I believe you have misquoted our leader. The quote goes:
"There's an old saying in Tennessee -- I know it's in Texas, probably in Tennessee -- that says, fool me once, shame on - shame on you. Fool me -- you can't get fooled again." Pres. G. W. Bush Nashville, Tenn., Sept. 17, 2002.
It should be noted that the Journal of Applied Physics is regarded by much of the physics community as one of the crappiest journal extant in terms of standards for publication. I would honestly be surprised if, during the last calendar year, even one article showed up in JAP that wasn't first rejected by one (or a host) of other journals.
If there were articles on the DOE's site that are not available anywhere else, couldn't the DOE ditch the rest and keep just those available. At the very least, it would irritate the money-grubbing assholes who wanted the entire site shutdown.
The short answer: Yes. The DOE should have only removed the offending articles and left the rest of the database intact.
I have written many articles that were in the OSTI database. As a matter of fact, at my organization publication entails, as a matter of institutional and DOE policy, sending a draft copy of each article or report to the library for OSTI dissemination. Frankly, I'd always wondered about the legality of this, given that in order to publish the article in scientific journals I need to sign a copyright transfer form that assigns all copyright rights to the publisher.
Upon transferral, the publishing company owns all rights to the work in question and to any "revised or expanded derivative works based on the Work." (I'm quoting from the IEEE copyright form in front of me). In my reading, the reports I send to the OSTI database would constitute a derivative work based on the Work submitted for publication. The publisher's request that the DOE not disseminate this material for free is, I'm afraid to say, their legal prerogative, however much you or I may disagree with the principle of their doing so. However, I would argue that no single publisher has the right to shut down the entire database; they only may exercise rights over their own copyrighted works and not over any other work in the database.
If I were running the DOE, I would have played hardball with the publisher: I would have requested the publisher to list, article by article, specifically those works that they own the copyrights to and then excise only those articles from the database (after they've demonstrated their ownership of the copyrights) while leaving the rest intact. Then, DOE policy would be modified to require all DOE-supported research to be published in only those journals that authorize OSTI dissemination of the work. Moreover, "public service" (i.e., editorial service or referee duties) for "for-profit" publishing companies (defined as those companies which do not permit OSTI dissemination) would not be permitted by employees who bill their time to DOE grants and contracts.
Such a boycott would honor both the legal copyrights of the scientific work as well as the notion of public access to taxpayer-funded research. In time, as other scientific organizations (NSF, NIH, etc.) follow suit, publishers who choose to exercise their copyrights in this fashion can expect to become progressively more marginalized by the community at large. The government has already, in most cases, footed 99.9% of the cost of producing a scientific work. Retaining the right to list the work in the OSTI database should not be too much to ask.
The founders also wished to restrict voting rights to white land-owning male citizens, so I wouldn't be too quick to adopt something merely because of their stamp of approval.
The 12th Amendment was ratified shortly after the "Crisis of 1800." In 1800, the U.S. government came precariously close to falling apart when the presidential/vice presidential balloting produced a tie. Jefferson and Burr tied in the 1800 presidential election with 73 votes apiece in the electoral college. The vote moved to the House to resolve the tie. 9 states were needed to win, but in ballot after ballot, Jefferson got 8 votes, Burr, 6. The succession procedure defined by the U.S. Constitution appeared to be failing, and one can only speculate on what chaos would have ensued had the Constitutional crisis not been resolved by some careful maneuvering of Jefferson's chief rival, Alexander Hamilton. Only after Hamilton persuaded key Federalists to vote for Jefferson were the President and Vice President selected. Hamilton was thanked for his service to the country--he was shot dead in a duel four years later by, you guessed it, Aaron Burr.
Bad design in the succession procedure almost destroyed the United States once, and the People, reconizing inferiorities in the process, were moved to change the way the system worked. A quick perusal of the list of amendments to the Constitution shows that changes in the election process have been made many times when what was in the highest law of the land was understood to be inferior or obsolete. One wonders if it may be possible for this to happen again in light of advances in voting theory. (I'm doubtful of the practicality of this, however. As one can tell by a quick read of the/. posts, most U.S. citizens lack the mathematical sophistication and the objectiveness to recognize deficiencies in the process and to evaluate alternatives).
Multiple postdocs weren't necessarily the kiss of death in physics when I went through the process not that many years back; I did two postdocs and had the luxury of turning down a couple of faculty positions before taking a staff position at a national lab. Either things have changed somewhat, or else much must depend on the field and the circumstances.
Alas, you never never never (never!) want to get passed over for tenure. If ever there were a genuine kiss of academic death, this would be it--no self-respecting university would opt to wear Princeton's or Stanford's hand-me-downs when they could just advertise in Physics Today and get 50+ highly qualified applicants who don't carry this baggage.
Personally, I just don't see what the allure is of academic life over, say, being a staff scientist at a national lab. I'd much rather just do research full-time and advise the occasional grad student and postdoc than put up with 300+ student intro courses, faculty committees from hell, petty university politics, and the pressure of bringing in enough research cash to satisfy a tenure committee. Your research ends up being very narrowly focussed while you try to carve a niche for yourself; this is the antithesis of creative scientific enterprise, IMO, and a great way of burning out during your most creative and productive years. All this while wearing the obligatory "I'm an assistant professor--kick me in the teeth" smile around the more senior faculty who control your fate.
I second the recommendation of "A Ph.D. is Not Enough!." It is an outstanding read and valuable information for those masochists who wish to try the academic route.
Who are you to decide what can and can't be compared? A language is a language, and I can compare and contrast the features of languages as I please. Note that I talked about FORTRAN 77 in my post, not F90 or F95. It was a fair comparison - F77 is deficient when compared to a modern language such as C++.
Read my comment on comparing modern automobiles with the Model A. Fine by me if you want to make the comparison--you just look a bit silly is all. Shall we dismiss C++ out of hand because pre-ANSI/ISO-Standard C++ had no convergence on template functionality? (In practice, compiler vendors still often don't implement the language standard fully, but I digress...).
These issues have been so well debated that I'm not even going to go into them here. You're wrong. Time and experience have shown this to be true. The "cute" features that bother you are important to writing large, maintainable projects. Can they be misused? Yes. Sure. But to programmers skilled with the language, they provide valuable tools that FORTRAN just doesn't offer.
