For instance, you can do anything you like with Linux in the comfort of your own home, so long as you don't distribute the result. But distribution can become a thorny issue if you're careless, and as a result the GPL offers relatively weak protection of your company's super-secret algorithms. After all, anyone who *legally* acquires part of the code, now has full GPL rights to the whole thing, tasty bits included.
Of course, this is only very slightly less protection than, say, trade secret law offers. In either case, the rule of thumb is to keep your secrets secret.
Companies do occasionally get stung by this when they fail to realize that they are actually "distributing" code. Witness Linksys, which was forced to give away its changes to the Linux kernel, probably because it simply never occurred to the developers that selling a device constituted "distribution" of the image in the firmware. The poster's company (or its customers) could conceivably run into the same problem down the line by selling POS units with the customized GPLed drivers installed.
Compare to TiVo, which had the sense to put the interesting code in separate executables that don't derive from anything GPLed.
Don't talk about it just build it. Personally I think someone should sit down and hack together a install package builder program based on something like gdialog and python that outputs a executable compressed image into a single bin file.
Now associate.deb files with that script, and you're good to go. Happy? (I'm sure a RedHat/SuSE/Gentoo user could devise similarly trivial solutions for their systems as well.) (The astute user will throw in a sed script as well to toss version specifier strings.)
Which is, of course, why projects like this never go anywhere. Because if you're going to require that every piece of software be repackaged anyway, to conform to your standard, there's no additional burden in packing them all in the same format. In which case, you've got the Debian project all over again, except that Debian has a ten-year head start and a lot more package maintainers in place.
Sure, there's scads of obscure software lying around out there in TGZs. But the fact that it's still a TGZ would seem to indicate that nobody has a burning interest in repackaging it.
I don't have any concrete information on this at all, but would comparing NASA funding with defence spending be useful as a first estimate?
Well, not quite. Remember that the vast majority of military spending goes to mundane things like salaries, fuel, food, and the like. This doesn't affect the economy any differently than anyone else spending money on those things, and it's actually worse for the economy if it's the government doing the spending, since the tax rate effectively reduces the mean return on investment, which discourages business spending.
However, money spend on research-intensive activity is quite another matter; many people would say that the NSF's budget is some of the best-spent money in the world. This is because research spending not only funds jobs (and with a larger multiplier effect, since most of the money is spent on stuff, which takes jobs to produce), but directly drives investment in technology, as well as potentially creating new technology.
Therefore, if you want to gauge the effect of a new space program, you'd be better off comparing it to the military's R&D budget. Which is, unfortunately, a little harder to actually pin down. Definitely larger than NASA's budget, though.
Have you looked into a folding bicycle? There are many models out there, most of which are designed with the specific goal of getting down to a size where they will go on a train. A bit pricier than a regular bicycle, but all the same advantages in terms of operating cost.
One of the cooler items in this year's
Scavenger Hunt -- each team had to set up a WASTE-alike (physical, not electronic) with drop boxes all over the neighborhood and a maximum 7 hour delivery time (to basically anyone on campus). Since everyone was too busy to check their email very regularly, and the boxes were everywhere, it actually became a preferred way for people doing the Hunt to communicate.
Not so much encrypted, though. Reading the mail your team was delivering became a very funny pastime. I did briefly consider printing out an ascii-armored GPG message, but those things are a bitch to type in.
I don't suppose you have a link handy for that simulation? I'm curious to know what you expect the decay products of a black hole event to actually be.
Furthermore, I think the scenario in which this happens is pretty far fetched. More realistic models (which can avoid proton decay, for instance) usually have to put the scale at which black holes form much higher (100 TeV or higher).
I'm familiar with the argument that you need large extra dimensions to be able to produce gravitational objects in the low TeV domain, but I've never heard proton decay cited as a constraint on this before. I'd be really interested in seeing a reference to this. Can you post one? (To be fair, most every theory these days seems to have something to say about proton stability, eventually.)
If black holes exist sans Hawking Radiation, we'd be in quite a bit of trouble upon the production of even the smallest one. Probably wise to check that little problem out. I'm not advising doing anything wacky and superparanoid, like building it on the Moon
Not really a problem, even without Hawking decay to destroy the little guys. Laboratory black holes would be substantially smaller than a proton, electrically neutral, and thus unable to interact with other matter at all, except via gravity. Moreover, they would almost certainly be moving at relativistic velocities upon production. Or certainly well above escape velocity, at any rate.
