It turns out that terrorism in western countries is a very rare thing, outside of a few hot areas like Spain's Basque area. This is very good, by the way.
Mining for rare event is extremely difficult. Bayes' s rule indicates that if in a database there are 0.01% actually suspicious events and your mining algorithms are extremely effective at 99% accuracy, then you still have an approximately 100:1 false positive ratio, which makes the mining still useless.
Your ideas sound good but are probably unworkable. Independent re-invention does not prove obviousness, as it happens all the time even for very difficult subjects (e.g. in maths, physics, not just engineering). The first to invent should get the credit as in research.
Your second point, I wholeheartedly agree with however. Patents are supposed to make something "patent", i.e. totally obvious to make with the knowledge of the patent. Some patents are like that but not all.
The third point is the opposite of reality. CSIRO are the one who worked their butt off to make wireless communication more reliable, effective and efficient, but they are not a manufacturing entity, they are a government research agency. They patented their discoveries as they should have. Even if Buffalo or others certainly made contributions into the manufacturing process, the basic scientific principles of their device is based on CSIRO's invention.
Nobody complains when IBM (say) patents a new silicon manufacturing process like SOI and makes wafer makes like Applied Materials pay licences. In effect the research arm of IBM makes billions a year from patent licences. People know that IBM has legal werewithals and don't think they can get away with cheating with IBM's inventions.
Here I'm pretty sure the Dells, HPs, Netgears of this world knew they were infringing but thought "who is this CSIRO thing? an Aussie gov. body? come on, can they even patent things in the US? can they even prosecure?" and tried to cheat knowing full well that at worse they might be liable for some fees down the line and might be able to settle for cheap somewhere down the line. They probably thought it would be a nice way for their legal department to make them some money for a change.
If patents like these are invalidated this is pretty much closing the lid on a whole body of strategies for applied research to fund itself through licences and commercial agreements. CSIRO is not a submarine patent shop. They employ around 6000 people and do very fine research in all areas of science. If they made a fine invention like it sounds they have, way in advance of everybody else, they should get a fine reward. This is the heart of the patent system.
If this doesn't work then the patent system is not only ineffective, it is simply evil.
1- Photons have no rest mass but they have energy and momentum (or relativistic mass). They are affected by gravity as they follow geodesics in spacetime, which are defined by the local mass-energy.
They original experimental proof of general relativity was by Eddington who showed that the mass of the Sun bends light in the way predicted by GR.
2- Mass by itself is not sufficient to "trap" light. A given mass contained within a given volume is, however, i.e a given density. A sufficient density creates an event horizon. GR predicts that all mass hidden within a event horizon must end up in a singularity.
Even if someone proved that P=NP and exhibited a P algorithm to solve all the NP problems, it may not be efficient in classical terms. If the complexity of that algorithm is O(N^1000) say then we are still better off using the meta-heuristics that are in used right now to solve them approximately.
FYI there are efficient approximate solutions to the TSP that are guaranteed to come within a few percent of the true optimum. A P=NP proof may not be worth that much in real dollars.
Linear programming with n variables and m constraints (n not necessarily equal to m) involves solving up to $C_n^m$ times a series of p x p matrices, with p= min(m,n). The usual algorithm is the simplex , which is indeed not that hard but not up to high school level. You need a good command of linear algebra to really understand it, especially the corner cases. A very bright 10-12th grader could probably understand it and program it too. However, an efficient way to solve LP was not found until 1984, so LP is not that easy.
Now the article is about LP with integer constraints (i.e. integer programming or IP). This is much harder. There are no known efficient way to solve IP. The whole field of combinatorial optimisation is not taught until university after a good course on linear algebra and discrete mathematics.
People might teach 2D linear programming using geometrical means to some high school but they are decidedly NOT teaching fully blown arbitrary dimension LP with integer constraints like this article is using. Integer programming is an NP-hard problem. I teach this to university seniors.
This is partly true. Microsoft is funding actual research (like in math, computer vision, research OS, etc). This is long-term stuff, I'm sure that if MS is smart they can make a lot of money from that just like IBM and others did. If they are not they can have yet another Xerox PARC on their hand.
On the other hand Apple is funding lots of development but not research. I'm not aware of anything that Apple does that would be publishable in a peer-reviewed academic journal. This is a shame BTW.
Actually many doctors (for instance) work for free or at least very little money.
I write free software but I'm not paid for writing software in general, I have another, academic job. I would not work for a software company anymore: I have done it in the beginning of my career and it would pain me to recall the experience.
The software I wrote is useful to a small subset of people, who would probably not pay (much) for what I did, but I know my software is used a little over the place and I'm happy with this.
Writing it was fun, even writing the documentation. I spend on it the time I can afford, which is not much these days with family obligations.
Many people in F/OSS are like me. They contribute to the level of their talent, time and commitment, mostly for fun, but every little bit helps.
Actually Microsoft themselves want to drop XP dinosaur altogether. For the moment they can't completely, but XP is stuck in an evolutionary dead end. For a start its 32-bit only, and I already have 4GB in my bog-standard desktop PC, and its security record is not good. Noticed all the spam you are getting, routed from insecure XP boxes ? Thanks MS.