Two obvious comments: First, language selection is entirely subordinate to design and planning of a project. A skilled programmer with an understanding of modern data structures and algorithms can, with very few exceptions, implement any of them in any sufficiently powerful language. You don't need C++ to do linked lists or hash tables or to have OO structure, or even to design large maintainable software projects. I have seen many large C++ projects fail because of poor design, and I have seen many large F90/F95 projects succeed because of superior design. The design decision on what language to use for a given project should be based solely on technical criteria and not kneejerk prejudices.
Second comment: Those "skilled programmers" you refer to rarely seem to end up writing the code I have to read and maintain. I myself don't work in a software house--I am a scientist and I work with scientists, people who, unlike me, often don't have formal computer science training or any compelling inclination to learn to write beautiful, well-structured C++. We're not talking about people who have read, understood, and know how not to abuse the information in Design Patterns--we're talking people with PhDs in physics or engineering who have taken a 1-week "How to get stuff to compile with a C++ compiler" syntax course who suddenly feel qualified to code up the world only to end up, more often than not, reinventing (badly) features already in the STL. They have only a nebulous understanding of the subtle aspects of C++, and they tend to cling to "natural-looking" constructs and idioms that just end up bogging down their code factors of two to ten compared to equivalent Fortran. These people are much less of a hassle if you let them program in Fortran.
I maintain, and I have considerable experience to back this claim, that because C++ is so large a language, to have any measure of refactorability you have to have programmers who have a solid understanding of the language as well as significant training in software design. Otherwise you get an unmaintainable mess, and the power and flexibility of C++ is overkill. Even if you do have such a team of programmers, it is not clear that C++ is always the language of choice: if you are not leveraging old C++ class libraries, you anticipate no need for refactorability of the current class libraries, if you have specific performance objectives (e.g., heavy array manipulation, which other languages handle better) or design objectives (a deadline that cannot tolerate the longer design cycle of designing/testing C++ class libraries), then C++ is simply not the best choice.
As for introsort, I concur with your technical assessment of the algorithm. My only reason for broaching the subject was to comment that Press et al. had presented a set of technical reasons for preferring heapsort in many settings (not the least of which is that the algorithm is dead easy to understand) and not some misguided attempt to crown an "ultimate sorting routine" as your original caricature suggested.
"Don't drive a Ford. Fords suck because my 2003 Caddy has way more features than a Model A." Let's compare apples and apples, shall we? It is unfair to compare Fortran 77 with modern languages such as C++. Either you compare Fortran 77 on its own merits with C++ circa 1977 (i.e. nothing) or else you should consider the most modern incarnation of Fortran (Fortran 95) with C++.
Fortran 95 has pointers; I have implemented essentially all "modern" data structures--hashes, heaps, tress, linked lists--with it. You can program object oriented F90/F95 (with one caveat--see "*" below). If you are averse to the ugly language features of F77 and its predecessors and their inclusion in modern Fortran for backwards compatability reasons, then you are welcome and encouraged to program in a beautified and simplified subset of the language such as the language "F" and forget about Holleriths and EQUIVALENCEs and the like.
As has been commented by others, Fortran is very efficient and relatively simple to optimize for certain kinds of tasks seen frequently by those in the science and engineering fields. Fortran 95 has nice array syntax that is dead easy for compilers to detect and optimize. Fortran also allows for a fairly rapid development cycle--the language doesn't have as much capacity for "being cute." You may hate the simplemindedness, but in the real world people have to read, understand, and maintain other people's code. All those cute syntactic tricks one feels obliged to play in C or C++ just add to the time it takes for someone to read over your code and figure out what the heck it is supposed to be doing. In my experience other peoples' Fortran 90 is so much easier to human-parse than other peoples' C or C++ that that alone would be enough to justify the language on any kind of nontrivial project. (That, and you don't have to piss away your time worrying about whether a method should be virtual or pure virtual, but I digress...).
As for your slam of heapsort in Numerical Recipes, in my recollection Press et al. identified many specific technical reasons for suggesting heapsort for many numerical applications over quicksort: it is an "in-place" sort, it has the same computational complexity as quicksort, the worst case scenario is only about 20-30% slower than the best case scenario, and, yes, its implementation in some languages is more transparent. (People who use numerical algorithms that they don't take time to understand should be forced to take the maiden ride in the airplanes or drive over the bridges they design). In many numerical applications one needs to sort large "nearly sorted" datasets; naked quicksort without some kind of shuffling dies a painful N-squared death on these cases.
(*) To be fair, I should comment that "object oriented" F95 doesn't support runtime polymorphism, but I've yet to encounter a numerical analysis task that would benefit in any significant way from this language feature.
There is some decent new stuff, though you have to look a bit if you want quality "hard" science fiction rather than space opera fluff. Stephen Baxter (Moonseed, Titan, Manifold Time), for example, writes some fairly decent classic science fiction. Greg Bear (Darwin's Radio among others) has his moments. And a new guy, Ken Wharton (Divine Intervention, has injected some fresh, if somewhat provocative, ideas into the genre. (Ken was a friend of mine from grad school, so I can't help but plug his book).
That said, of course not every application is going to achieve that level of performance.
I agree. It should be noted that 10^6 by 10^6 matricies are rather puny, really, considering the scope of the problem they envision running on the machine. I would think that to do high fidelity hydro effectively on the machine they'd need near-peak performance on 10^9 by 10^9 matricies at the smallest. (Even this isn't that big, really--it would correspond to a box 1000 cells to a side for a non-AMR hydro calculation). It will be interesting to see in the years to come what kind of "real world" performance they get on their climate models.
It's actually more restrictive if you work for the government--playing computer games on government property, even "after hours," falls under the rubric "Waste, Fraud, and Abuse," and could potentially (though it seldom does) lead to criminal penalties as well as administrative punishment. On the bright side, there are rewards for turning in those who defraud the government. You have a coworker who likes to play solitaire during his lunch break? Kaching!
Censorship of free speech has no purpose at any time.