V_esc ~ c/100000 so for a singularity with a mass of "hundreds of protons", the excess energy needed to reach escape velocity is on the order of a few GeV. At LHC, that's muon noise.
Therefore, I would expect a non-decaying black hole generated in a collider to simply escape into space harmlessly. With an interaction cross section probably comparable to that of a neutrino, it could even fly right through the Earth and come out the other side, probably without even disturbing many atoms in its path.
(Recall that a flux of a few trillion neutrinos per square meter passes right through the Earth each second. You can calculate that the stopping distance in lead for an average solar neutrino is... about a parsec. I'm not worried.)
Another reason not to worry: if black holes don't decay via the Hawking process, then the universe should be chock-full of primordial black holes with masses ranging from atomic to small planet. None of them seems to have swallowed the Earth yet (or, even more conviningly, the Sun).
There is a reason the pdf uses bit-mapped fonts, and we are it. Our SavHunt team (the Lush Puppies Mark III, FIST Deux, Deleuzian Potato) uses a computer database with a web interface to keep track of our itemss. So, the judges used bit-mapped fonts this year to force us to type in the list by had. Bastards!
(Aside: Anyone know of a method that would allow for a 'any n of m keyholders needed to decrypt' schema? It's something that has advantages, but I've no idea how to go about it)
See Applied Cryptography, chapter 3, "Secret Splitting", for an introduction. I found a toy implementation and mathematical explanation here. Or just Google for "secret splitting".
The evolution of languages differs from the evolution of species because branches can converge. The Fortran branch, for example, seems to be merging with the descendants of Algol. In theory this is possible for species too, but it's so unlikely that it has probably never happened.
No kidding. The evolution of languages is probably different from the evolution of species because, well, they're languages. Paul Graham seems to have overlooked the fact that there's already a fully-formed field of study devoted to things that evolve like languages. It's called linguistics. He's probably heard of it.
When I was a kid, I had a grand old time flipping through the back pages of the unabridged American Heritage dictionary. One item of interest in there is a (then current) language tree illustrating the radiation, convergence, and development of linguistic families derived from Indo-European. The dictionary's editors assured their readers that, while informative, the printed tree was a gross simplification of the real state of affairs.
My point being, I am inclined to suspect that a field that can examine modern languages and scraps of historical writing and extrapolate back to reconstruct languages dead for thousands of years, has the tools to make sense of fifty years of computer languages and get a sense of the general trends going forwards.
However, you should be doing all your homework without being forced! School is already too easy, and if you skip any of it you'll be the only one at McDonald's who can't make change! You should be asking your teachers for extra homework!
Extra homework, eh? Have you actually looked at homework recently? While there are exceptional schools and exceptional teachers, from what I've seen, most of the work being handed out these days strikes me as designed to encourage a slavish devotion to work for work's sake. Mostly it's the same busywork as ever, but there's a hell of a lot more of it then when you were in school. Which means that the occasionally interesting assignments are basically history, since no teacher has the time to grade responses more complex than numeric or multiple-choice now.
So frankly, I think it would be time better spent for anyone still a guest of the mandatory pre-college babysit^H^H^H^H^H^H^H education system to do just enough homework to get by and keep the grades up, and pursue projects of actual interest independently. Or even better, find a good mentor. Heaven knows I didn't get into college on the strength of my grades. Or grad school, for that matter.
Hi there. Nice work on USITE Crerar, then. I've been in there a number of times since the lab was installed. Impressive Seti numbers, too.
Since you ask, I work for Dr. Foster under the rubrick of the Computation Institute on campus. Where on campus that is currently is in a state of flux between Ryerson and the RI.
I think that the USITE computers will continue running Seti for the time being, as we are not suffering from a lack of heterogenous resources, but a lack of software to pull it all together. We've currently got I think twelve grid sites lined up to accept Sloan cluster finding work, not counting the WorldGrid, which we should (hopefully) be able to tie into as well before long. iVDGL might be of interest to you on that front. My group, the Grid Physics Network, is less concerned with these management issues, than with taking the existing tools and wrapping them around real physics problems.
On the other hand, I do know people who wouldn't mind if those lab computers got set up with some Condor daemons, which would be a lot more useful than some single-purpose cluster finding code, anyway, and more in keeping with the grid computing style of doing things. I don't know how much demand there is for WinNT pools, though.
