Do yourself a favor and go read some studies on the history of climate change over the last 100,000 years (taken from ice samples). I don't argue that dumping stuff into the atmosphere is bad, but fluctuations in global climate is rather common. There have been times in the planet's past (within the last 100,000) years where the climate was MUCH warmer with much higher concentrations of C02.
I think you could extend that to "if you want security you must never run any executable file that didn't come with your OS outside of a VM sandbox".
Anything that can be executed is a security threat. Random executables received from mails with "3bl4rg3 yu0r p3n1s" more than others, but few softwares actually have a bug/issue count of 0.
There are things that can be done that allow for significant improvement over the current situation. The only catch is that it requires significant change to the base system, and some changes to applications that wish to run on the system.
Instead of forcing everything outside of core OS software into a VM sandbox that completely restricts access to everything, you can have the system itself mandatorily mediate all access to system resources - in a sense have everything (except the kernel) sandboxed, but each sandbox flexibly defined according to a policy declaring what the software running inside the sandbox requires access to. Doing such a thing, of course, would require you to build such a security system into the kernel itself, preferably with some separation of policy and enforcement. You would also have to build a robust policy for the core system services, restricting each to their least privilege, and fixing anything that has unnecessary demands for access to system resources. At step two you would need to encourage more general user software to minimise unnecessary access demands, and generally making use of this new access control system (which would replace and/or augment file permissions, groups, etc.); hopefully applications would also provide suggested policy requests as well.
That sounds like an awful lot of work, but thankfully we already have the kernel security structure and core sytem utility fixes. That means that we are at step two - we need to start promoting this system as the new access control mechanism, and try and get application developers to play nice with, and use, the new system.
Someone else posted "Intelligent Falling". I'll go the other way and point out that it is entirely possible to (with a little selective presentation) construct an argument, almost identical to the ID argument against evolution, that shows gravity to be a flawed theory: Uncaused Force.
The point is that any scientific theory is going to have some things that cannot presently be explained, be it through lack of evidence or an incomplete theory. If all you have to do to "refute" a theory is show it doesn't currently explain everything, and then in turn claim this as positive evidence for your theory (which explains everything by invoking magic), then pretty much everything is up for grabs. Once you start down the ID path, it's easy enough to deny everything.
OK. If you prefer Linux to OSX as a desktop GUI... WHY BUY A MAC? Similarly equipped Wintels are cheaper.
Not that I'm in the category of someone who bought a Mac to run Linux, but I can see the potential appeal.
As the Apple fans will regularly claim, particularly when it comes to laptops, Macs are not that much more expensive. Add to that the fact that, again particularly in the laptop world, Macs come with a very specific predesigned hardware configuration making a working Linux install that much easier - getting Linux running on an average Wintel laptop isn't that hard, but you have to shop around to make sure all the hardware is going to be decently supported; with a Mac laptop your odds are much better, and you don't have to check system configuration details: if MacBooks are supported by distro X then it will work.
Finally there's the fact that Apple makes nicely designed hardware. Some people are willing to pay money for a nicely designed piece of hardware. Given that the chances of it "just working" with Linux are better, and you aren't paying that much more for it - surely you can see some potential appeal to various market segments.
The problem I have with this line of thinking is that it tends to get taken too far and quickly lead to an escalation of punishment to the point where everything is punishable by death. Is it reasonable to shoot someone spray-painting graffiti on your wall? Presumably you should only be allowed to do what is required to restrain the other person. But then they may well be armed, and your effort to restrain them may well result in them shooting you, so do you have to risk being shot before you are allowed to shoot them? How about you draw your gun and warn them you'll shoot them if they continue? What if they do continue?
How about we try the same example except this time it's someone parking in your parking space. They've infringed upon your property, so presumably you can use whatever force ncessary to defend that property.
The point is that while the basic idea has some merit, at some point you have to give up trying to reason from some pure ideal and inject some practicality into things.
Nothing builds more respect amongst an intelligent audience than someone in argument admitting "Yes, actually you're right, I (hadn't considered it that way/didn't know that fact)". Being able to consider the information and make an about face given sufficient cause is a sign of maturity and that you actually think about what you are discussing.
Of course I said "intelligent audience" - there are plenty who will see such change as a sign of weakness.
And just as a pre-emptive point: I'm not talking about, for example, Kerry's "flip-flopping". While he did apparently change his mind, part of the whole problem was that he wasn't willing to admit that he changed his mind. Had he said "At the time I believed in the war in Iraq, but I have now come to see that I was wrong in that belief and oppose it" things might have been different.
Remember, three days elapsed between the bombing of Nagasaki and Hiroshima. Yet, in those three days they didn't surrender even after KNOWING what our bomb could do.
Three days after Hiroshima the Japanese were still trying to ascertain what had happened. The city was levelled, the survivors were scattered amongst the outlying villages and communications to the area were poor at best. It's not really fair to say that the Japanese knew what the bomb could do by that point. Second guessing history is a fool's game, but it is quite possible that Japan would have surrendered even if Nagasaki had not been bombed once the full extent of what had happened in Hiroshima had been made clear. As it stands we will never know.
Often times a non-science major is required to take no more than what amounts to a single survey course in math as an undergraduate.