This is a bit naive. Political canvassing outside polling places, for example, is censored speech in the U.S., an abridgement of civil liberties for the distinct purpose of supporting democracy through fair elections. As another example, one who has access to national security information (take nuclear weapons information, for example) cannot share this information without going to jail. This has the distinct purpose (we may quibble about the validity of the purpose, but it is a purpose nonetheless) of preserving U.S. sovereignty and the safety of its citizens and allies. However, it entails an abridgement of civil liberties for those "in the know." One cannot stand up in synagogue and scream "Death to kikes!" without some expectation of legal recourse.
Freedom of speech is a lofty goal, and indeed it is treated with more reverence in the U.S. than in most other countries, but it is not and cannot be considered an absolute. Even now, with the "War on Terror," this freedom is being curtailed in the name of "national security" in ways we may well have considered impossible just one year ago. China, for reasons of its own national security, has even more severe restrictions on speech.
Personally, the most alarming aspect of this is that these companies were able to produce technology which satisfies, to a large degree, the rather tight-fisted control of information required by the PRC government. This bodes badly for those who hoped that the practical, technical difficulties of censorship would effectively block attempts to censor speech in the U.S.
My college roommate got a B.S. in Physics with minors in CS and math, and he went on to law school to study intellectual property/patent law. This has always struck me as a good way to put a physics degree to work without becoming a physicist and dealing with the requisite 5+ years of grad school getting a union card^]^]^]^]^]^]^]Ph.D.
Of course, with many subfields of physics having way more jobs than people to fill them, you might just opt to just stay in the more traditional physics tracks. Many of the national labs, e.g. LANL and LLNL, pay in the 90-100k+ range with nice benefits packages for entry-level Ph.D physicists, and they are aggressively trying to hire people. With the demographics of the labs as they are, they will likely continue being as aggressive in their hiring for another decade or so at least.
With one small typo you have juxtaposed the two great thinkers of modern times: Jon Katz and Kant. (Just between you and me, I would have ordered Kant before Katz, however. The former's acumen was, albeit slightly, more powerful than that of our esteemed/. colleague).
Re:Is Intelink More Secure Than Enigma?
on
Real Cyber-Spying
·
· Score: 1
Try using the "Hey, I'm a taxpayer, I PAID for this thing..." line with, say, Los Alamos National Lab weapons designers regarding plans and the detailed operation of the B61 or W76. Simple curiosity does not entail a "need to know." Nor does filing a 1040.
If you can get the information you want from the NSA, then more power to you. I'd suspect, however, that you won't get all the nitty gritty details you're after. (I could be wrong though--this is far abreast of either my expertise or whatever experience I may or may not have with classified computing systems).
Re:Is Intelink More Secure Than Enigma?
on
Real Cyber-Spying
·
· Score: 1
Yo, some Slashdot user who has access to this thing - tell us what administrative security is in force!
You're asking alot--I'd guess that you won't get a discussion in the open about details of admin security on Intelink. I'm no expert on the matter, but I'd suspect that this is classified national security information that would entail a loss of employment and/or jail time for the offender. Not to mention the fact that it would likely compromise the security that you are so concerned about in your post.
Remember back in the day when you could just walk into a store, drop a not-insignificant amount of money at the register, and then walk out of the store with an item you just bought? Seems quaint, no? Nowadays, after your purchase you get to stand in yet another line while a puke with an attitude and a pink magic marker signs his name ("X") on your receipt and "authorizes" your exit with your personal property. The entire legal concept of quid pro quo has been turned on its ear to accommodate these pink X's--we apparently no longer own the item when we exchange money for it, but rather the store can demand that you produce proof that your property didn't magically turn back into the store's property in the 10-foot walk from the register to the door. Remember the indignation we all used to feel at being treated like criminals just so a store Fry's Electronics could cut down on cash-register fraud? (Apparently, it's much less expensive to alienate customers than to just pay the employees enough to make them value their jobs). Remember how we all vowed never to shop in such a place anymore? Now this behavior is endemic--like the sheep we are, we accept it for that extra 5% off the purchase price.
Be sure you get rankled now. Five years from now, when the only place that'll sell you food is a urine-stained 7-11 in Compton because your face is a 92.4% match to a convicted felon in Joliet, you'll be forced to accept it. By then it'll be too late.
"If you don't do anything wrong, you have nothing to fear."...
I find this excerpt from the linked article to be interesting: Although ChoicePoint says it has records on nearly every American with a credit card, it doesn't always provide access to that data. The company's Autotrack service is popular with many agencies and businesses and is also used by reporters at The Wall Street Journal. But entering the name of FBI Director Louis Freeh into the Autotrack database produces an error message. A company spokesman says ChoicePoint intentionally blocks Mr. Freeh's records as an act of good corporate citizenship.
Translation of the last line: "A company spokesman says that the publicly held firm, ChoicePoint, is not so stupid as to endanger its stockholder's investments by providing information on the man heading one of ChoicePoint's biggest client organizations." Apparently this comment by the ChoicePoint drone is intended to make us all feel better, as if we all hobnob with politically heavyweights of Freeh's standing.
(Pardon my speculation as to the content of the article; at the present time it is inaccessible to me). Using nuclear weapons to guide asteroids and comets is an interesting idea, but the choice of using it for warmaking is odd. From a practical sense, the object would be far from "stealthy;" an intense X-ray source emmanating from a chunk of sky would almost certainly be detected, given our current capability, and it would tell us exactly where to look. We'd certainly wonder about that object coming towards us, especially when it changes course en route to Earth.
Any state with the technological savvy and nuclear arsenal to conduct such an activity would be able to dispose of its enemies in another fashion, and its enemies no doubt could dispose of the aggressor. (If not, then more conventional modes of attack could be used by the aggressor with greater precision, flexibility, and lower cost). Besides, the number of possible adversaries a target state would have that could conduct such a mission is very limited, so the target state would know full well who lofted a big snowball at them, and they would merely respond in kind prior to impact with whatever arsenal they had available. The only real use I can see for such a technology is to somehow coordinate it with a first-strike nuclear attack, with the big space rock knocking out hard targets such as underground command centers.