And here you have the conspiracy-theorist motive for such an action: MS isn't afraid of India switching to Linux, but of the millions of engineers India turns out becoming millions of open source Linux programmers. But if MS can ensure that they will all have seen the Windows sources at some point, then they'll never again be able to contribute code to any major project, lest MS get all litigious about the possibility of misappropriated code. Might not win in the courts, but raise your hand if you'd like to see a federal judge slap a preliminary injunction on any distribution of the Linux kernel until the mess is sorted out!
True? Nah, likely not. Would it work? Just possibly. We've all heard about Samba developers who treat MS code like a toddler running around with ebola milkshakes (cover eyes and run).
No, not mistaken. In fact, you're pretty much spot on. I'm sure Dr. Foster appreciated the fact that you were paying attention instead of fantasizing about Beowulf clusters in the sky. He really isn't driven by the vast (potential) computational powers of the Grid, but by what it enables researchers to accomplish.
One of these, as you said, is management of distributed resources. Importantly, we want such management to be transparent to the end user. Thus the terminology "Grid computing" -- the system should be like the power grid, except that you plug in your task and get the computational resources that you need, instead of electricity. We'd really like it if you could just submit a request for a given file (or other object representing a desired data product) and have the Grid spit it back at you, without your having to have any knowledge about what resources were needed or what secondary computations were run in order to satisfy your job's dependencies. Of course, the whole system is based on open technologies, and there is lots of monitoring technology built in, but that's more for the benefit of the software and developers than the end users.
The other great potential benefit of Grid computing is for collaboration. Several researchers can easily examine the same data by requesting the same data product from the Grid, and can all run the same processor-intensive visualization tools using Grid-supplied resources. If things are arranged correctly, the visualization itself never needs to be run more than once. One project I'm involved with, by way of example, is producing a sky browser that can, on demand, dispatch Grid jobs to crunch through Sloan survey data and identify the galaxy clusters in the field you're looking at. Cosmologists think this is really neat.
By and large, new science codes being deployed on the Grid are written in portable things like Java, Python, TCL(!), and so on, precisely because you don't need to care that much what the underlying system looks like. There are a lot of old binaries written in FORTRAN and worse, especially in the particle physics community, that make no end of grief for people setting up production pipelines on the Grid. This is made a little easier by the fact that Linux-based clusters are popping up everywhere these days, so often getting something to run on Linux (or even as specific as i386 Linux) will get you far enough.
Anybody who wants, can play. And plenty of people who didn't necessarily want, also wind up playing.:-) If you have special talents, cool tools, or unusual friends, there's a good chance that you will be found and recruited.
If you live in a dorm, your choice of team will probably be quite clear; look for signs going up in late spring as to meeting places for planning powwows and the like.
If you want to have more fun, and avoid the regimentation of the big dorms, check out the Federation of Independent Scavhunt Teams (F.I.S.T. for really short, our full name is currently running towards a paragraph.) Our philosophy is, let's have major fun, as cheaply as possible, and devote ourselves religiously to what we consider one of the greatest games ever invented for four days. We're looking to start recruiting new talent sometime shortly after winter break, so keep your eyes open.
We're also probably the most geographically diverse team: all over campus, of course, but we've even got some hardcore players who drive in from other states.
wanting to build? Fred & Justin had a lab on the 3rd floor of Kirsten; they used to spend nights in there drinking, smoking cigarettes, and building low-budget lasers, plasma cannons, and other implements of destruction.
And oddly enough, the physics dept. basically sealed off the 3rd floor "student lab" after they left. By the time I managed to get back in there, it had been stripped bare and was going to be used for teaching space. Of course, one of them apparently left bits of their stuff as "presents" hidden around the Research Institute. Took me two years, but I eventually found the guts of the pulse forming network (I think) stashed in the sub-basement of the Accelerator building next to some discarded-crated-and-encased-in-fiberglass NASA hardware.
Never did find much of great use, though. On the other hand, claiming to have some leftover Fred TechTM on hand is still a good way to scare a few points out of the judges. A schematic for the plasma cannon was all it took to get partial credit for the "Deface the surface of the moon" item a couple of years ago.
This story comes up every so often, and is met with the same incredulity. I was there, on the team that built the reactor, so take it from me when I say that we did it. (Well, not so much we, as Fred and Justin did it with their mighty ninja atomic physicist powers; I was a first-year at the time, so my major contribution there was listening to them explain the scheme at breakfast.)
