of the problems of a high energy scale indeed go away.


Of course, you can ask what the hell any of this has to do with reality. The thing is that all of this is consistent with all experiments to date and explains several tricky points. More importantly, it is experimentally testable; part of the testing happens in tabletop experiments (groups at Stanford and at NIST in Boulder are working on measuring gravity at distances down to about 10^-6 m) and part of it in accelerators. The final tests (thumbs up or thumbs down) will come from experiments at the LHC accelerator in Geneva, which should (knock on wood) be up to spin around 2004/5. Final results should take a few more years after the machine comes on-line.

But disclaimer: At this point this entire scenario is conjecture. People are already working out "observational experiments" to check these models -- for instance, whether these are consistent with the known spectrum of cosmic rays -- which are strong experimental constraints. But until the final experiments happen we can't be certain, one way or the other.

Also, since the original paper came out there have been several modified versions of the conjecture, which differ essentially in technical (but very important) points. The Randall-Sundrum model is especially important, and today's model looks to join the list of candidates.

So what does this mean for us? First of all, if it's right then the underlying scale of gravity is only 10^3 GeV, which is definitely accessible with the next generation (LHC) of accelerators. This means we can start to directly monkey around with the processes associated with black hole formation and the origins of the universe. Apart from completely changing physics (by making quantum gravity experiments practical) this is one of those things that creates more applications than we know what to do with. Making small black holes (and no, they wouldn't eat up the planet. :) is one thing. In some of the models effective FTL travel may be possible.

But possibly the most interesting thing is that there's no reason at all for our brane -- the one that our universe lives on -- is the only one. In fact, the most reasonable model suggests that there is some unbelievable number of branes floating out there, maybe 10^24 of them. It's not clear that the laws of physics would be the same on all of them -- e.g. the speed of light may be different, or the charge of the electron, or whatever -- but if the scenario turns out to be true, it is possible (though difficult) to communicate between two different worlds.

And for my money, that's the neatest thing of all.

I'm thinking in particular of the Blue Mountain machine at LANL -- this was/is a big set of SGI Origin 2000 boxes; I don't actually remember offhand, but I seem to recall that they ran Irix, etc, and used typical MPI message-passing crap for most applications. Never mind. Anyway, point is that there were some problems with getting a linear performance increase on the machine as more and more processors were added; we're talking, incidentally, of thousands of processors. Some problems were to be expected -- clearly it's naive to expect totally linear speedup as you increase the number of processors, even for extremely well parallelized code. But some of the things we saw were just ridiculous -- there was, for instance, a sort of catastrophic breakdown at about 4000 processors; this has probably been solved for some time now, but not without an awful lot of work being done to do so. For some more info about this, check out this page at LANL, among others.

Don't get the impression, btw, that I'm against SGI getting into this area -- I happen to like the company quite a bit, think they have some extraordinary products, etc etc. I'm just a little wary these days when I see them talking about massively parallel machines.

Cheers.

The real action is lower down Avalon was top 100 now is down to 265. The top cluster goes to cplant take the award for top cluster now.

Here's a link to the real Loki.

16 Pentium Pro Processors
x 5 Fast Ethernet interfaces
+ 2 Gbytes RAM
+ 50 Gbytes Disk
+ 2 Fast Ethernet switches
+ Linux
------------------------------
= 1.2 real Gflops for $63,000


---

An interesting paper from August titled Ultimate Physical Limits to Computation by Seth Lloyd of MIT has been submitted to Nature. The abstract and article in TeX, PS and more can be found at http://xxx.lanl.gov/abs/quant-ph/9908043 .
It's a math heavy, but still accessible to the lay audience, and a fun read insomuch as Lloyd goes so far as to talk about outlandish theoreticals as black hole computers.

Blue Mountain has two parts, an open and a secure side. As far as I can tell (and this is not terribly surprising), only details on the open side are available from press releases, etc. Anyway, it's a big beefy SGI Origin2000 system, with lots and lots of boxes each holding lots and lots of processors. (Sorry about the vagueness here -- you can probably find details if you look hard enough.) We're talking thousands of processors here, in case that wasn't abundantly clear.