An alternative, peaceful use for such a technology would be to bring resources such as H2O to places we'd like to colonize. Slam a comet into the moon or Mars to bring water there, for example. Unlike trying to precisely control the descent of a chunk of ice onto Earth, a dicey game at best, one could instead direct the comet toward a different celestial body and have a much larger margin of error.
Re:On the subject of "cold fusion"...
on
Excess Heat
·
· Score: 2
The confinement scheme you are talking about (commonly referred to as an "electrostatic confinement" or a "Penning trap") is actually closer to hot fusion than cold fusion. The device works by making a spherical electrostatic potential that both confines ions and causes them to oscillate about the center of the device. If the electric field is sufficiently strong that the relative speeds of the ions is high, then collisions among the particles can, on occasion, lead to fusion reactions.
This link takes you to web pages describing (and showing nice color photos of) an ongoing Penning trap fusion experiment being conducted at the Los Alamos National Laboratory. The main drawback to devices of this type for power production is that one is limited in the number of ions that can be confined (the Brillouin limit) due to space-charge effects. Even without achieving breakeven, these devices are still very interesting, however, as enough fusion reactions take place in a small enough package to make the Penning trap devices attractive neutron sources: unlike radionucleides, you have both an on-off switch (making them easier and safer to handle and operate), and you can dial the flux of neutrons you want (making them more flexible).
Re:It's not the calorimetry, it's the neutrons
on
Excess Heat
·
· Score: 3
Now, as it happens, all of the candidate reactions using deuterium as a fuel generate neutrons as a side product. It doesn't really matter how you combine the Deuteriums, whether you combine them with other Deuteriums or with Tritium, or how you manage to overcome the electrical repulsion of the nucleus. If you induce fusion in the fuels they used, you will get neutrons. A lot of neutrons.
Sorry to pick nits, but your claim that every fusion reaction involving deuterium leads to neutron production is not true. The following are the most obvious examples:
D + D -> T (1.01 MeV) + p (3.02 MeV)
D + He3 -> He4 (3.5 Mev) + p (14.7 MeV)
D + Li6 -> 2 He4 + 22.4 Mev
All of these reactions are exothermic, so they could, at least in principle, be candidates for anomalous energy production. However, upon closer examination, they each have problems explaining the data, so your original point--that one has no business calling something fusion without being able to measure any of the fusion products--still stands. Take the first reaction: it produces tritium, which could then react with the deuterium and make 14 MeV neutrons. Moreover, about half the time the D + D reaction branches instead into an He3 and a neutron, so neutrons seem to be inextricably linked with the first candidate reaction. The third, D + Li6, is not a good candidate if one has no lithium in the system to begin with, and if one doesn't produce any measurable He4 at the end of the day. Regarding the second candidate reaction, a drawback to explaining the Pons and Fleischmann results as being this reaction chain is that, given He3's relative paucity on Earth, one finds it difficult to imagine how He3 got into the experiment by accident.
The moon's mass is about 7.4 * 10^22 kg. The earth's mass is about 6.0 * 10^24 kg. The moon is about 3.8 * 10^5 km away. This works out to 7.6 * 10^28 joules of gravitational energy. Assume that the earth draws 10 terawatts of power from waves. (Our getting so much is a very dubious assumption, but it's fine for the sake of making estimates). This means that we can draw power at this rate for on the order of 7.6 * 10^15 sec before a macroscopic change in the moon's position would result; this time intervals which works out to about 240 million years. I would suspect that before a quarter of a billion years passes someone will have figured out how to solve the problem.
Who knows, perhaps we'll want to keep the moon closer to Earth for power generation purposes--the lunar tides would be alot stronger then, and ostensibly the energy-from-tides mechanism would be more efficient.
If you believe that there is no right to property after death, but there is a right to property before death, surely you oppose the portions of the IRS code which prohibit people from giving away their wealth while they're still alive.
Your entire argument is based on a straw man: I never said that I oppose the portions of the IRS code which tax non-charitable gifts, a portion of the tax law that is entirely consistent with the position that one does not have property rights after death. It eliminates precisely the loophole that you use in your argument.
I believe that in a representative government based on democratic principles that there is both a moral authority for the government to tax and a moral authority for a person to own property. I would propose that these two beliefs are not mutually exclusive. The government provides services which require revenue to support, and, being a representative democracy, these are the services that the population by-and-large wants. This quid pro quo is entirely consistent with one's moral authority to own property.
Whether the U.S.A. is such a government is perhaps debatable....
Getting bent out of shape over the "multiple taxation" of estate taxes is silly. Practically every dollar you make is taxed multiply. First you have payroll taxes, social security, and medicare, all of which deduct a percentage of the net dollar you earn, so all three "multiply tax" the money. Then you take that remaining 50 or so cents (if you are in the top tax bracket, as I am), and you buy a candy bar with it, and in so doing you pay sales tax on the already-multiply-taxed money. Where's the moral outrage?
Please explain the moral foundation of one's "entitlement to property" after death, because I've never heard a convincing moral argument for the deceased to indeed have a right to property. I think (and I'm in good company, as so did Thomas Jefferson and Adam Smith) that of all the possible taxes, the principle of taxing someone who is no longer in existence is perhaps the fairest of all taxes. (A recent Salon.com article makes this point very eloquently).
I'm continually amused by those who claim that social programs to aid the poor should be abolished, that the "American Way" is for everyone to be self-made men and women, yet in the very next breath they feel entitled to wealth (their parents wealth) that they themselves did not earn. What could be more egalitarian than requiring that everyone make his or her way in the world without the privilege of wealth greasing the wheels overmuch? Wealth already buys one's way into the best schools, it purchases face time in the best social circles, it provides one with access to the highest reaches of government. The children of wealth already enjoy numerous advantages over the poor. As well they should--one incentive for parents to work their tails off to become wealthy is to provide for their children in the best way possible, to give them the biggest possible head start in life. However, once their children are no longer children, it is only fair--nay, it's the American Way--for them to make their own way in life. They have the best preparation for success that money can buy. Now let them go contribute to society and build their own fortunes.
No doubt the estate tax needs to be modified so that the family farmer's family, the object of so much pithy rhetoric, doesn't lose the farm just because the hapless farmer fell into his combine. But this is an entirely different matter from setting up a government system that perpetuates an aristrocracy built upon old money. A need for reform is not a call to abolish a system. Social security needs reform--does that mean we should abolish it altogether as well?