The fact is, a breeder reactor is just anything that is making plutonium, at least as far as the judges were concerned. So they made plutonium, by irradiating thorium from lantern mantles with a source they "borrowed" from the student labs. The tricky part was convincing the physics department to lend them a $20K proportional counter so they could detect the relaxation photons and thus prove plutonium production. After 36 hours of running they had a few hundred events that we figured corresponded to a total yield of 100K atoms or so.
Yes, purification would have been harder. No, we're not actually sure what eventually happened to the reactor.
It's a negligable amount. About $500 for first place, which is at most 1/5th of the teams funds spent on the competition.
Not so, my friend! Dare you slander the mighty F.I.S.T.* so? We took fourth place, which came with an (I think, payout still pending AFAIK) $150 prize. Whatever it was, we got together after the film screening and decided that the prize money came out to about half of our total budget. Goes to show what real dumpster-diving legwork can get you.
*F.I.S.T. = The Lush Puppies Mk. II: Federation of Independent Scavhunt Teams. Basically we decided that the little five-person teams that never stood a chance against the big dorms were being oppressed, so rather than be Borged by the big guys, we banded together to form a team of our own. I think our best moment (sadly not caught on the film) was after judging and cleanup finished, some of the FIST headed out behind Ida Noyes to dumpster dive the other teams' scraps for use next year.
As you say, Turing proved that the halting problem is undecidable; to be more precise, there is no general algorithm to decide whether a program will halt, that will provide an answer in finite time for any program.
That's okay. Detecting whether a DVD menu program is trying to die, the vast majority of the time, only involves special cases about which the halting theorem is silent. For instance, if the program halts by exuting the HALT machine instruction, you simply search for that instruction and replace it with your own code. If you don't trust the program not to modify its own code, you can run it in an emulator that constantly checks whether it is about to execute a HALT instruction. Consider that this is roughly what any modern CPU does in order to enforce memory protection.
Detecting infinite loops is a bit harder, and while you can search for trivial LOOP: GOTO LOOP constructs, you really need to run the program in emulation to get anywhere. You can, for instance, take a snapshot of the code and virtual machine state from time to time; if both ever return to the exact same state, you must be in an infinite loop (assuming you don't have interrupt-driven code to worry about -- if you do, expand the code under analysis to include the interrupt handlers, and repeat).
Moral of the story: not allowing the menu to die doesn't require knowing whether the menu program will ever halt, only whether it is about to. And that, far from being a theoretical impossibility, is a fundamental technique of virtual machine design.
The number of atoms won't change if you change c, but the distance between them (which is determined by the interaction between charges, which is mediated by photons, which move at c) will--it will change in exactly the way to make you think that c hasn't changed.
As philosophically nice as that would be, this just isn't true. The spacing of atoms and molecules in bulk matter is mediated by electromagnetic forces, but is determined by the solution to the local Schrodinger equation in a periodic potential, and it is not at all straightforward to evaluate this spacing given only fundamental constants and the composition of the material. In particular, the spacing in question definitely does not vary linearly with c.
To take an example from a previous comment, suppose you wanted to express c in terms of the radius of a U-238 nucleus and a Lymann-alpha photon period. Using only first-order effects, the energy of the Ly-alpha photon varies as 1/c. Since h is experimentally determined, we'll suppose it doesn't change (you'd notice immediately if it did), in which case the Ly-alpha period is linear in c.
For the nuclear radius the matter is more complex. For the droplet model it becomes a matter of QCD to determine the effective nucleon radius. If we use the less realistic but easier point-particle non-local-potential model, it depends primarily on the mass of the pion: the volume depends (to rough first order) on c^2, so the radius goes as c^(2/3).
Finally, the measured value of c goes as length over time, so this measurement varies as c^(2/3-1) = c^(-1/3). So you really can't claim that every possible measurement of the value of c is invariant under changes in c, because this one clearly isn't.
Umm... sorta. The feather and the bowling ball experience identical forces, and thus acquire the same velocity in a given time. However, as you say, we assume M_2 = 0 for the *other* force equation, the gravitational force of the feather/bowling ball acting on the earth.
So if we put the M_2 term back in, we find that the bowling ball and the earth meet slightly sooner than for the feather, because the earth is more strongly attracted to the bowling ball. But obviously, this is a *very* small effect.
At least, I think that's what you were getting at. I'm not sure what the moon had to do with it.
Done carefully, there is no other explanation for the two-slit experiment. If you send electrons one at a time at a double slit, you will still see an interference pattern, because the electron is interfering with itself. Place an electron detector in the path before the slits, and you instead see two spots where the electrons land. This is a pretty standard seocnd- or third-year physics lab experiment.