I believe the sort of math they were testing is the kind that one would reasonably expect most students to have covered prior to entering college in mandatory math courses. I think the issue is more to do with the fact that we take illiteracy seriously, but innumeracy is just written off: if a child fails to learn to read, much effort is thrown into remedial education to try and bring them up to standard; if a child fails to learn to do basic arithmetic, and mathematics, then very little is actually done.
Of course, as the article points out, illiteracy is still a problem. As long as a child can read and write the problem is considered solved - the quality of their reading, and more importantly their writing, is not generally considered.
An interesting point I read elsewhere involves the changing writing style of students as computers and word processing become more prevalent. Due to the ease of editing on computer, as opposed to handwritten work, students tend to learn to construct more complex sentences. On the other hand students also tend to commit to writing sooner and with less thought, which results in more stream of consciousness writing and poorer quality in general.
And this conjecture is based on what? Certainly not any real world evidence. The Cell architecture is completely untested and a radically new design for a commodity chip. On paper it looks decent, but so did Itanium (and, technically, Itanium is quite good... except that the software has never been able to properly exploit it. Much the same may be true for Cell).
Software is, of course, a key issue. There are some possible solutions though, and I think IBM is already working on them. This is one of the areas where IBMs support for Linux may start to pay off. There's a lot of work going in to get Linux on Cell working well and taking full advantage of the chip. By the time Cell comes onto the market as a commodity server and workstation chip it may have pretty decent software support - at least if you are willing to use Linux. That, of course, has pros and cons of it's own. It makes it much harder for Cell chips to make a dent in the desktop market, but the server market and high-end workstation market might well be quite viable. I do agree that, pending actual results in the market, it is hard to predict how things will go. At best Cell looks like an interesting idea that may manage to take off and would be somewhat revolutionary if it did. We will have to see.
Slashdot Size Effects: The large size of the slashdot community means that only the first hour of postings receives any sort of attention. Thoughtful posts that come late to the thread are lost. The dupe gives a second set of respondents a chance to be among the first 100 posts and contribute their comments.
I actually think this is a serious issue and deserves more than just occasional dupe postings. I've heard a variety of suggestions for how to mitigate the issue somewhat, but I don't think I heard anything really good yet. Hopefully this can be one of the discussion topics CmdrTaco gets to eventually. The best suggestions I've heard so far include:
Extending the comment sorting options to include "Randomised" which preserves threading but sorts threads randomly rather than oldest first. Make "Randomised" the default sorting option.
Providing some way to "spend" Karma, be it when posting or moderating, to increase the score of a post. Presumably that should work on an exponential scale so that bumping the score by 1 extra is only expensive, while sending a post straight to score 5 takes you from the Karma cap to 0.
Neither really quite solve the problem, and both have issues of their own, but hopefully they might spur some other more interesting and successful ideas.
My greatest hope, however, is not that a revitalized Perl will squash the other dynamic languages (Python, Ruby, PHP, ECMAScript, etc) but will instead bring them into a state of interoperability. I really, really want Parrot to succeed so well that the other languages decide to target it as a backend so I can trivially call Python or C libraries from Perl and vice versa.
I'm really looking forward to Parrot as well. It has the potential to unify Perl, Python, and Ruby in a very nice way - the languages all keep their distinctive qualities, but anyone has access to all the wonderful libraries available for any of them. Moreover, it opens the door for new scripting languages: where before a new language had to manage to rally enough support to get sufficient libraries developed they can now hook into Parrot and get instant access to the richness there. For instance I very much like the look of Amber as a scripting language with maintainability in mind. More choices is a good thing.
An unfortunate consequence is that his brand of extremism is likely to make more realistic claims and analyses less acceptable to the mainstream.
A fortunate possible consequence is that such extremism may shift the "center-of-opinion" towards (but not into) extremist alarmism
There's more at work here than just that I suspect. Various studies (try this for an example) have seen that like-minded groups tend to naturally polarise. That is, if you get a group of people who have simliar opinions on subject "x" together for a discussion then the members will tend to come away with a more extremist or polarised opinions on that subject. As we have more information at our disposal, usually far more than we can easily consume, there is a tendency to consume material from generally like minded individuals, thus further polarising your own opinions. Worse, as opinions move further to extremes it becomes increasingly easy to find differing opinions utterly unpalatable: they tend to be extremist in the opposite direction, and speak, effectively, in a different language. It is entirely possible that the "centre-of-opinion" doesn't move at all, but instead anyone on either side of the issue simply gets drawn to one extreme or the other. You can certainly witness this effect in US politics.
Apple doesn't sell computers. They don't sell operating systems. They sell integrated systems. To do that, they must control the hardware and software.
Which, as other here have been pointing out, could be done by strict constraints on licensing of OS X. Technically they don't have to build the hardware to control the hardware.
I find it interesting that so many people, many of them apparently Apple fans, have so little faith in the quality and design of Apple hardware that they think Apple would crash if they didn't have OS X as a carrot to make people buy the hardware.
That said I'm not suggesting that Apple needs to license OS X. They seem to be making money doing what their doing so there's not necessarily any reason to change. To move to licensing the OS would be to change business models and focus, and I can understand if Apple doesn't care to do that. I just don't think it is quite the terrible and crippling move that many here seem to make it out to be.