Slashdot Mirror
I believe you have misquoted our leader. The quote goes:
"There's an old saying in Tennessee -- I know it's in Texas, probably in Tennessee -- that says, fool me once, shame on - shame on you. Fool me -- you can't get fooled again." Pres. G. W. Bush Nashville, Tenn., Sept. 17, 2002.
It should be noted that the Journal of Applied Physics is regarded by much of the physics community as one of the crappiest journal extant in terms of standards for publication. I would honestly be surprised if, during the last calendar year, even one article showed up in JAP that wasn't first rejected by one (or a host) of other journals.
If there were articles on the DOE's site that are not available anywhere else, couldn't the DOE ditch the rest and keep just those available. At the very least, it would irritate the money-grubbing assholes who wanted the entire site shutdown.
The short answer: Yes. The DOE should have only removed the offending articles and left the rest of the database intact.
I have written many articles that were in the OSTI database. As a matter of fact, at my organization publication entails, as a matter of institutional and DOE policy, sending a draft copy of each article or report to the library for OSTI dissemination. Frankly, I'd always wondered about the legality of this, given that in order to publish the article in scientific journals I need to sign a copyright transfer form that assigns all copyright rights to the publisher.
Upon transferral, the publishing company owns all rights to the work in question and to any "revised or expanded derivative works based on the Work." (I'm quoting from the IEEE copyright form in front of me). In my reading, the reports I send to the OSTI database would constitute a derivative work based on the Work submitted for publication. The publisher's request that the DOE not disseminate this material for free is, I'm afraid to say, their legal prerogative, however much you or I may disagree with the principle of their doing so. However, I would argue that no single publisher has the right to shut down the entire database; they only may exercise rights over their own copyrighted works and not over any other work in the database.
If I were running the DOE, I would have played hardball with the publisher: I would have requested the publisher to list, article by article, specifically those works that they own the copyrights to and then excise only those articles from the database (after they've demonstrated their ownership of the copyrights) while leaving the rest intact. Then, DOE policy would be modified to require all DOE-supported research to be published in only those journals that authorize OSTI dissemination of the work. Moreover, "public service" (i.e., editorial service or referee duties) for "for-profit" publishing companies (defined as those companies which do not permit OSTI dissemination) would not be permitted by employees who bill their time to DOE grants and contracts.
Such a boycott would honor both the legal copyrights of the scientific work as well as the notion of public access to taxpayer-funded research. In time, as other scientific organizations (NSF, NIH, etc.) follow suit, publishers who choose to exercise their copyrights in this fashion can expect to become progressively more marginalized by the community at large. The government has already, in most cases, footed 99.9% of the cost of producing a scientific work. Retaining the right to list the work in the OSTI database should not be too much to ask.
The founders also wished to restrict voting rights to white land-owning male citizens, so I wouldn't be too quick to adopt something merely because of their stamp of approval.
/. posts, most U.S. citizens lack the mathematical sophistication and the objectiveness to recognize deficiencies in the process and to evaluate alternatives).
The 12th Amendment was ratified shortly after the "Crisis of 1800." In 1800, the U.S. government came precariously close to falling apart when the presidential/vice presidential balloting produced a tie. Jefferson and Burr tied in the 1800 presidential election with 73 votes apiece in the electoral college. The vote moved to the House to resolve the tie. 9 states were needed to win, but in ballot after ballot, Jefferson got 8 votes, Burr, 6. The succession procedure defined by the U.S. Constitution appeared to be failing, and one can only speculate on what chaos would have ensued had the Constitutional crisis not been resolved by some careful maneuvering of Jefferson's chief rival, Alexander Hamilton. Only after Hamilton persuaded key Federalists to vote for Jefferson were the President and Vice President selected. Hamilton was thanked for his service to the country--he was shot dead in a duel four years later by, you guessed it, Aaron Burr.
Bad design in the succession procedure almost destroyed the United States once, and the People, reconizing inferiorities in the process, were moved to change the way the system worked. A quick perusal of the list of amendments to the Constitution shows that changes in the election process have been made many times when what was in the highest law of the land was understood to be inferior or obsolete. One wonders if it may be possible for this to happen again in light of advances in voting theory. (I'm doubtful of the practicality of this, however. As one can tell by a quick read of the
Multiple postdocs weren't necessarily the kiss of death in physics when I went through the process not that many years back; I did two postdocs and had the luxury of turning down a couple of faculty positions before taking a staff position at a national lab. Either things have changed somewhat, or else much must depend on the field and the circumstances.
Alas, you never never never (never!) want to get passed over for tenure. If ever there were a genuine kiss of academic death, this would be it--no self-respecting university would opt to wear Princeton's or Stanford's hand-me-downs when they could just advertise in Physics Today and get 50+ highly qualified applicants who don't carry this baggage.
Personally, I just don't see what the allure is of academic life over, say, being a staff scientist at a national lab. I'd much rather just do research full-time and advise the occasional grad student and postdoc than put up with 300+ student intro courses, faculty committees from hell, petty university politics, and the pressure of bringing in enough research cash to satisfy a tenure committee. Your research ends up being very narrowly focussed while you try to carve a niche for yourself; this is the antithesis of creative scientific enterprise, IMO, and a great way of burning out during your most creative and productive years. All this while wearing the obligatory "I'm an assistant professor--kick me in the teeth" smile around the more senior faculty who control your fate.
I second the recommendation of "A Ph.D. is Not Enough!." It is an outstanding read and valuable information for those masochists who wish to try the academic route.
Who are you to decide what can and can't be compared? A language is a language, and I can compare and contrast the features of languages as I please. Note that I talked about FORTRAN 77 in my post, not F90 or F95. It was a fair comparison - F77 is deficient when compared to a modern language such as C++.
Read my comment on comparing modern automobiles with the Model A. Fine by me if you want to make the comparison--you just look a bit silly is all. Shall we dismiss C++ out of hand because pre-ANSI/ISO-Standard C++ had no convergence on template functionality? (In practice, compiler vendors still often don't implement the language standard fully, but I digress...).
