You just spent $130 for the privilege of learning a proprietary system. Because you will now invest months and years in learning that system, you'll then basically be forced to spend thousands of dollars on the full version once you graduate, since you will already have sunk the cost of learning something.
Are there open source alternatives? You bet. No, not quite the same bundle of functionality, but overall better: Maxima (symbolic math+functional programming), OCAML and Haskell (functional programming), R (graphics, interactive numerical programming), Python (graphics, 3D visualization, interactive numerical programming), and many others.
There are lots of tools like Mathematica. Maxima will do symbolic math and general programming. OCAML, Haskell, and a bunch of other functional language give you the functional programming part of Mathematica. R is more numerically and statistically oriented. Python with VTK and several other packages is more than competitive for numerical and visualization applications.
You can't get exactly Mathematica, but you can get better functionality overall with different open source packages.
But when you have to spend 75% of your time reading websites and manuals and going back and forth to websites and trying to figure out the terminal, and... Well, it's frustrating. Too frustrating.WindowsXP makes things easier for the average, not so bright computer user.
You spent years getting familiar with Windows. You can't expect to pick up Linux in a day. This says nothing about the relative quality or utility of the different OSes.
In fact, learning Linux probably is initially harder than learning Windows. On the other hand, learning Linux is probably a more valuable skill: you learn to use software that doesn't change every year. And once you understand the command line tools and scripting, you can do really amazing things very quickly.
The biggest problem I had was KDE or Gnome? But then I started using it...
I think the answer is: it doesn't matter. Learn to use LaTeX, Emacs, xterm, and the standard POSIX tools. Learn Python or Perl. And if you are an "economist in training", learn R (for data analysis).
You might think of spectrum roughly as a rectangle. Currently, we allocate "vertical strips" (frequency bands) in this rectangle through the FCC. With UWB, you allocate "horizontal strips" (sequences) through this rectangle through a haphazard process. (More traditional spread spectrum methods occupy smaller, bounded rectangles, spreading across a limited range of frequencies).
Either way, the capacity is limited, and either way, as you allocate resources to one way of transmitting data, you create interference for the other way. In small numbers, UWB interference will be largely unnoticeable, but if it caches on as widely as its proponents claim it will, it may drown out traditional frequency based allocations. You get a situation roughly analogous to the interference problems between Bluetooth and 802.11.
I think UWB is basically an attempt to circumvent current frequency-based allocation schemes, and to replace our cheap, non-proprietary frequency-based infrastructure with a proprietary, patented, and more costly sequence-based scheme. Once millions of these devices are deployed and we are starting to see interference, manufacturers will whine and complain that they can't be banned anymore because the economic cost is too high.
In short, let's not fall into that trap. We already have spread spectrum technologies that are more sophisticated: they limited transmissions not only by sequence but also to a given, allocated frequency band. That works fine. We don't need UWB, and adopting UWB now would probably lead to bigger problems down the road.
Of course, I shouldn't have to mention that both QPE and Konq/e are fully-fledged GPL'ed projects
Qt costs lots of money if you want to write commercial apps for it. Fortunately, there is a cheaper and better choice for writing embedded apps: just use FLTK. It's much more compact than Qt, you can use it freely for writing commercial apps, and it runs on Linux.
Nobody will ever compete with Microsoft Office head on--not because people can't technically produce anything better, but because Microsoft sets the agenda. Lotus SmartSuite probably had the best shot, and it failed even in situations where people got it for free.
But Office is a cumbersome dinosaur. Office-based business applications are flaky, difficult to use, and unreliable. Office can be dethroned.
What we really need to do is to figure out how to get the same jobs done with something that is compellingly better: software that enables web-based collaboration, software consisting of small, specialized, downloadable applications, software that's much easier than Office to extend and program, even for non-programmers.
I think this is another existence proof for a nice, non-Microsoft OS on the iPaq. I'm not quite sure why I would want to run it, though: Linux for handhelds is quite functional (even if its icons are not as nice), and it is free, open, and standard.
The biggest problem I have had with running non-WinCE operating systems on the iPaq is the installation, which is a very laborious and slow process that takes hours to download stuff over the serial line. What is really needed is the ability to overlay a new OS from Flash and/or to install a new OS by clicking on an application in Flash memory. Or, of course, Compaq might finally preinstall Linux on the iPaq; even HP will be shipping a Linux PDA.