They had the problem before with win98, which was OK in its way, but at least XP was better in almost every respect. MS has the means to force people to change, but they are causing people to take a look at alternatives.
In this instance the advantage is to the installed base.
Nothing wrong with egos in this context. People can express themselves and compete on merit, this is good. If a distro does something good it will get used by others eventually.
Competitions amongst distros is not so much for the end-user. Of course many end-users may find things they like in each of them, but the main point is to foster best practices. Linux has many problems to solve, and if we only had Debian we would never have gotten a proper installer, for example. Usually each distribution will solve a particular problem well, and ideas can percolate to the top.
The tune recognition task is easier than it sounds (ha). In fact it's enough to hum the *contour* of the music, i.e. whether it simply goes up or down, for a couple of bars, ignoring the rythm even.
This way of indexing and recognising music is called the Parson Code and is quite effective.
This is better than most people think : string theory is a generalisation of *both* general relativity and quantum mechanics. It is consistent with both, therefore all QM and GR results can be found in ST (now M-theory). Now for ST to predict something *new* and therefore interesting is difficult, because it must happen at the QM-GR interface, i.e in very extreme environments like near black hole.
In addition the theory is difficult to handle mathematically. It would be good to be able to find things like Hawkings black-hole radiation from ST, but I'm not sure anyone has been able to pull it off yet, although the entropy of black holes have been derived from ST in some cases.
Sorry for the grammar, it was late, and english is not my first language.
Qt does a very good job of providing a cross-platform toolkit. It looks reasonably like the native toolkit on each platform (at least OS/X, Windows and Linux), while not going the X11 route, which is a tough thing to do.
The toolkit itself is rather pleasant to use, in spite of the nonstandard signal/slots.
For all the others there's various libraries and environments like cygwin that support various degrees of Posix for Windows. It is easy to find a fork() for windows, or indeed posix threads, etc.
Not that different really. Qt does a very cross-platform good job, and several other toolkits do a more-than-passable job as well such wxWidget or FLTK.
To me cross-platform is the only way to go, if you don't want your code to obsolete itself in no time at all.
Slashdot Mirror
Have they fixed the problem ?
Is it safe to buy a MBP now ?
It turns out that terrorism in western countries is a very rare thing, outside of a few hot areas like Spain's Basque area. This is very good, by the way.
Mining for rare event is extremely difficult. Bayes' s rule indicates that if in a database there are 0.01% actually suspicious events and your mining algorithms are extremely effective at 99% accuracy, then you still have an approximately 100:1 false positive ratio, which makes the mining still useless.
Your ideas sound good but are probably unworkable. Independent re-invention does not prove obviousness, as it happens all the time even for very difficult subjects (e.g. in maths, physics, not just engineering). The first to invent should get the credit as in research.
Your second point, I wholeheartedly agree with however. Patents are supposed to make something "patent", i.e. totally obvious to make with the knowledge of the patent. Some patents are like that but not all.
The third point is the opposite of reality. CSIRO are the one who worked their butt off to make wireless communication more reliable, effective and efficient, but they are not a manufacturing entity, they are a government research agency. They patented their discoveries as they should have. Even if Buffalo or others certainly made contributions into the manufacturing process, the basic scientific principles of their device is based on CSIRO's invention.
Nobody complains when IBM (say) patents a new silicon manufacturing process like SOI and makes wafer makes like Applied Materials pay licences. In effect the research arm of IBM makes billions a year from patent licences. People know that IBM has legal werewithals and don't think they can get away with cheating with IBM's inventions.
Here I'm pretty sure the Dells, HPs, Netgears of this world knew they were infringing but thought "who is this CSIRO thing? an Aussie gov. body? come on, can they even patent things in the US? can they even prosecure?" and tried to cheat knowing full well that at worse they might be liable for some fees down the line and might be able to settle for cheap somewhere down the line. They probably thought it would be a nice way for their legal department to make them some money for a change.
If patents like these are invalidated this is pretty much closing the lid on a whole body of strategies for applied research to fund itself through licences and commercial agreements. CSIRO is not a submarine patent shop. They employ around 6000 people and do very fine research in all areas of science. If they made a fine invention like it sounds they have, way in advance of everybody else, they should get a fine reward. This is the heart of the patent system.
If this doesn't work then the patent system is not only ineffective, it is simply evil.
1- Photons have no rest mass but they have energy and momentum (or relativistic mass). They are affected by gravity as they follow geodesics in spacetime, which are defined by the local mass-energy.
They original experimental proof of general relativity was by Eddington who showed that the mass of the Sun bends light in the way predicted by GR.
2- Mass by itself is not sufficient to "trap" light. A given mass contained within a given volume is, however, i.e a given density. A sufficient density creates an event horizon. GR predicts that all mass hidden within a event horizon must end up in a singularity.
Nothing fundamental as far as we know, but we are not there yet (as far as I know, at leat :-)
Even if someone proved that P=NP and exhibited a P algorithm to solve all the NP problems, it may not be efficient in classical terms. If the complexity of that algorithm is O(N^1000) say then we are still better off using the meta-heuristics that are in used right now to solve them approximately.