These issues have been so well debated that I'm not even going to go into them here. You're wrong. Time and experience have shown this to be true. The "cute" features that bother you are important to writing large, maintainable projects. Can they be misused? Yes. Sure. But to programmers skilled with the language, they provide valuable tools that FORTRAN just doesn't offer.
Two obvious comments: First, language selection is entirely subordinate to design and planning of a project. A skilled programmer with an understanding of modern data structures and algorithms can, with very few exceptions, implement any of them in any sufficiently powerful language. You don't need C++ to do linked lists or hash tables or to have OO structure, or even to design large maintainable software projects. I have seen many large C++ projects fail because of poor design, and I have seen many large F90/F95 projects succeed because of superior design. The design decision on what language to use for a given project should be based solely on technical criteria and not kneejerk prejudices.
Second comment: Those "skilled programmers" you refer to rarely seem to end up writing the code I have to read and maintain. I myself don't work in a software house--I am a scientist and I work with scientists, people who, unlike me, often don't have formal computer science training or any compelling inclination to learn to write beautiful, well-structured C++. We're not talking about people who have read, understood, and know how not to abuse the information in Design Patterns--we're talking people with PhDs in physics or engineering who have taken a 1-week "How to get stuff to compile with a C++ compiler" syntax course who suddenly feel qualified to code up the world only to end up, more often than not, reinventing (badly) features already in the STL. They have only a nebulous understanding of the subtle aspects of C++, and they tend to cling to "natural-looking" constructs and idioms that just end up bogging down their code factors of two to ten compared to equivalent Fortran. These people are much less of a hassle if you let them program in Fortran.
I maintain, and I have considerable experience to back this claim, that because C++ is so large a language, to have any measure of refactorability you have to have programmers who have a solid understanding of the language as well as significant training in software design. Otherwise you get an unmaintainable mess, and the power and flexibility of C++ is overkill. Even if you do have such a team of programmers, it is not clear that C++ is always the language of choice: if you are not leveraging old C++ class libraries, you anticipate no need for refactorability of the current class libraries, if you have specific performance objectives (e.g., heavy array manipulation, which other languages handle better) or design objectives (a deadline that cannot tolerate the longer design cycle of designing/testing C++ class libraries), then C++ is simply not the best choice.
As for introsort, I concur with your technical assessment of the algorithm. My only reason for broaching the subject was to comment that Press et al. had presented a set of technical reasons for preferring heapsort in many settings (not the least of which is that the algorithm is dead easy to understand) and not some misguided attempt to crown an "ultimate sorting routine" as your original caricature suggested.
"Don't drive a Ford. Fords suck because my 2003 Caddy has way more features than a Model A." Let's compare apples and apples, shall we? It is unfair to compare Fortran 77 with modern languages such as C++. Either you compare Fortran 77 on its own merits with C++ circa 1977 (i.e. nothing) or else you should consider the most modern incarnation of Fortran (Fortran 95) with C++.
Fortran 95 has pointers; I have implemented essentially all "modern" data structures--hashes, heaps, tress, linked lists--with it. You can program object oriented F90/F95 (with one caveat--see "*" below). If you are averse to the ugly language features of F77 and its predecessors and their inclusion in modern Fortran for backwards compatability reasons, then you are welcome and encouraged to program in a beautified and simplified subset of the language such as the language "F" and forget about Holleriths and EQUIVALENCEs and the like.
As has been commented by others, Fortran is very efficient and relatively simple to optimize for certain kinds of tasks seen frequently by those in the science and engineering fields. Fortran 95 has nice array syntax that is dead easy for compilers to detect and optimize. Fortran also allows for a fairly rapid development cycle--the language doesn't have as much capacity for "being cute." You may hate the simplemindedness, but in the real world people have to read, understand, and maintain other people's code. All those cute syntactic tricks one feels obliged to play in C or C++ just add to the time it takes for someone to read over your code and figure out what the heck it is supposed to be doing. In my experience other peoples' Fortran 90 is so much easier to human-parse than other peoples' C or C++ that that alone would be enough to justify the language on any kind of nontrivial project. (That, and you don't have to piss away your time worrying about whether a method should be virtual or pure virtual, but I digress...).
As for your slam of heapsort in Numerical Recipes, in my recollection Press et al. identified many specific technical reasons for suggesting heapsort for many numerical applications over quicksort: it is an "in-place" sort, it has the same computational complexity as quicksort, the worst case scenario is only about 20-30% slower than the best case scenario, and, yes, its implementation in some languages is more transparent. (People who use numerical algorithms that they don't take time to understand should be forced to take the maiden ride in the airplanes or drive over the bridges they design). In many numerical applications one needs to sort large "nearly sorted" datasets; naked quicksort without some kind of shuffling dies a painful N-squared death on these cases.
(*) To be fair, I should comment that "object oriented" F95 doesn't support runtime polymorphism, but I've yet to encounter a numerical analysis task that would benefit in any significant way from this language feature.
There is some decent new stuff, though you have to look a bit if you want quality "hard" science fiction rather than space opera fluff. Stephen Baxter (Moonseed, Titan, Manifold Time), for example, writes some fairly decent classic science fiction. Greg Bear (Darwin's Radio among others) has his moments. And a new guy, Ken Wharton (Divine Intervention, has injected some fresh, if somewhat provocative, ideas into the genre. (Ken was a friend of mine from grad school, so I can't help but plug his book).
That said, of course not every application is going to achieve that level of performance.
I agree. It should be noted that 10^6 by 10^6 matricies are rather puny, really, considering the scope of the problem they envision running on the machine. I would think that to do high fidelity hydro effectively on the machine they'd need near-peak performance on 10^9 by 10^9 matricies at the smallest. (Even this isn't that big, really--it would correspond to a box 1000 cells to a side for a non-AMR hydro calculation). It will be interesting to see in the years to come what kind of "real world" performance they get on their climate models.
It's actually more restrictive if you work for the government--playing computer games on government property, even "after hours," falls under the rubric "Waste, Fraud, and Abuse," and could potentially (though it seldom does) lead to criminal penalties as well as administrative punishment. On the bright side, there are rewards for turning in those who defraud the government. You have a coworker who likes to play solitaire during his lunch break? Kaching!