It depends on how you look at it. To construct the real numbers you generally need to have first constructed the natural numbers, but to axiomitize the real numbers you simply need to provide a set of rules that will uniquely define the real numbers, and those rules may not uniquely define the natural numbers - that is, you can describe the real numbers as a whole, but require extra information/rules to actually be able to pick out the natural numbers. If you can't actually pick out the natural numbers then natural number arithmetic can't be constructed. What do I mean? Well take the Goldbach conjecture as a nice "hard to prove" problem that could well fall into the "unproveable" category: Every even number greater then 2 can be expressed as the sum of two (not necessarily different) prime numbers. While all even numbers and primes are real numbers, an axiomitization of real number arithmetic need not define "even" or "prime" - these are properties that require you to be able to pick out the natural numbers: without defining 2 you can't define "even", and every real number is divisible by every other real number, so without being able to say "divisible by a natural number" you can't define "prime". Within the domain of real numbers, with no way to pick out numbers as being natural numbers, the problem makes no sense. In this sense the behaviour of natural numbers, even though they are a subset of the real numbers, is, in fact, considerably more complicated.
How big a bird? Would an eagle with a 3 meter wingspan (that's just shy of 10' for you Americans) that hunted moa (flightless birds that weighed as much as 250kg (more than 500 pounds)) count? Because Haast's eagle was exactly that, and only became extinct around 500 years ago. One of those would have had no trouble taking down a man.
New Zealand has very unique fauna, and unlike nearby Australia (which seems to have the finest array of deadly creatures in the world) it's almost all harmless. Haast's eagle was one glaring exception, and would certainly have been a truly fearsome creature had it survived.
I'll put it another way: applied math has a tendency to look at things through a keyhole, or through the particular lens of the application being used. As long as you focus on the applications you only understand the small amount you can see clearly through that particular lens. Pure math is about trying to take the lens away and study what is behind it. If you understand clearly what you're actually looking at, rather than one particular view of it. By understanding what's actually there you have a far better chance of understanding that something else you're doing, looking through some other lens, is actually looking at exactly the same thing. In some sense it is precisely because you were taught applied math that you failed to make the connections between what you were being taught and what you were doing.
Would the sort of teaching I have in mind told you that linear algebra was the same as what you were doing with your graphics programming? Probably not no. But really the issue is not whether the math was applied or not, because apparently what you were taught was taught in a very applied mode. The reason it didn't click with you, from the sound of it, is becausse there are lots of different applications, and the ones the teachers chose weren't the ones that happened to fit with you.
What I would have preferred to see happen is more time spent trying to understand the deeper questions and less on the mechanics of solving immediate applications. By trying to figure out what a matrix is, what it means how it acts as a linear transform on a vector space (and what that really means), and doing the same things with algebraic equations, I would hope that you would develop the tools yourself to see the connections between matrices and the equations you were programming on computer rather than having to have someone tell you. If you understand how and why these things behave and work the way they do you have a much better chance of seeing how the underlying structure connects with so many other things, and how that structure can be applied in different, new, and potentially unexpected ways.
Fact: 98% of students don't enjoy pure mathematics.
Really? You've done studies have you? A large percentage of people have never really been exposed to pure math, and often when they have its been taught in very poor fashion. I've done a fair bit of teaching in mathematics, and I've dealt with plenty of struggling students who find it boring and uninteresting. Oddly enough I've found that actually exposing them to some of the wonders of pure math inspires much more interest in mathematics than yet more arbitrary word problems.
That's what applications of math are: a chance to understand, think logically, and explore. It's not teaching formulas!
I think what we have here is a failure to communicate. Being asked to work out how much change Sue should get after buying her stereo at 30% off doesn't help kids understand, doesn't require them to think logically except possibly as an accidental byproduct, and really doesn't encourage exploration.
If you tell an Algebra student to derive the quadratic equation they will fall asleep. But if you tell them there is a secret formula that can solve any ballistics equation without factoring, and they will be intrigued.
Tell them to figure out how to compute the volume of a solid of rotation and they are miserable. Tell them to compute the volume of Mrs. StinkyTeacher, given the equation for her gut. Now they get it.
The juxtapositions you present there are all applied math problems, you're just putting the latter versions in fancy dress to make them appear "cool". There's no pure math there at all. The problem is that kids get sick of all the different ways you try to dress up a problem. Eventually it all starts to wear thin and you're trying to come up with ever more ouytlandish ways to couch the same problem, but not really achieving anything.
However, there are other, non-axiomatic, statements in any formal system that cannot be proven either true or false.
That's not actually true. For instance there exists a complete and consistent axiomitization of the real numbers, and of Euclidean geometry. Godel's proof required that the system provide the ability to contruct basic arithmetical (in the sen of the natural numbers) truths. It's a common misconception that Godel's theorems apply to any formal system.
Too many teachers don't teach applied mathematics, so math becomes boring and hard. Every single forumla you learn from Algebra I to Multivariable calculus and beyond can be used to solve a problem in the very room the student sits.
You know I would say that too many teachers don't teach pure mathematics, so the joy of exploration and discovery and logical thought is lost. Mathematics becomes rote mechanical rules that you unthinkingly chug through to produce some number which is supposed to be important. There is no questioning of why those rules are what they are, why the methods work, and what the structure actually is. The focus is on teaching kids the applications of math and they never get to understand how to think about math, how to think logically, how to explore the structure of our own mental creations. Mathematics is taught with absolutely no sense of wonder, or curiousity.