As a physics professor, you should be well acquainted with the concept that you don't need millions of dollars to be happy (or successful), which was my point.
Beyond that, I'm sorry you didn't cut it as a teacher and weren't interested in engineering. However, those professions are every bit as important and challenging as being a tenured physics professor, and they probably contribute quite a bit more to our society.
Now, those same geeks are complaining ON THOSE VERY SAME VENTURE CAPITALISTS?!?!
Nope, there are, in fact, different kinds of geeks. There are the clueless geeks who think that they can make a quick buck by developing big sounding technology with expensive tools and on impossible deadlines. Sometimes they get lucky, but more often, they get f*cked.
Clueful geeks never participated in this game. They work steady jobs, save money, run small consulting business, and generally are having a much better time. If a VC contacts them, they just politely refuse.
You don't need millions of dollars in order to be happy. Engineering and software development is a decent way to make a living, and you can be quite well off without ruining your health on startup dreams.
Oh, as for the open source startups, the VCs that invested in them were fools. But if VCs are going to waste their money, they might as well waste it on something that contributes to the common good. I suspect many engineers working for such companies weren't dreaming on getting rich but just liked the idea of creating open source software fulltime. (Support of open source by companies like Sun and IBM, on the other hand, makes business sense for them.)
Bankruptcy or GPL'ed code? If it were my company, I would be choosing the former.
A small software company with that kind of attitude doesn't stand a chance. Software buyers that make a commitment to some technology are looking at long-term availability and maintenance, availability and maintenance that may survive the shaking little company that produced the software in the first place. The LGPL and GPL ensure long-term availability. Few other licenses do.
Frankly, receiving anything at all above that would be downright generous. Don't expect companies to be generous, that's not their job.
I expect companies to live up to their commitments, whether implied or contractual. If they break contractual commitments, they get taken to court. If they break implied commitments, they get criticized. I consider changing a license from the GPL to the SPL breaking an implied commitment.
Why do people think commercial companies contribute to open source? It's because they develop tools to keep their business running, but they don't have any interest of getting into the software business. The GPL addresses this.
The SPL removes most of the incentive to contribute. Why would I want to contribute software to a company, just to be forced to pay for my own contribution forever more? Sure, it may give me the functionality I need in the short term, but in the long term, I would have been better off at contributing to an open source project.
Doesn't sound like a bad licensing scheme to me at all.
It wouldn't have been a bad licensing scheme if they had used it from the beginning. But changing terms after benefitting from bug reports and suggestions for enhancements from end users who thought they were contributing to a GPL'ed project is wrong.
If we're profiting from their work I see this as only fair.
People suggested bug fixes and suggested enhancements assuming they were working on an GPL'ed product and then Sistina changed the terms. That is the part that isn't fair.
I think looking at percentages isn't very meaningful. There are a huge number of non-profit, alternative, and diverse voices on the net. You can start publishing to the world at essentially no cost. That's a big change from what we had just 10 years ago, and it's very different from the traditional media.
Of course, it was inevitable that any successful medium causes advertising, PR, and corporate interests to move in. But while they may dominate percentage-wise, the fact is that in absolute numbers, we have access to much more diverse content than ever before.
International organizations can both help and hurt international freedoms. The EU and UN have done a lot of good things for human rights, for example. Overall, the US would likely also be quite a bit better off if it stopped obstructing international initiatives on the environment, human rights, etc.
Internationalization is coming--driven by technology and mobility. The choice we have is to try and make it work for the people. Of course, that's an uphill struggle: wealthy private institutions have a leg up. But as long as voter participation in the US is still somewhere around 50%, I don't think anybody in the US has a right to complain about the politicians or policies they are getting.
The problems with the Qt license as are serious for ISVs today as they were then:
Another problem at the time was the licensing of QT from TrollTech. The LGPL license of GTK+ was more attractive as it offers our ISVs a cheaper alternative to a costly toolkit.
TrollTech now allows redistribution under the GPL, but ISVs still need to pay a steep licensing fee. What possible interest would Sun have to commit their customers to writing to a toolkit that costs them a lot of money and that Sun has no control over? The KDE/Qt licensing is worse than the CDE/Motif licensing from the point of view of ISVs.
You see, in the real world, the license issues involved in Gtk+ vs. Qt aren't about "freedom", they are about economics. Qt just doesn't cut it there--if people had wanted to pay that kind of money for a toolkit, they didn't have to wait for Troll Tech or KDE to come around.