My slightly-biased opinion would be that, in light of the many millions of dollars which were undoubtedly spent on said machines, it is extremely unlikely that the cluster would be ditched anytime in the near future, even if we end up getting a faster cluster -- you can always use more computing power. :-)

For lots more info, check out this press release, which gives some (now outdated) details on nirvana, the open part of blue mountain. Also, the ACL site at Los Alamos is pretty good, though a big PR-y. It also has details about the (currently extant) Linux cluster, in case you're interested. Finally, if you're curious about the real details of the Blue Mountain operating environment, you can take a look at this page, which has lots of good info.

Have fun.

Blue Mountain has two parts, an open and a secure side. As far as I can tell (and this is not terribly surprising), only details on the open side are available from press releases, etc. Anyway, it's a big beefy SGI Origin2000 system, with lots and lots of boxes each holding lots and lots of processors. (Sorry about the vagueness here -- you can probably find details if you look hard enough.) We're talking thousands of processors here, in case that wasn't abundantly clear.

My slightly-biased opinion would be that, in light of the many millions of dollars which were undoubtedly spent on said machines, it is extremely unlikely that the cluster would be ditched anytime in the near future, even if we end up getting a faster cluster -- you can always use more computing power. :-)

For lots more info, check out this press release, which gives some (now outdated) details on nirvana, the open part of blue mountain. Also, the ACL site at Los Alamos is pretty good, though a big PR-y. It also has details about the (currently extant) Linux cluster, in case you're interested. Finally, if you're curious about the real details of the Blue Mountain operating environment, you can take a look at this page, which has lots of good info.

Have fun.

Blue Mountain has two parts, an open and a secure side. As far as I can tell (and this is not terribly surprising), only details on the open side are available from press releases, etc. Anyway, it's a big beefy SGI Origin2000 system, with lots and lots of boxes each holding lots and lots of processors. (Sorry about the vagueness here -- you can probably find details if you look hard enough.) We're talking thousands of processors here, in case that wasn't abundantly clear.

My slightly-biased opinion would be that, in light of the many millions of dollars which were undoubtedly spent on said machines, it is extremely unlikely that the cluster would be ditched anytime in the near future, even if we end up getting a faster cluster -- you can always use more computing power. :-)

For lots more info, check out this press release, which gives some (now outdated) details on nirvana, the open part of blue mountain. Also, the ACL site at Los Alamos is pretty good, though a big PR-y. It also has details about the (currently extant) Linux cluster, in case you're interested. Finally, if you're curious about the real details of the Blue Mountain operating environment, you can take a look at this page, which has lots of good info.

Have fun.

In spite of the intresting "nitro" comments it should only take an icecube to cool the sensor down if it's exposed.
One would think it would not be burryed deep within the soda machine where it could cool the sensor down itself along with the sodas.

However one of the news reports I heard on this suggests that this is all handled by remote controll and the machines are not indupendent.
This could mean the soda machines get the tempiture for the area the same way I do.. over the Internet.
I use wmWeather to get my local weather and it shouldn't be a big effort for a soda machine to do the same.

Accually reading Cokacolas responce I wonder if this has anything to do with price. It's posable Coke is just extending the old coke finger into giving vending maching owners Internet controll.
One of the things an owner COULD do is jack up the price when it gets hot but he can do that anyway just not by remote.

However a hack for the remote controlled coke machine would be to break it's Internet connection late at night when it's cold.
Or better yet.. just stock up late at night. There are enough night geeks that your local night geek can get the sodas for the day geeks.

If the soda machine is at your office and has an internal sensor then just make shure it's inside where it's nice and cold.

Work done by Ford and Roman showed that to have a wormhole you need stupendously large amounts of negative energy density to maintain the throat of the wormhole (what you would travel through). Think millions of galaxies worth of mass stretched over an area smaller than 10^(-19) meters thick.

Work by Taylor, Hiscock, and Anderson (yes I'm Taylor) showed that even the simplest kind of matter field you can imagine, a scalar field, will not support the wormhole throat. The other fields' (fermionic, electromagnetic, and gravitational) effect have not been calculated because they're much too difficult in this case, but in simpler cases they vary from the scalar field by a factor of 2-4. This isn't nearly enough to maintain the wormhole throat.

Because of all these problems (and others), most everyone has abandoned wormhole research for the time being. There are hints (work by Tanaka) that it may be possible, but it seems very unlikely in the general case that we can create or maintain a wormhole (if we happened to find one suddenly).

Work done by Ford and Roman showed that to have a wormhole you need stupendously large amounts of negative energy density to maintain the throat of the wormhole (what you would travel through). Think millions of galaxies worth of mass stretched over an area smaller than 10^(-19) meters thick.