Teaching kids how to apply mathematics is important, but really not that hard. Teaching kids to see math as something other than a whole list of rules and methods and mechanical applications of formulas - now that takes some real effort. That, however, is what pure mathemathatics can get you.
Not to say that all engineering departments are like that - obviously there are quite a few exceptions. However, that's how it is - Engineering is applied mathematics after all. My CS degree consisted of probably just as much math as computers, if not more. Calc 1/2/3 and lots of mathematical electives.
It's interesting that math teaching hasn't caught up with modern needs. Engineers need math, and there is a lot of focus on engineering mathematics. In practice that means lots of calculus and probably some linear algebra. That's ideal if you're going into civil or mechanical engineering, but modern electrical engineering or software engineering has very different needs in the way of mathematics. There is a real need to get serious abstract algebra considered as vital engineering mathematics for electrical and software engineers. If you're doing more CS focused software engineering then courses in category theory really ought to be in there too. Hopefully such a shift will be occurring in the next few years (or decades - sometimes it's baffling how slow things move).
Thankfully, most Americans understand that if they have no affiliation with terrorist groups, they have nothing to worry about.
Sounds very good, but is utterly wrong. Americans have nothing to worry about as long as the authorities do not know, believe or suspect that they (or anyone they have sufficient similarity (like first and last name) with someone who is known, believed or suspected to) have affiliations with groups that are known, believed or suspected to be terrorist groups.
The real danger is that action such as this sets a precendent. Let's presume for a minute that you're a rabidly loyal Bush fan and can't imagine the current administration doing anything wrong. You should still object to this. Even if we assume that currently the wiretap authority is not being abused in any way shape or form, and that this administration will see to it that they aren't abused, it lays down legal precedent that some future administration can use to justify their own wiretapping which may not be quite so palatable to you. Just imagine: 50 years from now an evil Democrat administration gets in and decides that Republicans are a threat to national security and need to tap all international phone calls from anyone with ties to the Republican party. Unless we obect to what is happening now, there will be perfectly reasonable precendent to allow such a thing to be secretly authorised.
It's just bad no matter how you cut it. You don't have to object to the current administrations policy, nor believe them capable of abusing their authority to see that.
et's say that the next Jose Padilla is not dumb enough to get caught, and does manage to work with some people to scrounge up some radiological material from some used dental x-ray machines or industrial hardware.
Then we should all... not be terribly disconcerted.
Let's say he and some buddies, living in Queens, have bought a few pounds of black powder or other modest explosive/propellant
Then you at least have something to worry about.
Let's be clear here: "dirty bombs" are, in practice, no more dangerous than conventional explosives. The radiological part doesn't really make a damn bit of difference. To keep the radioactive material sufficiently concentrated to actually be capable of doing any harm you have to have such a small explosion and spread it over such a small area that's its pointless. Spread it over a wide area with a decent explosion and it simply isn't that dangerous. The only difference a dirty bomb makes is: (1) A little extra cleanup expense, but given that we're cleaning up after a bomb already the extra expense to clean up the radiological component sufficiently is marginal; (2) People will PANIC!
So the single most dangerous part of a dirty is that people will panic. And why will people panic? Because they keep getting told that a dirty bomb is some grave and terrible threat. It isn't. And if we actually spent more time explaining the reality rather than using scare tactics we could easily and safely remove the only real threat that dirty bombs have over conventional explosives - the panic from the public.
counter = 1 while counter < 10:
do something...
couner = counter +1
in a statically typed language that is picked up at compile time. in a dynamically typed language you won't notice it until your app hangs at runtime and you'd better hope your in a situation where you can easilly attatch a debugger.
In some ways the confusion here arises from the fact that static types had benefits in optimization, and hence were necessary information for the compiler to parse correctly. At the same time static types are used as a means of checking (to a certain extent) the correctness of code. Because the compiler ends up doing the correctness checking because it needs to sort out the types it has become common to conflate compilation and static error checking into a single task, when really they are two quite separate tasks. What am I getting at here? That compiling or interpreting a program isn't really how you should be checking for correctness - it's nice but hardly necessary. If you ran python static correctness checking tool pychecker (which tries to statically catch many common errors) over that code it would throw a warning that "couner" was set but never used, happily catching the typo/spelling error. In fact if statically checking your code is important you'll find pychecker will catch a great many common errors for you.
Now as to the separation of static checking and compilation - if you're willing to see that static checking is a separate task and devote tools to just doing that then you can do a whole lot better than just static types. Check out JML which, by adding extra annotations beyond just static type annotation, allows you to use far more powerful static checking like ESC/Java2 which makes standard static type checking via javac look as poor at catching errors as a dynamically typed language.
Actually I was going to say that I think classical sheet music compared to free form jazz would be a far better analogy. Put together a trio of classical musicians with a score and they'll do passably, but will seem quite stiff and uninteresting compared to a jazz trio just jamming. On the other hand you can put together a symphony orchestra with a score and get something very respectable. Try to get 50 jazz musicians to all do a free form jam together and mostly you'll get a mess.
Slashdot Mirror
Do yourself a favor and go read some studies on the history of climate change over the last 100,000 years (taken from ice samples). I don't argue that dumping stuff into the atmosphere is bad, but fluctuations in global climate is rather common. There have been times in the planet's past (within the last 100,000) years where the climate was MUCH warmer with much higher concentrations of C02.