You can do this with more massive objects, too. From the point of view of an alpha particle, a heavy nucleus is a large, opaque disk. Firing alphas one by one at a thin target of heavy nuclei also gets you a diffraction pattern. Why? Because the alpha is interfering with itself in going around the nucleus it interacted with -- in effect, it went around both sides of the nucleus.
In more recent news, it turns out that you can do this with Bose condensates, too. Which means you are sending an entire atom through multiple paths simultaneously.
Slashdot Mirror
For instance, you can do anything you like with Linux in the comfort of your own home, so long as you don't distribute the result. But distribution can become a thorny issue if you're careless, and as a result the GPL offers relatively weak protection of your company's super-secret algorithms. After all, anyone who *legally* acquires part of the code, now has full GPL rights to the whole thing, tasty bits included.
Of course, this is only very slightly less protection than, say, trade secret law offers. In either case, the rule of thumb is to keep your secrets secret.
Companies do occasionally get stung by this when they fail to realize that they are actually "distributing" code. Witness Linksys, which was forced to give away its changes to the Linux kernel, probably because it simply never occurred to the developers that selling a device constituted "distribution" of the image in the firmware. The poster's company (or its customers) could conceivably run into the same problem down the line by selling POS units with the customized GPLed drivers installed.
Compare to TiVo, which had the sense to put the interesting code in separate executables that don't derive from anything GPLed.
Well, okay, if you insist.Now associate
Which is, of course, why projects like this never go anywhere. Because if you're going to require that every piece of software be repackaged anyway, to conform to your standard, there's no additional burden in packing them all in the same format. In which case, you've got the Debian project all over again, except that Debian has a ten-year head start and a lot more package maintainers in place.
Sure, there's scads of obscure software lying around out there in TGZs. But the fact that it's still a TGZ would seem to indicate that nobody has a burning interest in repackaging it.
Well, not quite. Remember that the vast majority of military spending goes to mundane things like salaries, fuel, food, and the like. This doesn't affect the economy any differently than anyone else spending money on those things, and it's actually worse for the economy if it's the government doing the spending, since the tax rate effectively reduces the mean return on investment, which discourages business spending.
However, money spend on research-intensive activity is quite another matter; many people would say that the NSF's budget is some of the best-spent money in the world. This is because research spending not only funds jobs (and with a larger multiplier effect, since most of the money is spent on stuff, which takes jobs to produce), but directly drives investment in technology, as well as potentially creating new technology.
Therefore, if you want to gauge the effect of a new space program, you'd be better off comparing it to the military's R&D budget. Which is, unfortunately, a little harder to actually pin down. Definitely larger than NASA's budget, though.
Have you looked into a folding bicycle? There are many models out there, most of which are designed with the specific goal of getting down to a size where they will go on a train. A bit pricier than a regular bicycle, but all the same advantages in terms of operating cost.
Google will get you started nicely.
Not so much encrypted, though. Reading the mail your team was delivering became a very funny pastime. I did briefly consider printing out an ascii-armored GPG message, but those things are a bitch to type in.
How sweet!
I don't suppose you have a link handy for that simulation? I'm curious to know what you expect the decay products of a black hole event to actually be.
Oh, great. The sci.physics loons have discovered /.
On the other hand, sci.physics is a lot funnier than the trolls that hang out here, so it's probably an improvement.
Furthermore, I think the scenario in which this happens is pretty far fetched. More realistic models (which can avoid proton decay, for instance) usually have to put the scale at which black holes form much higher (100 TeV or higher).
I'm familiar with the argument that you need large extra dimensions to be able to produce gravitational objects in the low TeV domain, but I've never heard proton decay cited as a constraint on this before. I'd be really interested in seeing a reference to this. Can you post one? (To be fair, most every theory these days seems to have something to say about proton stability, eventually.)
If black holes exist sans Hawking Radiation, we'd be in quite a bit of trouble upon the production of even the smallest one. Probably wise to check that little problem out. I'm not advising doing anything wacky and superparanoid, like building it on the Moon
... about a parsec. I'm not worried.)
Not really a problem, even without Hawking decay to destroy the little guys. Laboratory black holes would be substantially smaller than a proton, electrically neutral, and thus unable to interact with other matter at all, except via gravity. Moreover, they would almost certainly be moving at relativistic velocities upon production. Or certainly well above escape velocity, at any rate.