FYI there are efficient approximate solutions to the TSP that are guaranteed to come within a few percent of the true optimum. A P=NP proof may not be worth that much in real dollars.
No. Remember Knuth's aphorism :
Linear programming with n variables and m constraints (n not necessarily equal to m) involves solving up to $C_n^m$ times a series of p x p matrices, with p= min(m,n). The usual algorithm is the simplex , which is indeed not that hard but not up to high school level. You need a good command of linear algebra to really understand it, especially the corner cases. A very bright 10-12th grader could probably understand it and program it too. However, an efficient way to solve LP was not found until 1984, so LP is not that easy.
Now the article is about LP with integer constraints (i.e. integer programming or IP). This is much harder. There are no known efficient way to solve IP. The whole field of combinatorial optimisation is not taught until university after a good course on linear algebra and discrete mathematics.
People might teach 2D linear programming using geometrical means to some high school but they are decidedly NOT teaching fully blown arbitrary dimension LP with integer constraints like this article is using. Integer programming is an NP-hard problem. I teach this to university seniors.
This is partly true. Microsoft is funding actual research (like in math, computer vision, research OS, etc). This is long-term stuff, I'm sure that if MS is smart they can make a lot of money from that just like IBM and others did. If they are not they can have yet another Xerox PARC on their hand.
On the other hand Apple is funding lots of development but not research. I'm not aware of anything that Apple does that would be publishable in a peer-reviewed academic journal. This is a shame BTW.
Actually many doctors (for instance) work for free or at least very little money.
I write free software but I'm not paid for writing software in general, I have another, academic job. I would not work for a software company anymore: I have done it in the beginning of my career and it would pain me to recall the experience.
The software I wrote is useful to a small subset of people, who would probably not pay (much) for what I did, but I know my software is used a little over the place and I'm happy with this.
Writing it was fun, even writing the documentation. I spend on it the time I can afford, which is not much these days with family obligations.
Many people in F/OSS are like me. They contribute to the level of their talent, time and commitment, mostly for fun, but every little bit helps.
Eradicate the geek culture and convert everyone to business types? To sell what? In order to sue each other for fun and profit?
You don't have to like the geek culture, but what are you doing on slashdot then?
Yes, microsoft/PCs definitely lost marketshare to Apple recently.
Actually Microsoft themselves want to drop XP dinosaur altogether. For the moment they can't completely, but XP is stuck in an evolutionary dead end. For a start its 32-bit only, and I already have 4GB in my bog-standard desktop PC, and its security record is not good. Noticed all the spam you are getting, routed from insecure XP boxes ? Thanks MS.
They had the problem before with win98, which was OK in its way, but at least XP was better in almost every respect. MS has the means to force people to change, but they are causing people to take a look at alternatives.
In this instance the advantage is to the installed base.
Nothing wrong with egos in this context. People can express themselves and compete on merit, this is good. If a distro does something good it will get used by others eventually.
Competitions amongst distros is not so much for the end-user. Of course many end-users may find things they like in each of them, but the main point is to foster best practices. Linux has many problems to solve, and if we only had Debian we would never have gotten a proper installer, for example. Usually each distribution will solve a particular problem well, and ideas can percolate to the top.
The tune recognition task is easier than it sounds (ha). In fact it's enough to hum the *contour* of the music, i.e. whether it simply goes up or down, for a couple of bars, ignoring the rythm even.
This way of indexing and recognising music is called the Parson Code and is quite effective.
In fact the most likely explanation is a typo. 50% perhaps ?
oh, I don't know : Afghanistan, Iraq, Iran, France even :-)
Ruthless exploitation at both ends is the big deal.
This is better than most people think : string theory is a generalisation of *both* general relativity and quantum mechanics. It is consistent with both, therefore all QM and GR results can be found in ST (now M-theory). Now for ST to predict something *new* and therefore interesting is difficult, because it must happen at the QM-GR interface, i.e in very extreme environments like near black hole.
In addition the theory is difficult to handle mathematically. It would be good to be able to find things like Hawkings black-hole radiation from ST, but I'm not sure anyone has been able to pull it off yet, although the entropy of black holes have been derived from ST in some cases.
Sorry for the grammar, it was late, and english is not my first language.
Qt does a very good job of providing a cross-platform toolkit. It looks reasonably like the native toolkit on each platform (at least OS/X, Windows and Linux), while not going the X11 route, which is a tough thing to do.
The toolkit itself is rather pleasant to use, in spite of the nonstandard signal/slots.
Enterprise-level versions of Windows are fairly posix-compliant.
For all the others there's various libraries and environments like cygwin that support various degrees of Posix for Windows. It is easy to find a fork() for windows, or indeed posix threads, etc.
Chromium the rendering engine and associated libraries looks to be the backend of Chrome, the browser.
The latter is not available to any other platform than Windows.
Not that different really. Qt does a very cross-platform good job, and several other toolkits do a more-than-passable job as well such wxWidget or FLTK.
To me cross-platform is the only way to go, if you don't want your code to obsolete itself in no time at all.