It turns out that KDE suffered because of the well-known failures of C++ standardization.
Don't shift blame. This isn't a "failure", the kind of standardization you are talking about was never promised for C++. The KDE project should have known that. In fact, overall, ANSI C++ has probably turned out to be a much better language design than one might have thought a few years ago, but as the basis for a GUI toolkit, it is still a third rate choice.
The KDE project made two seriously bad choices at the start (choosing Qt and C++), and it doesn't matter how nice the KDE desktop becomes (and it is pretty nice) those issues will just not go away.
AMD can fix this easily: add a 1:2 divider circuit on their clock lines, and voila, instant 2.8GHz Athlon. And if that's too blatant, they can just call it their "hyperthreading NOP engine"; that even doubles the MIPS.
GNU, Debian, RedHat, Gnome, KDE, BSD, Linux, and lots of other open source software were created without a SuSE-like business-model. Open source software predates SuSE by many decades. So, evidently, open source software doesn't need SuSE-like attempts at generating income for its creation. In fact, I would say that if open source software could only succeed by becoming proprietary software, as your argument suggests, open source would have failed by definition.
plea bargain and police state
on
Sklyarov Indicted
·
· Score: 2, Insightful
You know that this is going to be settled: prosecutors know that going through with the trial is risky, but they can sure intimidate Sklyarov with 25 years in jail (he wouldn't survive it, so it might as well be the death penalty) and a $2.5M fine (try paying that back in rubles).
But, hey, why stop there? If we have such unbounded trust in our legal officials, why do we bother with laws at all? Why not give police complete freedom to pick up people that are engaged in unsavory activities according to their judgement, give prosecutors complete freedom to craft punishments, and give judges complete freedom to impose whatever they see fit for whatever action they see as illegal or detrimental to society? Given the penalties currently on the books, we might as well.
What this comes down to is that the US is increasingly not a country of laws but a country run by the law enforcement and justice system. The distinction is profound and it is very important to keep it in mind. In fact, we have a name for the latter situation: it's called a "police state".
This will make countries even more trigger happy, since they have even less time to react to an incoming missile.
this compounds problems of proprietary formats
on
Quicktime In Linux
·
· Score: 2, Insightful
Proprietary formats are bad: they restrict fair use rights unreasonably, and they mean that data simply becomes inaccessible in a few years, when the companies that created those formats have moved on to the next thing or gone out of business. Giving companies that kind of control over content threatens the foundations of our academic, social, and political institutions.
Using the kinds of workarounds CodeWeavers is making available only perpetuates such formats. You should instead ask web sites and content providers to use open, documented formats. Even if the open formats are encumbered by patents (like MPEG), that is still a better deal than using something proprietary; patents eventually expire, but undocumented proprietary formats never become open, they become obsolete and forgotten.
Besides, don't fool yourself for a moment: Microsoft and Apple will only allow this sort of thing to go on if they see it as being either useless or in their advantage. Otherwise, they have plenty of legal and technical means for stopping it.
X-windows is still a hole and a barrier to the pervasive Linux desktop. It's still around because it's a virus, like Unix itself, that had corporate support in its infancy.
You got your history mixed up. Sun, the predominant workstation manufacturer when X11 came out, was fighting X11 tooth and nail, first with SunView (a direct-to-framebuffer system like Windows), then with NeWS (a Postscript-based system). Both failed miserably, as customers overwhelmingly preferred to install MIT X11 themselves, and Sun eventually gave up and started shipping X11. Several other vendors tried their own window systems as well and failed.
X11's network transparency and separation of functionality and policy made it a far more versatile and useful system than any of the alternatives; while these design choices clearly do not result in the most efficient rendering (although X11 is still very good) and clearly do not promote a consistent user experience, the overall advantage is still on the X11 side. X11 has turned out to be a reliable foundation on which to build a wide variety of applications and systems.
and point them to Berlin
You can clearly improve on X11 in many ways. It's too bad that Berlin is looking to the past and repeating the errors of previous systems, rather than coming up with a new, innovative design.
SuSE seems like a nice distro, but I got the feeling that they are trying much harder than, say, RedHat, to force people to buy the CDs. I think it would be problematic for Linux if SuSE became the predominant distribution.