Censorship of free speech has no purpose at any time.
This is a bit naive. Political canvassing outside polling places, for example, is censored speech in the U.S., an abridgement of civil liberties for the distinct purpose of supporting democracy through fair elections. As another example, one who has access to national security information (take nuclear weapons information, for example) cannot share this information without going to jail. This has the distinct purpose (we may quibble about the validity of the purpose, but it is a purpose nonetheless) of preserving U.S. sovereignty and the safety of its citizens and allies. However, it entails an abridgement of civil liberties for those "in the know." One cannot stand up in synagogue and scream "Death to kikes!" without some expectation of legal recourse.
Freedom of speech is a lofty goal, and indeed it is treated with more reverence in the U.S. than in most other countries, but it is not and cannot be considered an absolute. Even now, with the "War on Terror," this freedom is being curtailed in the name of "national security" in ways we may well have considered impossible just one year ago. China, for reasons of its own national security, has even more severe restrictions on speech.
Personally, the most alarming aspect of this is that these companies were able to produce technology which satisfies, to a large degree, the rather tight-fisted control of information required by the PRC government. This bodes badly for those who hoped that the practical, technical difficulties of censorship would effectively block attempts to censor speech in the U.S.
My college roommate got a B.S. in Physics with minors in CS and math, and he went on to law school to study intellectual property/patent law. This has always struck me as a good way to put a physics degree to work without becoming a physicist and dealing with the requisite 5+ years of grad school getting a union card^]^]^]^]^]^]^]Ph.D.
Of course, with many subfields of physics having way more jobs than people to fill them, you might just opt to just stay in the more traditional physics tracks. Many of the national labs, e.g. LANL and LLNL, pay in the 90-100k+ range with nice benefits packages for entry-level Ph.D physicists, and they are aggressively trying to hire people. With the demographics of the labs as they are, they will likely continue being as aggressive in their hiring for another decade or so at least.
Only if you open with 1. e4 rather than 1. d4.
With one small typo you have juxtaposed the two great thinkers of modern times: Jon Katz and Kant. (Just between you and me, I would have ordered Kant before Katz, however. The former's acumen was, albeit slightly, more powerful than that of our esteemed /. colleague).
Try using the "Hey, I'm a taxpayer, I PAID for this thing..." line with, say, Los Alamos National Lab weapons designers regarding plans and the detailed operation of the B61 or W76. Simple curiosity does not entail a "need to know." Nor does filing a 1040.
If you can get the information you want from the NSA, then more power to you. I'd suspect, however, that you won't get all the nitty gritty details you're after. (I could be wrong though--this is far abreast of either my expertise or whatever experience I may or may not have with classified computing systems).
Yo, some Slashdot user who has access to this thing - tell us what administrative security is in force!
You're asking alot--I'd guess that you won't get a discussion in the open about details of admin security on Intelink. I'm no expert on the matter, but I'd suspect that this is classified national security information that would entail a loss of employment and/or jail time for the offender. Not to mention the fact that it would likely compromise the security that you are so concerned about in your post.
Remember back in the day when you could just walk into a store, drop a not-insignificant amount of money at the register, and then walk out of the store with an item you just bought? Seems quaint, no? Nowadays, after your purchase you get to stand in yet another line while a puke with an attitude and a pink magic marker signs his name ("X") on your receipt and "authorizes" your exit with your personal property. The entire legal concept of quid pro quo has been turned on its ear to accommodate these pink X's--we apparently no longer own the item when we exchange money for it, but rather the store can demand that you produce proof that your property didn't magically turn back into the store's property in the 10-foot walk from the register to the door. Remember the indignation we all used to feel at being treated like criminals just so a store Fry's Electronics could cut down on cash-register fraud? (Apparently, it's much less expensive to alienate customers than to just pay the employees enough to make them value their jobs). Remember how we all vowed never to shop in such a place anymore? Now this behavior is endemic--like the sheep we are, we accept it for that extra 5% off the purchase price.
Be sure you get rankled now. Five years from now, when the only place that'll sell you food is a urine-stained 7-11 in Compton because your face is a 92.4% match to a convicted felon in Joliet, you'll be forced to accept it. By then it'll be too late.
"If you don't do anything wrong, you have nothing to fear."...
I find this excerpt from the linked article to be interesting: Although ChoicePoint says it has records on nearly every American with a credit card, it doesn't always provide access to that data. The company's Autotrack service is popular with many agencies and businesses and is also used by reporters at The Wall Street Journal. But entering the name of FBI Director Louis Freeh into the Autotrack database produces an error message. A company spokesman says ChoicePoint intentionally blocks Mr. Freeh's records as an act of good corporate citizenship.
Translation of the last line: "A company spokesman says that the publicly held firm, ChoicePoint, is not so stupid as to endanger its stockholder's investments by providing information on the man heading one of ChoicePoint's biggest client organizations." Apparently this comment by the ChoicePoint drone is intended to make us all feel better, as if we all hobnob with politically heavyweights of Freeh's standing.
(Pardon my speculation as to the content of the article; at the present time it is inaccessible to me). Using nuclear weapons to guide asteroids and comets is an interesting idea, but the choice of using it for warmaking is odd. From a practical sense, the object would be far from "stealthy;" an intense X-ray source emmanating from a chunk of sky would almost certainly be detected, given our current capability, and it would tell us exactly where to look. We'd certainly wonder about that object coming towards us, especially when it changes course en route to Earth.
Any state with the technological savvy and nuclear arsenal to conduct such an activity would be able to dispose of its enemies in another fashion, and its enemies no doubt could dispose of the aggressor. (If not, then more conventional modes of attack could be used by the aggressor with greater precision, flexibility, and lower cost). Besides, the number of possible adversaries a target state would have that could conduct such a mission is very limited, so the target state would know full well who lofted a big snowball at them, and they would merely respond in kind prior to impact with whatever arsenal they had available. The only real use I can see for such a technology is to somehow coordinate it with a first-strike nuclear attack, with the big space rock knocking out hard targets such as underground command centers.