V_esc ~ c/100000 so for a singularity with a mass of "hundreds of protons", the excess energy needed to reach escape velocity is on the order of a few GeV. At LHC, that's muon noise.
Therefore, I would expect a non-decaying black hole generated in a collider to simply escape into space harmlessly. With an interaction cross section probably comparable to that of a neutrino, it could even fly right through the Earth and come out the other side, probably without even disturbing many atoms in its path.
(Recall that a flux of a few trillion neutrinos per square meter passes right through the Earth each second. You can calculate that the stopping distance in lead for an average solar neutrino is
Another reason not to worry: if black holes don't decay via the Hawking process, then the universe should be chock-full of primordial black holes with masses ranging from atomic to small planet. None of them seems to have swallowed the Earth yet (or, even more conviningly, the Sun).
There is a reason the pdf uses bit-mapped fonts, and we are it. Our SavHunt team (the Lush Puppies Mark III, FIST Deux, Deleuzian Potato) uses a computer database with a web interface to keep track of our itemss. So, the judges used bit-mapped fonts this year to force us to type in the list by had. Bastards!
(Aside: Anyone know of a method that would allow for a 'any n of m keyholders needed to decrypt' schema? It's something that has advantages, but I've no idea how to go about it)
See Applied Cryptography, chapter 3, "Secret Splitting", for an introduction. I found a toy implementation and mathematical explanation here. Or just Google for "secret splitting".
The evolution of languages differs from the evolution of species because branches can converge. The Fortran branch, for example, seems to be merging with the descendants of Algol. In theory this is possible for species too, but it's so unlikely that it has probably never happened.
No kidding. The evolution of languages is probably different from the evolution of species because, well, they're languages. Paul Graham seems to have overlooked the fact that there's already a fully-formed field of study devoted to things that evolve like languages. It's called linguistics. He's probably heard of it.
When I was a kid, I had a grand old time flipping through the back pages of the unabridged American Heritage dictionary. One item of interest in there is a (then current) language tree illustrating the radiation, convergence, and development of linguistic families derived from Indo-European. The dictionary's editors assured their readers that, while informative, the printed tree was a gross simplification of the real state of affairs.
My point being, I am inclined to suspect that a field that can examine modern languages and scraps of historical writing and extrapolate back to reconstruct languages dead for thousands of years, has the tools to make sense of fifty years of computer languages and get a sense of the general trends going forwards.
However, you should be doing all your homework without being forced! School is already too easy, and if you skip any of it you'll be the only one at McDonald's who can't make change! You should be asking your teachers for extra homework!
Extra homework, eh? Have you actually looked at homework recently? While there are exceptional schools and exceptional teachers, from what I've seen, most of the work being handed out these days strikes me as designed to encourage a slavish devotion to work for work's sake. Mostly it's the same busywork as ever, but there's a hell of a lot more of it then when you were in school. Which means that the occasionally interesting assignments are basically history, since no teacher has the time to grade responses more complex than numeric or multiple-choice now.
So frankly, I think it would be time better spent for anyone still a guest of the mandatory pre-college babysit^H^H^H^H^H^H^H education system to do just enough homework to get by and keep the grades up, and pursue projects of actual interest independently. Or even better, find a good mentor. Heaven knows I didn't get into college on the strength of my grades. Or grad school, for that matter.
Hi there. Nice work on USITE Crerar, then. I've been in there a number of times since the lab was installed. Impressive Seti numbers, too.
Since you ask, I work for Dr. Foster under the rubrick of the Computation Institute on campus. Where on campus that is currently is in a state of flux between Ryerson and the RI.
I think that the USITE computers will continue running Seti for the time being, as we are not suffering from a lack of heterogenous resources, but a lack of software to pull it all together. We've currently got I think twelve grid sites lined up to accept Sloan cluster finding work, not counting the WorldGrid, which we should (hopefully) be able to tie into as well before long. iVDGL might be of interest to you on that front. My group, the Grid Physics Network, is less concerned with these management issues, than with taking the existing tools and wrapping them around real physics problems.
On the other hand, I do know people who wouldn't mind if those lab computers got set up with some Condor daemons, which would be a lot more useful than some single-purpose cluster finding code, anyway, and more in keeping with the grid computing style of doing things. I don't know how much demand there is for WinNT pools, though.