Has it been warmer in the last 100,000 years? No, but if they stretch that to 150,000 then yes it has been warmer than it is now. "Much" warmer is perhaps an exageration though. Has there been much higher CO2 concentrations? Not even close. Current CO2 concentrations represent spike in CO2 levels twice as large as any previous spike in the last 650,000 years. What des that have to do with temperature? Take a look at a chart of and see how closely they correlate. Now realise that in that plot current CO2 levels are at 5.5 on that scale! Yes, correlation doesn't imply causation. Basic physics regarding absroption spectra of CO2 is hat implies causation, the correlation just shows that theory bears out in practice.
Jedidiah.
I think you could extend that to "if you want security you must never run any executable file that didn't come with your OS outside of a VM sandbox".
Anything that can be executed is a security threat. Random executables received from mails with "3bl4rg3 yu0r p3n1s" more than others, but few softwares actually have a bug/issue count of 0.
There are things that can be done that allow for significant improvement over the current situation. The only catch is that it requires significant change to the base system, and some changes to applications that wish to run on the system.
Instead of forcing everything outside of core OS software into a VM sandbox that completely restricts access to everything, you can have the system itself mandatorily mediate all access to system resources - in a sense have everything (except the kernel) sandboxed, but each sandbox flexibly defined according to a policy declaring what the software running inside the sandbox requires access to. Doing such a thing, of course, would require you to build such a security system into the kernel itself, preferably with some separation of policy and enforcement. You would also have to build a robust policy for the core system services, restricting each to their least privilege, and fixing anything that has unnecessary demands for access to system resources. At step two you would need to encourage more general user software to minimise unnecessary access demands, and generally making use of this new access control system (which would replace and/or augment file permissions, groups, etc.); hopefully applications would also provide suggested policy requests as well.
That sounds like an awful lot of work, but thankfully we already have the kernel security structure and core sytem utility fixes. That means that we are at step two - we need to start promoting this system as the new access control mechanism, and try and get application developers to play nice with, and use, the new system.
Jedidiah.
Someone else posted "Intelligent Falling". I'll go the other way and point out that it is entirely possible to (with a little selective presentation) construct an argument, almost identical to the ID argument against evolution, that shows gravity to be a flawed theory: Uncaused Force.
The point is that any scientific theory is going to have some things that cannot presently be explained, be it through lack of evidence or an incomplete theory. If all you have to do to "refute" a theory is show it doesn't currently explain everything, and then in turn claim this as positive evidence for your theory (which explains everything by invoking magic), then pretty much everything is up for grabs. Once you start down the ID path, it's easy enough to deny everything.
Jedidiah.
OK. If you prefer Linux to OSX as a desktop GUI... WHY BUY A MAC? Similarly equipped Wintels are cheaper.
Not that I'm in the category of someone who bought a Mac to run Linux, but I can see the potential appeal.
As the Apple fans will regularly claim, particularly when it comes to laptops, Macs are not that much more expensive. Add to that the fact that, again particularly in the laptop world, Macs come with a very specific predesigned hardware configuration making a working Linux install that much easier - getting Linux running on an average Wintel laptop isn't that hard, but you have to shop around to make sure all the hardware is going to be decently supported; with a Mac laptop your odds are much better, and you don't have to check system configuration details: if MacBooks are supported by distro X then it will work.
Finally there's the fact that Apple makes nicely designed hardware. Some people are willing to pay money for a nicely designed piece of hardware. Given that the chances of it "just working" with Linux are better, and you aren't paying that much more for it - surely you can see some potential appeal to various market segments.
Jedidiah.
The problem I have with this line of thinking is that it tends to get taken too far and quickly lead to an escalation of punishment to the point where everything is punishable by death. Is it reasonable to shoot someone spray-painting graffiti on your wall? Presumably you should only be allowed to do what is required to restrain the other person. But then they may well be armed, and your effort to restrain them may well result in them shooting you, so do you have to risk being shot before you are allowed to shoot them? How about you draw your gun and warn them you'll shoot them if they continue? What if they do continue?
How about we try the same example except this time it's someone parking in your parking space. They've infringed upon your property, so presumably you can use whatever force ncessary to defend that property.
The point is that while the basic idea has some merit, at some point you have to give up trying to reason from some pure ideal and inject some practicality into things.
Jedidiah.
Nothing builds more respect amongst an intelligent audience than someone in argument admitting "Yes, actually you're right, I (hadn't considered it that way/didn't know that fact)". Being able to consider the information and make an about face given sufficient cause is a sign of maturity and that you actually think about what you are discussing.
Of course I said "intelligent audience" - there are plenty who will see such change as a sign of weakness.
And just as a pre-emptive point: I'm not talking about, for example, Kerry's "flip-flopping". While he did apparently change his mind, part of the whole problem was that he wasn't willing to admit that he changed his mind. Had he said "At the time I believed in the war in Iraq, but I have now come to see that I was wrong in that belief and oppose it" things might have been different.
Jedidiah.
Remember, three days elapsed between the bombing of Nagasaki and Hiroshima. Yet, in those three days they didn't surrender even after KNOWING what our bomb could do.