Work by Taylor, Hiscock, and Anderson (yes I'm Taylor) showed that even the simplest kind of matter field you can imagine, a scalar field, will not support the wormhole throat. The other fields' (fermionic, electromagnetic, and gravitational) effect have not been calculated because they're much too difficult in this case, but in simpler cases they vary from the scalar field by a factor of 2-4. This isn't nearly enough to maintain the wormhole throat.

Because of all these problems (and others), most everyone has abandoned wormhole research for the time being. There are hints (work by Tanaka) that it may be possible, but it seems very unlikely in the general case that we can create or maintain a wormhole (if we happened to find one suddenly).

We conclude that Murphy's proposal (radiation of the power of the main-bus electrical systems from the rear of the craft) can not explain the anomalous Pioneer acceleration.

For those interested, the original articles and articles that cite (and comment) it can be found from the web: gr-qc/9808081

Running this one down took a little leg work, seeing as how the BBC did not elect to give the names of the researchers involved. As best I can tell the BBC has mixed up two separate lensing events. The paper that appears to have triggered the story is probably this paper on MACHO-97-BLG-41, since that is the most recent paper claiming a gravitational lensing planet detection. However, that paper is about a 3-Jupiter mass planet orbiting a binary star system, an interesting find, to be sure, but a far cry from an earth-sized planet. So, even if that is the article the BBC is responding to, it's not the one they're talking about.

The article mentions that the event was observed in 1998 and involved an earth-sized planet, so that sounds suspiciously like MACHO-98-BLG-35, but that paper came out (as a preprint) back in May, and it was announced at the January AAS meeting, so it's a little surprising to see a news article on it just now, unless it's just now appearing in the journals.

Anyhow, assuming the event is 98-BLG-35, there's more to the story. The PLANET collaboration also monitored this event, and they found no evidence of a planet in this system. As far as I know, the status of this system is still under dispute. Unless some problem has been found with PLANET's data, I think it's a little early to claim that an earthlike planet has been detected.

To get the scoop on microlensing, its application in planet searches, and the other things we can learn from it, I recommend PLANET's web page. Among other things, they talk about why microlensing is more sensitive than radial velocity studies (the technique that has produced most of the other extrasolar planet detections) to planets in star systems similar to our own solar system.

-r

Running this one down took a little leg work, seeing as how the BBC did not elect to give the names of the researchers involved. As best I can tell the BBC has mixed up two separate lensing events. The paper that appears to have triggered the story is probably this paper on MACHO-97-BLG-41, since that is the most recent paper claiming a gravitational lensing planet detection. However, that paper is about a 3-Jupiter mass planet orbiting a binary star system, an interesting find, to be sure, but a far cry from an earth-sized planet. So, even if that is the article the BBC is responding to, it's not the one they're talking about.

The article mentions that the event was observed in 1998 and involved an earth-sized planet, so that sounds suspiciously like MACHO-98-BLG-35, but that paper came out (as a preprint) back in May, and it was announced at the January AAS meeting, so it's a little surprising to see a news article on it just now, unless it's just now appearing in the journals.

Anyhow, assuming the event is 98-BLG-35, there's more to the story. The PLANET collaboration also monitored this event, and they found no evidence of a planet in this system. As far as I know, the status of this system is still under dispute. Unless some problem has been found with PLANET's data, I think it's a little early to claim that an earthlike planet has been detected.

To get the scoop on microlensing, its application in planet searches, and the other things we can learn from it, I recommend PLANET's web page. Among other things, they talk about why microlensing is more sensitive than radial velocity studies (the technique that has produced most of the other extrasolar planet detections) to planets in star systems similar to our own solar system.

-r

BQP (the class of problems "efficiently" solved by a QC) is not known to equal NP. There has been evidence either way, so it is not a closed problem yet.

Another introduction can be found off the execelent e-print archive of LALN: http://xxx.lanl.gov/abs/quant-ph/9809016

Polynomial problems are said to be in NP since they are a subset of NP, so it is not right to say factoring is not in NP, since all polynomial problems are "in" NP. There are also problems that are NP hard but not NP complete, meaning they are harder than polynomial, but that a fast solution to one of them doesn't solve all other NP problems.

Factoring is known to NOT be NP-complete.

Factoring might be NP hard, or might be polynomial.

P might be the same as NP.

So fast factoring on a quantum computer doesn't mean all NP problems are easy.