And here you have the conspiracy-theorist motive for such an action: MS isn't afraid of India switching to Linux, but of the millions of engineers India turns out becoming millions of open source Linux programmers. But if MS can ensure that they will all have seen the Windows sources at some point, then they'll never again be able to contribute code to any major project, lest MS get all litigious about the possibility of misappropriated code. Might not win in the courts, but raise your hand if you'd like to see a federal judge slap a preliminary injunction on any distribution of the Linux kernel until the mess is sorted out!
True? Nah, likely not. Would it work? Just possibly. We've all heard about Samba developers who treat MS code like a toddler running around with ebola milkshakes (cover eyes and run).
No, not mistaken. In fact, you're pretty much spot on. I'm sure Dr. Foster appreciated the fact that you were paying attention instead of fantasizing about Beowulf clusters in the sky. He really isn't driven by the vast (potential) computational powers of the Grid, but by what it enables researchers to accomplish.
One of these, as you said, is management of distributed resources. Importantly, we want such management to be transparent to the end user. Thus the terminology "Grid computing" -- the system should be like the power grid, except that you plug in your task and get the computational resources that you need, instead of electricity. We'd really like it if you could just submit a request for a given file (or other object representing a desired data product) and have the Grid spit it back at you, without your having to have any knowledge about what resources were needed or what secondary computations were run in order to satisfy your job's dependencies. Of course, the whole system is based on open technologies, and there is lots of monitoring technology built in, but that's more for the benefit of the software and developers than the end users.
The other great potential benefit of Grid computing is for collaboration. Several researchers can easily examine the same data by requesting the same data product from the Grid, and can all run the same processor-intensive visualization tools using Grid-supplied resources. If things are arranged correctly, the visualization itself never needs to be run more than once. One project I'm involved with, by way of example, is producing a sky browser that can, on demand, dispatch Grid jobs to crunch through Sloan survey data and identify the galaxy clusters in the field you're looking at. Cosmologists think this is really neat.
By and large, new science codes being deployed on the Grid are written in portable things like Java, Python, TCL(!), and so on, precisely because you don't need to care that much what the underlying system looks like. There are a lot of old binaries written in FORTRAN and worse, especially in the particle physics community, that make no end of grief for people setting up production pipelines on the Grid. This is made a little easier by the fact that Linux-based clusters are popping up everywhere these days, so often getting something to run on Linux (or even as specific as i386 Linux) will get you far enough.
Anybody who wants, can play. And plenty of people who didn't necessarily want, also wind up playing. :-) If you have special talents, cool tools, or unusual friends, there's a good chance that you will be found and recruited.
If you live in a dorm, your choice of team will probably be quite clear; look for signs going up in late spring as to meeting places for planning powwows and the like.
If you want to have more fun, and avoid the regimentation of the big dorms, check out the Federation of Independent Scavhunt Teams (F.I.S.T. for really short, our full name is currently running towards a paragraph.) Our philosophy is, let's have major fun, as cheaply as possible, and devote ourselves religiously to what we consider one of the greatest games ever invented for four days. We're looking to start recruiting new talent sometime shortly after winter break, so keep your eyes open.
We're also probably the most geographically diverse team: all over campus, of course, but we've even got some hardcore players who drive in from other states.
And oddly enough, the physics dept. basically sealed off the 3rd floor "student lab" after they left. By the time I managed to get back in there, it had been stripped bare and was going to be used for teaching space. Of course, one of them apparently left bits of their stuff as "presents" hidden around the Research Institute. Took me two years, but I eventually found the guts of the pulse forming network (I think) stashed in the sub-basement of the Accelerator building next to some discarded-crated-and-encased-in-fiberglass NASA hardware.
Never did find much of great use, though. On the other hand, claiming to have some leftover Fred TechTM on hand is still a good way to scare a few points out of the judges. A schematic for the plasma cannon was all it took to get partial credit for the "Deface the surface of the moon" item a couple of years ago.
This story comes up every so often, and is met with the same incredulity. I was there, on the team that built the reactor, so take it from me when I say that we did it. (Well, not so much we, as Fred and Justin did it with their mighty ninja atomic physicist powers; I was a first-year at the time, so my major contribution there was listening to them explain the scheme at breakfast.)
The fact is, a breeder reactor is just anything that is making plutonium, at least as far as the judges were concerned. So they made plutonium, by irradiating thorium from lantern mantles with a source they "borrowed" from the student labs. The tricky part was convincing the physics department to lend them a $20K proportional counter so they could detect the relaxation photons and thus prove plutonium production. After 36 hours of running they had a few hundred events that we figured corresponded to a total yield of 100K atoms or so.