Three days after Hiroshima the Japanese were still trying to ascertain what had happened. The city was levelled, the survivors were scattered amongst the outlying villages and communications to the area were poor at best. It's not really fair to say that the Japanese knew what the bomb could do by that point. Second guessing history is a fool's game, but it is quite possible that Japan would have surrendered even if Nagasaki had not been bombed once the full extent of what had happened in Hiroshima had been made clear. As it stands we will never know.
Jedidiah.
Often times a non-science major is required to take no more than what amounts to a single survey course in math as an undergraduate.
I believe the sort of math they were testing is the kind that one would reasonably expect most students to have covered prior to entering college in mandatory math courses. I think the issue is more to do with the fact that we take illiteracy seriously, but innumeracy is just written off: if a child fails to learn to read, much effort is thrown into remedial education to try and bring them up to standard; if a child fails to learn to do basic arithmetic, and mathematics, then very little is actually done.
Of course, as the article points out, illiteracy is still a problem. As long as a child can read and write the problem is considered solved - the quality of their reading, and more importantly their writing, is not generally considered.
An interesting point I read elsewhere involves the changing writing style of students as computers and word processing become more prevalent. Due to the ease of editing on computer, as opposed to handwritten work, students tend to learn to construct more complex sentences. On the other hand students also tend to commit to writing sooner and with less thought, which results in more stream of consciousness writing and poorer quality in general.
Jedidiah.
And this conjecture is based on what? Certainly not any real world evidence. The Cell architecture is completely untested and a radically new design for a commodity chip. On paper it looks decent, but so did Itanium (and, technically, Itanium is quite good... except that the software has never been able to properly exploit it. Much the same may be true for Cell).
Software is, of course, a key issue. There are some possible solutions though, and I think IBM is already working on them. This is one of the areas where IBMs support for Linux may start to pay off. There's a lot of work going in to get Linux on Cell working well and taking full advantage of the chip. By the time Cell comes onto the market as a commodity server and workstation chip it may have pretty decent software support - at least if you are willing to use Linux. That, of course, has pros and cons of it's own. It makes it much harder for Cell chips to make a dent in the desktop market, but the server market and high-end workstation market might well be quite viable. I do agree that, pending actual results in the market, it is hard to predict how things will go. At best Cell looks like an interesting idea that may manage to take off and would be somewhat revolutionary if it did. We will have to see.
Jedidiah.
Slashdot Size Effects: The large size of the slashdot community means that only the first hour of postings receives any sort of attention. Thoughtful posts that come late to the thread are lost. The dupe gives a second set of respondents a chance to be among the first 100 posts and contribute their comments.
I actually think this is a serious issue and deserves more than just occasional dupe postings. I've heard a variety of suggestions for how to mitigate the issue somewhat, but I don't think I heard anything really good yet. Hopefully this can be one of the discussion topics CmdrTaco gets to eventually. The best suggestions I've heard so far include:
Extending the comment sorting options to include "Randomised" which preserves threading but sorts threads randomly rather than oldest first. Make "Randomised" the default sorting option.
Providing some way to "spend" Karma, be it when posting or moderating, to increase the score of a post. Presumably that should work on an exponential scale so that bumping the score by 1 extra is only expensive, while sending a post straight to score 5 takes you from the Karma cap to 0.
Neither really quite solve the problem, and both have issues of their own, but hopefully they might spur some other more interesting and successful ideas.
Jedidiah.
My greatest hope, however, is not that a revitalized Perl will squash the other dynamic languages (Python, Ruby, PHP, ECMAScript, etc) but will instead bring them into a state of interoperability. I really, really want Parrot to succeed so well that the other languages decide to target it as a backend so I can trivially call Python or C libraries from Perl and vice versa.
I'm really looking forward to Parrot as well. It has the potential to unify Perl, Python, and Ruby in a very nice way - the languages all keep their distinctive qualities, but anyone has access to all the wonderful libraries available for any of them. Moreover, it opens the door for new scripting languages: where before a new language had to manage to rally enough support to get sufficient libraries developed they can now hook into Parrot and get instant access to the richness there. For instance I very much like the look of Amber as a scripting language with maintainability in mind. More choices is a good thing.
Jedidiah.
An unfortunate consequence is that his brand of extremism is likely to make more realistic claims and analyses less acceptable to the mainstream.
A fortunate possible consequence is that such extremism may shift the "center-of-opinion" towards (but not into) extremist alarmism
There's more at work here than just that I suspect. Various studies (try this for an example) have seen that like-minded groups tend to naturally polarise. That is, if you get a group of people who have simliar opinions on subject "x" together for a discussion then the members will tend to come away with a more extremist or polarised opinions on that subject. As we have more information at our disposal, usually far more than we can easily consume, there is a tendency to consume material from generally like minded individuals, thus further polarising your own opinions. Worse, as opinions move further to extremes it becomes increasingly easy to find differing opinions utterly unpalatable: they tend to be extremist in the opposite direction, and speak, effectively, in a different language. It is entirely possible that the "centre-of-opinion" doesn't move at all, but instead anyone on either side of the issue simply gets drawn to one extreme or the other. You can certainly witness this effect in US politics.
Jedidiah.
Apple doesn't sell computers. They don't sell operating systems. They sell integrated systems. To do that, they must control the hardware and software.
Which, as other here have been pointing out, could be done by strict constraints on licensing of OS X. Technically they don't have to build the hardware to control the hardware.