It is not known if quantum computers can solve NP complete problems in polynomial time. Most people doubt it.

Computer scientists certainly are working to try to understand the new complexity classes which result from the presence of quantum computers. I'm not quite sure if they have a class which is NP hard to a quantum computer, but even harder to a classical computer as Kreep suggests, but I am sure people think about such things.

I am not aware of crypto systems based on NP complete problems. Even these might turn out to be breakable, if P=NP or if P=NP when you have a quantum computer. If anyone knows of a crypto system based on an NP complete problem (rather than just one that is NP hard) I'd like to take a look at it.

Nearly all the papers in the field come out first on the web at quant-phys.

John

Whats wrong with the damn thing - it eats brackets...

forgot http:// .. in the link?

I read from the article (e.g. the Postscript version of the original article on LANL), that this protocol can only be used to secure real time communications only.

The protocol cannot be used for email or file encryption. With this protocol one can only make sure, that the line you are currently using is secure.

As timing is a critical of the security mechanism, standard internet with nondeterministic transmission times does not qualify for this protocol.

Nonetheless an interesting article (and as "smooth" to read as any other scientific article).

#define rant_mode
A classic case of hearsay (=RTFM failure) and media hype.
#undef rant_mode

Here's the preprint in ps from LANL, instead; only institutional subscribers have access to the PRL online article.

I agree that peer-review is vital to filter out crackpots and commercial propaganda.

Why?

The various LANL preprint servers have been running since 1991 with no such scoring system. Everything that's submitted from somebody associated with a university or research institution is simply accepted.

People are pretty knowledgeable about their own fields; it doesn't take more than skimming the abstract to see if its written by a crackpot, and only a bit more to see if its interesting to them. (And if it is to them, it need not be to someone even working on very similar problems.) People looking for info out of their fields shouldn't be reading the current state-of-the-art; they need to get up to speed first with, eg, review articles.

So I'd tend to just skip the `rating' totally; other than an ego boost/dasher, I think most scientists would skim with, in Slashdot terms, the threshold set to -2 anyway.

In astronomy, there is a very popular preprint server at Los Alamos where many people submit electronic copies of papers they expect to get published.

However, I usually don't peruse this server. I instead read journal tables of contents (often sorted by topic) or (even better!) newsletters focusing on my particular subfield (star formation). This is for the same reason that I read Slashdot: I don't have the time/energy to peruse everything. Instead, I put my trust in a "portal" of some nature to pre-filter my reading material for me.

I do think it is very valuable to have large archives of electronic materials for reference research (like the NASA/ADS abstract and article service) which is much more effecient than a trip to the library. But I believe the parallel between open source software and scientific reporting breaks down at the bleeding-edge. "Beta" versions of scientific papers are either not worth much (for lack of credibility) or even dangerous (as "bugs" may propagate in the literature).

In the various subfields of Physics, this idea
of a public "preprint server" has been implemented
for some time: check out the Los Alamos
Physics Preprint server.

I've been active in research (astronomy) for
the past ten years or so, and I've had many
conversations with other researchers on the
future of scientific publication. Some of the
main points are:

1. Review/moderation is necessary. There are
a _lot_ of people who have crackpot theories
about the universe, and some of them aren't
shy. Without refereeing of some sort,
the number of scientifically worthless --
see definition below -- papers will grow to
the point that they may swamp the worthwhile
papers. At that point, many users will stop
using the archive.

Note on "scientifically worthless": science
is an enterprise which depends on its
workers to adhere to a set of rules, such
as understanding basic physical principles,
checking the existing literature, creating
falsifiable hypotheses, verifying new
results, repeating experiments, etc. Papers
describing ideas which aren't developed
along these rules are, by definition,
scientifically worthless.

2. Scientists depend on their publication
records to land good jobs, and to advance
in those jobs. At the moment, in astronomy,
at least, the existing
electronic archives are NOT viewed as
"real publications". There's a little bit
of a chicken-and-egg problem: until the
electronic archives are taken seriously,
many people won't publish in them
exclusively. But if everyone publishes
elsewhere, why take electronic archives
seriously?

3. Many people, myself included, worry a great
deal about the use of electronic archives
10 or 20 years hence. I have paged through
bound journals dating back more than 100
years, and used them occasionally in my
research. I can interpret the information
easily. But I don't think it will be an
easy matter to keep electronic media up-to-
date over a century. The librarians to whom
I've talked are _very_ worried about this.