An alternative, peaceful use for such a technology would be to bring resources such as H2O to places we'd like to colonize. Slam a comet into the moon or Mars to bring water there, for example. Unlike trying to precisely control the descent of a chunk of ice onto Earth, a dicey game at best, one could instead direct the comet toward a different celestial body and have a much larger margin of error.
The confinement scheme you are talking about (commonly referred to as an "electrostatic confinement" or a "Penning trap") is actually closer to hot fusion than cold fusion. The device works by making a spherical electrostatic potential that both confines ions and causes them to oscillate about the center of the device. If the electric field is sufficiently strong that the relative speeds of the ions is high, then collisions among the particles can, on occasion, lead to fusion reactions.
This link takes you to web pages describing (and showing nice color photos of) an ongoing Penning trap fusion experiment being conducted at the Los Alamos National Laboratory. The main drawback to devices of this type for power production is that one is limited in the number of ions that can be confined (the Brillouin limit) due to space-charge effects. Even without achieving breakeven, these devices are still very interesting, however, as enough fusion reactions take place in a small enough package to make the Penning trap devices attractive neutron sources: unlike radionucleides, you have both an on-off switch (making them easier and safer to handle and operate), and you can dial the flux of neutrons you want (making them more flexible).
Now, as it happens, all of the candidate reactions using deuterium as a fuel generate neutrons as a side product. It doesn't really matter how you combine the Deuteriums, whether you combine them with other Deuteriums or with Tritium, or how you manage to overcome the electrical repulsion of the nucleus. If you induce fusion in the fuels they used, you will get neutrons. A lot of neutrons.
Sorry to pick nits, but your claim that every fusion reaction involving deuterium leads to neutron production is not true. The following are the most obvious examples:
D + D -> T (1.01 MeV) + p (3.02 MeV)
D + He3 -> He4 (3.5 Mev) + p (14.7 MeV)
D + Li6 -> 2 He4 + 22.4 Mev
All of these reactions are exothermic, so they could, at least in principle, be candidates for anomalous energy production. However, upon closer examination, they each have problems explaining the data, so your original point--that one has no business calling something fusion without being able to measure any of the fusion products--still stands. Take the first reaction: it produces tritium, which could then react with the deuterium and make 14 MeV neutrons. Moreover, about half the time the D + D reaction branches instead into an He3 and a neutron, so neutrons seem to be inextricably linked with the first candidate reaction. The third, D + Li6, is not a good candidate if one has no lithium in the system to begin with, and if one doesn't produce any measurable He4 at the end of the day. Regarding the second candidate reaction, a drawback to explaining the Pons and Fleischmann results as being this reaction chain is that, given He3's relative paucity on Earth, one finds it difficult to imagine how He3 got into the experiment by accident.
The moon's mass is about 7.4 * 10^22 kg. The earth's mass is about 6.0 * 10^24 kg. The moon is about 3.8 * 10^5 km away. This works out to 7.6 * 10^28 joules of gravitational energy. Assume that the earth draws 10 terawatts of power from waves. (Our getting so much is a very dubious assumption, but it's fine for the sake of making estimates). This means that we can draw power at this rate for on the order of 7.6 * 10^15 sec before a macroscopic change in the moon's position would result; this time intervals which works out to about 240 million years. I would suspect that before a quarter of a billion years passes someone will have figured out how to solve the problem.
Who knows, perhaps we'll want to keep the moon closer to Earth for power generation purposes--the lunar tides would be alot stronger then, and ostensibly the energy-from-tides mechanism would be more efficient.
If you believe that there is no right to property after death, but there is a right to property before death, surely you oppose the portions of the IRS code which prohibit people from giving away their wealth while they're still alive.
Your entire argument is based on a straw man: I never said that I oppose the portions of the IRS code which tax non-charitable gifts, a portion of the tax law that is entirely consistent with the position that one does not have property rights after death. It eliminates precisely the loophole that you use in your argument.
I believe that in a representative government based on democratic principles that there is both a moral authority for the government to tax and a moral authority for a person to own property. I would propose that these two beliefs are not mutually exclusive. The government provides services which require revenue to support, and, being a representative democracy, these are the services that the population by-and-large wants. This quid pro quo is entirely consistent with one's moral authority to own property.
Whether the U.S.A. is such a government is perhaps debatable....
Thanks for the link--it is very interesting reading and excellent food for thought re: the purpose and future of NASA.
Getting bent out of shape over the "multiple taxation" of estate taxes is silly. Practically every dollar you make is taxed multiply. First you have payroll taxes, social security, and medicare, all of which deduct a percentage of the net dollar you earn, so all three "multiply tax" the money. Then you take that remaining 50 or so cents (if you are in the top tax bracket, as I am), and you buy a candy bar with it, and in so doing you pay sales tax on the already-multiply-taxed money. Where's the moral outrage?
Please explain the moral foundation of one's "entitlement to property" after death, because I've never heard a convincing moral argument for the deceased to indeed have a right to property. I think (and I'm in good company, as so did Thomas Jefferson and Adam Smith) that of all the possible taxes, the principle of taxing someone who is no longer in existence is perhaps the fairest of all taxes. (A recent Salon.com article makes this point very eloquently).
I'm continually amused by those who claim that social programs to aid the poor should be abolished, that the "American Way" is for everyone to be self-made men and women, yet in the very next breath they feel entitled to wealth (their parents wealth) that they themselves did not earn. What could be more egalitarian than requiring that everyone make his or her way in the world without the privilege of wealth greasing the wheels overmuch? Wealth already buys one's way into the best schools, it purchases face time in the best social circles, it provides one with access to the highest reaches of government. The children of wealth already enjoy numerous advantages over the poor. As well they should--one incentive for parents to work their tails off to become wealthy is to provide for their children in the best way possible, to give them the biggest possible head start in life. However, once their children are no longer children, it is only fair--nay, it's the American Way--for them to make their own way in life. They have the best preparation for success that money can buy. Now let them go contribute to society and build their own fortunes.
No doubt the estate tax needs to be modified so that the family farmer's family, the object of so much pithy rhetoric, doesn't lose the farm just because the hapless farmer fell into his combine. But this is an entirely different matter from setting up a government system that perpetuates an aristrocracy built upon old money. A need for reform is not a call to abolish a system. Social security needs reform--does that mean we should abolish it altogether as well?