I find it interesting that so many people, many of them apparently Apple fans, have so little faith in the quality and design of Apple hardware that they think Apple would crash if they didn't have OS X as a carrot to make people buy the hardware.
That said I'm not suggesting that Apple needs to license OS X. They seem to be making money doing what their doing so there's not necessarily any reason to change. To move to licensing the OS would be to change business models and focus, and I can understand if Apple doesn't care to do that. I just don't think it is quite the terrible and crippling move that many here seem to make it out to be.
Jedidiah.
It depends on how you look at it. To construct the real numbers you generally need to have first constructed the natural numbers, but to axiomitize the real numbers you simply need to provide a set of rules that will uniquely define the real numbers, and those rules may not uniquely define the natural numbers - that is, you can describe the real numbers as a whole, but require extra information/rules to actually be able to pick out the natural numbers. If you can't actually pick out the natural numbers then natural number arithmetic can't be constructed. What do I mean? Well take the Goldbach conjecture as a nice "hard to prove" problem that could well fall into the "unproveable" category: Every even number greater then 2 can be expressed as the sum of two (not necessarily different) prime numbers. While all even numbers and primes are real numbers, an axiomitization of real number arithmetic need not define "even" or "prime" - these are properties that require you to be able to pick out the natural numbers: without defining 2 you can't define "even", and every real number is divisible by every other real number, so without being able to say "divisible by a natural number" you can't define "prime". Within the domain of real numbers, with no way to pick out numbers as being natural numbers, the problem makes no sense. In this sense the behaviour of natural numbers, even though they are a subset of the real numbers, is, in fact, considerably more complicated.
Jedidiah.
How big a bird? Would an eagle with a 3 meter wingspan (that's just shy of 10' for you Americans) that hunted moa (flightless birds that weighed as much as 250kg (more than 500 pounds)) count? Because Haast's eagle was exactly that, and only became extinct around 500 years ago. One of those would have had no trouble taking down a man.
New Zealand has very unique fauna, and unlike nearby Australia (which seems to have the finest array of deadly creatures in the world) it's almost all harmless. Haast's eagle was one glaring exception, and would certainly have been a truly fearsome creature had it survived.
Jedidiah.
I'll put it another way: applied math has a tendency to look at things through a keyhole, or through the particular lens of the application being used. As long as you focus on the applications you only understand the small amount you can see clearly through that particular lens. Pure math is about trying to take the lens away and study what is behind it. If you understand clearly what you're actually looking at, rather than one particular view of it. By understanding what's actually there you have a far better chance of understanding that something else you're doing, looking through some other lens, is actually looking at exactly the same thing. In some sense it is precisely because you were taught applied math that you failed to make the connections between what you were being taught and what you were doing.
Jedidiah.
Would the sort of teaching I have in mind told you that linear algebra was the same as what you were doing with your graphics programming? Probably not no. But really the issue is not whether the math was applied or not, because apparently what you were taught was taught in a very applied mode. The reason it didn't click with you, from the sound of it, is becausse there are lots of different applications, and the ones the teachers chose weren't the ones that happened to fit with you.
What I would have preferred to see happen is more time spent trying to understand the deeper questions and less on the mechanics of solving immediate applications. By trying to figure out what a matrix is, what it means how it acts as a linear transform on a vector space (and what that really means), and doing the same things with algebraic equations, I would hope that you would develop the tools yourself to see the connections between matrices and the equations you were programming on computer rather than having to have someone tell you. If you understand how and why these things behave and work the way they do you have a much better chance of seeing how the underlying structure connects with so many other things, and how that structure can be applied in different, new, and potentially unexpected ways.
Jedidiah.
Fact: 98% of students don't enjoy pure mathematics.
Really? You've done studies have you? A large percentage of people have never really been exposed to pure math, and often when they have its been taught in very poor fashion. I've done a fair bit of teaching in mathematics, and I've dealt with plenty of struggling students who find it boring and uninteresting. Oddly enough I've found that actually exposing them to some of the wonders of pure math inspires much more interest in mathematics than yet more arbitrary word problems.
That's what applications of math are: a chance to understand, think logically, and explore. It's not teaching formulas!
I think what we have here is a failure to communicate. Being asked to work out how much change Sue should get after buying her stereo at 30% off doesn't help kids understand, doesn't require them to think logically except possibly as an accidental byproduct, and really doesn't encourage exploration.
If you tell an Algebra student to derive the quadratic equation they will fall asleep. But if you tell them there is a secret formula that can solve any ballistics equation without factoring, and they will be intrigued.
Tell them to figure out how to compute the volume of a solid of rotation and they are miserable. Tell them to compute the volume of Mrs. StinkyTeacher, given the equation for her gut. Now they get it.
The juxtapositions you present there are all applied math problems, you're just putting the latter versions in fancy dress to make them appear "cool". There's no pure math there at all. The problem is that kids get sick of all the different ways you try to dress up a problem. Eventually it all starts to wear thin and you're trying to come up with ever more ouytlandish ways to couch the same problem, but not really achieving anything.
Jedidiah.
However, there are other, non-axiomatic, statements in any formal system that cannot be proven either true or false.
That's not actually true. For instance there exists a complete and consistent axiomitization of the real numbers, and of Euclidean geometry. Godel's proof required that the system provide the ability to contruct basic arithmetical (in the sen of the natural numbers) truths. It's a common misconception that Godel's theorems apply to any formal system.