Yes, I know that it may not be difficult
in THEORY to copy old materials to new
formats and new media every N years;
but in practice, it's a royal pain. In an
era of shrinking library budgets, it may
become fiscally impossible.

On the other hand, I do very much support the
idea of "Open Source" publications. It will
enable many more scientists to publish their
ideas. In my field, for example, the authors
have to pay the journals about $125 PER PAGE
for the papers they publish. My last paper cost
over $2000, and I had to pay for some of it
myself (since I work at a small university that
doesn't have a lot of money to support research).

The tricky thing will be to find a mechanism
which keeps the good points of the current
scientific journal system, while avoid the
pitfalls (some of which I've mentioned above).

There is no place for real discussions or peer reviews, however. Most online journals I've seen are pretty weak, just copying what a paper journal does except with lower quality and funding.

The physics archive is at http://xxx.lanl.gov/

In physics, authors have published on the web for a long time now. In Aug '91 (yes eight years ago), a pre-print server was set up at Los Alamos . It initially was just for high energy theoretical physics, but now covers a lot of subjects, and is heavily used. Authors send their papers in (normally before sending them to a normal journal), and everyone can see the new papers the following day. These days, every piece of new research (in high energy physics at least) can be found here. Interestingly, this hasn't stopped the journal publishers much. Papers are still sent to them, and most institutions involved in the subject still subscribe to them. The big difference is that now researchers see each others results as soon as they're ready to publish, and that the papers are much easier to get to. Seems to be a good thing.

xxx.lanl.gov preprint archive

Astrophysics Data System

Each major lab (SLAC, Fermilab, CERN, etc) usually also has a method of accessing articles, but I'll let you find those on your own. ;) -- Bob

Paul Ginsparg's e-print server began as an automated e-mail archive of high-energy physics theory papers in 1991. xxx.lanl.gov (now also known as arXiv.org is now the primary source of online literature in Physics and Mathematics.

It's a cluster of 48 128PE Origins tied together with an 800MB/s GSN (a.k.a. HiPPI) network. The individual Origin boxes are ccNUMA, of course, but the boxes communicate with each other using MPI over the GSN (much like Beowulf clusters do MPI over Ethernet or Myrinet or whatever). There used to be hardware specs on the web, but those disappeared after the China fiasco a few months back. This page gives some details.

--Troy

Tangentially related is this short preprint on "the diameter of the WWW". Talks about how many average hops it is from any given web page to any other, and how this might affect search engines.

I did some reasearch on this for my degree a couple of years ago. I got most of my info from the los-alamos pre-print archive under quant-ph.
http://xxx.lanl.gov/archive/quant-ph
Some useful background information is in Andrew Steane's paper, but you will get heaps more stuff with a quick search.

Remember this site: http://xxx.lanl.gov it's damn useful.

I did some reasearch on this for my degree a couple of years ago. I got most of my info from the los-alamos pre-print archive under quant-ph.
http://xxx.lanl.gov/archive/quant-ph
Some useful background information is in Andrew Steane's paper, but you will get heaps more stuff with a quick search.

Remember this site: http://xxx.lanl.gov it's damn useful.

I did some reasearch on this for my degree a couple of years ago. I got most of my info from the los-alamos pre-print archive under quant-ph.
http://xxx.lanl.gov/archive/quant-ph
Some useful background information is in Andrew Steane's paper, but you will get heaps more stuff with a quick search.

Remember this site: http://xxx.lanl.gov it's damn useful.

I read an artical a few years ago in the quant_physics archives http://xxx.lanl.gov/ about using quantum states to encrypt data communications.

Somebody (one of the telco companies) had demonstrated a configuration where they were using it over about 50km of standard optic fibre.

The idea went along the lines of using a some quantum state to transmit data. Because the other end also knew this state they could test for bits in this state. Probably some sort of polarising filters.

Somebody eaves dropping wouldn't know which space to measure in so would only be able to get a little information out. Since they are doing measurements on the fiber they are destroying the information on the way through so that the receiver would know what was happening and could stop the transmittion before secret information was lost. Long before the snooper had time to do any sort of traffic analysis and learn what code was being used. I guess classical encryption on top of this (giving a statisticaly random bit pattern) would make any analysis impossible.

Most of the discussion was about codes needed to detect intruders while still being able to correct random noise that is common on the quantum scale of things.