Yes, purification would have been harder. No, we're not actually sure what eventually happened to the reactor.
Not so, my friend! Dare you slander the mighty F.I.S.T.* so? We took fourth place, which came with an (I think, payout still pending AFAIK) $150 prize. Whatever it was, we got together after the film screening and decided that the prize money came out to about half of our total budget. Goes to show what real dumpster-diving legwork can get you.
*F.I.S.T. = The Lush Puppies Mk. II: Federation of Independent Scavhunt Teams. Basically we decided that the little five-person teams that never stood a chance against the big dorms were being oppressed, so rather than be Borged by the big guys, we banded together to form a team of our own. I think our best moment (sadly not caught on the film) was after judging and cleanup finished, some of the FIST headed out behind Ida Noyes to dumpster dive the other teams' scraps for use next year.
As you say, Turing proved that the halting problem is undecidable; to be more precise, there is no general algorithm to decide whether a program will halt, that will provide an answer in finite time for any program.
That's okay. Detecting whether a DVD menu program is trying to die, the vast majority of the time, only involves special cases about which the halting theorem is silent. For instance, if the program halts by exuting the HALT machine instruction, you simply search for that instruction and replace it with your own code. If you don't trust the program not to modify its own code, you can run it in an emulator that constantly checks whether it is about to execute a HALT instruction. Consider that this is roughly what any modern CPU does in order to enforce memory protection.
Detecting infinite loops is a bit harder, and while you can search for trivial LOOP: GOTO LOOP constructs, you really need to run the program in emulation to get anywhere. You can, for instance, take a snapshot of the code and virtual machine state from time to time; if both ever return to the exact same state, you must be in an infinite loop (assuming you don't have interrupt-driven code to worry about -- if you do, expand the code under analysis to include the interrupt handlers, and repeat).
Moral of the story: not allowing the menu to die doesn't require knowing whether the menu program will ever halt, only whether it is about to. And that, far from being a theoretical impossibility, is a fundamental technique of virtual machine design.
As philosophically nice as that would be, this just isn't true. The spacing of atoms and molecules in bulk matter is mediated by electromagnetic forces, but is determined by the solution to the local Schrodinger equation in a periodic potential, and it is not at all straightforward to evaluate this spacing given only fundamental constants and the composition of the material. In particular, the spacing in question definitely does not vary linearly with c.
To take an example from a previous comment, suppose you wanted to express c in terms of the radius of a U-238 nucleus and a Lymann-alpha photon period. Using only first-order effects, the energy of the Ly-alpha photon varies as 1/c. Since h is experimentally determined, we'll suppose it doesn't change (you'd notice immediately if it did), in which case the Ly-alpha period is linear in c.
For the nuclear radius the matter is more complex. For the droplet model it becomes a matter of QCD to determine the effective nucleon radius. If we use the less realistic but easier point-particle non-local-potential model, it depends primarily on the mass of the pion: the volume depends (to rough first order) on c^2, so the radius goes as c^(2/3).
Finally, the measured value of c goes as length over time, so this measurement varies as c^(2/3-1) = c^(-1/3). So you really can't claim that every possible measurement of the value of c is invariant under changes in c, because this one clearly isn't.
Umm ... sorta. The feather and the bowling ball experience identical forces, and thus acquire the same velocity in a given time. However, as you say, we assume M_2 = 0 for the *other* force equation, the gravitational force of the feather/bowling ball acting on the earth.
So if we put the M_2 term back in, we find that the bowling ball and the earth meet slightly sooner than for the feather, because the earth is more strongly attracted to the bowling ball. But obviously, this is a *very* small effect.
At least, I think that's what you were getting at. I'm not sure what the moon had to do with it.
Done carefully, there is no other explanation for the two-slit experiment. If you send electrons one at a time at a double slit, you will still see an interference pattern, because the electron is interfering with itself. Place an electron detector in the path before the slits, and you instead see two spots where the electrons land. This is a pretty standard seocnd- or third-year physics lab experiment.
You can do this with more massive objects, too. From the point of view of an alpha particle, a heavy nucleus is a large, opaque disk. Firing alphas one by one at a thin target of heavy nuclei also gets you a diffraction pattern. Why? Because the alpha is interfering with itself in going around the nucleus it interacted with -- in effect, it went around both sides of the nucleus.
In more recent news, it turns out that you can do this with Bose condensates, too. Which means you are sending an entire atom through multiple paths simultaneously.