Jedidiah.
Too many teachers don't teach applied mathematics, so math becomes boring and hard. Every single forumla you learn from Algebra I to Multivariable calculus and beyond can be used to solve a problem in the very room the student sits.
You know I would say that too many teachers don't teach pure mathematics, so the joy of exploration and discovery and logical thought is lost. Mathematics becomes rote mechanical rules that you unthinkingly chug through to produce some number which is supposed to be important. There is no questioning of why those rules are what they are, why the methods work, and what the structure actually is. The focus is on teaching kids the applications of math and they never get to understand how to think about math, how to think logically, how to explore the structure of our own mental creations. Mathematics is taught with absolutely no sense of wonder, or curiousity.
Teaching kids how to apply mathematics is important, but really not that hard. Teaching kids to see math as something other than a whole list of rules and methods and mechanical applications of formulas - now that takes some real effort. That, however, is what pure mathemathatics can get you.
Jedidiah.
Not to say that all engineering departments are like that - obviously there are quite a few exceptions. However, that's how it is - Engineering is applied mathematics after all. My CS degree consisted of probably just as much math as computers, if not more. Calc 1/2/3 and lots of mathematical electives.
It's interesting that math teaching hasn't caught up with modern needs. Engineers need math, and there is a lot of focus on engineering mathematics. In practice that means lots of calculus and probably some linear algebra. That's ideal if you're going into civil or mechanical engineering, but modern electrical engineering or software engineering has very different needs in the way of mathematics. There is a real need to get serious abstract algebra considered as vital engineering mathematics for electrical and software engineers. If you're doing more CS focused software engineering then courses in category theory really ought to be in there too. Hopefully such a shift will be occurring in the next few years (or decades - sometimes it's baffling how slow things move).
Jedidiah.
The real danger is that action such as this sets a precendent. Let's presume for a minute that you're a rabidly loyal Bush fan and can't imagine the current administration doing anything wrong. You should still object to this. Even if we assume that currently the wiretap authority is not being abused in any way shape or form, and that this administration will see to it that they aren't abused, it lays down legal precedent that some future administration can use to justify their own wiretapping which may not be quite so palatable to you. Just imagine: 50 years from now an evil Democrat administration gets in and decides that Republicans are a threat to national security and need to tap all international phone calls from anyone with ties to the Republican party. Unless we obect to what is happening now, there will be perfectly reasonable precendent to allow such a thing to be secretly authorised.
It's just bad no matter how you cut it. You don't have to object to the current administrations policy, nor believe them capable of abusing their authority to see that.
Jedidiah.
et's say that the next Jose Padilla is not dumb enough to get caught, and does manage to work with some people to scrounge up some radiological material from some used dental x-ray machines or industrial hardware.
Then we should all... not be terribly disconcerted.
Let's say he and some buddies, living in Queens, have bought a few pounds of black powder or other modest explosive/propellant
Then you at least have something to worry about.
Let's be clear here: "dirty bombs" are, in practice, no more dangerous than conventional explosives. The radiological part doesn't really make a damn bit of difference. To keep the radioactive material sufficiently concentrated to actually be capable of doing any harm you have to have such a small explosion and spread it over such a small area that's its pointless. Spread it over a wide area with a decent explosion and it simply isn't that dangerous. The only difference a dirty bomb makes is: (1) A little extra cleanup expense, but given that we're cleaning up after a bomb already the extra expense to clean up the radiological component sufficiently is marginal; (2) People will PANIC!
So the single most dangerous part of a dirty is that people will panic. And why will people panic? Because they keep getting told that a dirty bomb is some grave and terrible threat. It isn't. And if we actually spent more time explaining the reality rather than using scare tactics we could easily and safely remove the only real threat that dirty bombs have over conventional explosives - the panic from the public.
Jedidiah.
In some ways the confusion here arises from the fact that static types had benefits in optimization, and hence were necessary information for the compiler to parse correctly. At the same time static types are used as a means of checking (to a certain extent) the correctness of code. Because the compiler ends up doing the correctness checking because it needs to sort out the types it has become common to conflate compilation and static error checking into a single task, when really they are two quite separate tasks. What am I getting at here? That compiling or interpreting a program isn't really how you should be checking for correctness - it's nice but hardly necessary. If you ran python static correctness checking tool pychecker (which tries to statically catch many common errors) over that code it would throw a warning that "couner" was set but never used, happily catching the typo/spelling error. In fact if statically checking your code is important you'll find pychecker will catch a great many common errors for you.
Now as to the separation of static checking and compilation - if you're willing to see that static checking is a separate task and devote tools to just doing that then you can do a whole lot better than just static types. Check out JML which, by adding extra annotations beyond just static type annotation, allows you to use far more powerful static checking like ESC/Java2 which makes standard static type checking via javac look as poor at catching errors as a dynamically typed language.
Jedidiah.
Actually I was going to say that I think classical sheet music compared to free form jazz would be a far better analogy. Put together a trio of classical musicians with a score and they'll do passably, but will seem quite stiff and uninteresting compared to a jazz trio just jamming. On the other hand you can put together a symphony orchestra with a score and get something very respectable. Try to get 50 jazz musicians to all do a free form jam together and mostly you'll get a mess.
Jedidiah.