Slashdot Mirror
Are there open source alternatives? You bet. No, not quite the same bundle of functionality, but overall better: Maxima (symbolic math+functional programming), OCAML and Haskell (functional programming), R (graphics, interactive numerical programming), Python (graphics, 3D visualization, interactive numerical programming), and many others.
You can't get exactly Mathematica, but you can get better functionality overall with different open source packages.
You spent years getting familiar with Windows. You can't expect to pick up Linux in a day. This says nothing about the relative quality or utility of the different OSes.
In fact, learning Linux probably is initially harder than learning Windows. On the other hand, learning Linux is probably a more valuable skill: you learn to use software that doesn't change every year. And once you understand the command line tools and scripting, you can do really amazing things very quickly.
The biggest problem I had was KDE or Gnome? But then I started using it...
I think the answer is: it doesn't matter. Learn to use LaTeX, Emacs, xterm, and the standard POSIX tools. Learn Python or Perl. And if you are an "economist in training", learn R (for data analysis).
Either way, the capacity is limited, and either way, as you allocate resources to one way of transmitting data, you create interference for the other way. In small numbers, UWB interference will be largely unnoticeable, but if it caches on as widely as its proponents claim it will, it may drown out traditional frequency based allocations. You get a situation roughly analogous to the interference problems between Bluetooth and 802.11.
I think UWB is basically an attempt to circumvent current frequency-based allocation schemes, and to replace our cheap, non-proprietary frequency-based infrastructure with a proprietary, patented, and more costly sequence-based scheme. Once millions of these devices are deployed and we are starting to see interference, manufacturers will whine and complain that they can't be banned anymore because the economic cost is too high.
In short, let's not fall into that trap. We already have spread spectrum technologies that are more sophisticated: they limited transmissions not only by sequence but also to a given, allocated frequency band. That works fine. We don't need UWB, and adopting UWB now would probably lead to bigger problems down the road.
Qt costs lots of money if you want to write commercial apps for it. Fortunately, there is a cheaper and better choice for writing embedded apps: just use FLTK. It's much more compact than Qt, you can use it freely for writing commercial apps, and it runs on Linux.
But Office is a cumbersome dinosaur. Office-based business applications are flaky, difficult to use, and unreliable. Office can be dethroned.
What we really need to do is to figure out how to get the same jobs done with something that is compellingly better: software that enables web-based collaboration, software consisting of small, specialized, downloadable applications, software that's much easier than Office to extend and program, even for non-programmers.
The biggest problem I have had with running non-WinCE operating systems on the iPaq is the installation, which is a very laborious and slow process that takes hours to download stuff over the serial line. What is really needed is the ability to overlay a new OS from Flash and/or to install a new OS by clicking on an application in Flash memory. Or, of course, Compaq might finally preinstall Linux on the iPaq; even HP will be shipping a Linux PDA.
Beyond that, I'm sorry you didn't cut it as a teacher and weren't interested in engineering. However, those professions are every bit as important and challenging as being a tenured physics professor, and they probably contribute quite a bit more to our society.
Nope, there are, in fact, different kinds of geeks. There are the clueless geeks who think that they can make a quick buck by developing big sounding technology with expensive tools and on impossible deadlines. Sometimes they get lucky, but more often, they get f*cked.
Clueful geeks never participated in this game. They work steady jobs, save money, run small consulting business, and generally are having a much better time. If a VC contacts them, they just politely refuse.
You don't need millions of dollars in order to be happy. Engineering and software development is a decent way to make a living, and you can be quite well off without ruining your health on startup dreams.
Oh, as for the open source startups, the VCs that invested in them were fools. But if VCs are going to waste their money, they might as well waste it on something that contributes to the common good. I suspect many engineers working for such companies weren't dreaming on getting rich but just liked the idea of creating open source software fulltime. (Support of open source by companies like Sun and IBM, on the other hand, makes business sense for them.)
This seems like the ideal language for DeCSS--even a liberal arts educated judge must see that such a "program" constitutes some form of literature.
A small software company with that kind of attitude doesn't stand a chance. Software buyers that make a commitment to some technology are looking at long-term availability and maintenance, availability and maintenance that may survive the shaking little company that produced the software in the first place. The LGPL and GPL ensure long-term availability. Few other licenses do.
I expect companies to live up to their commitments, whether implied or contractual. If they break contractual commitments, they get taken to court. If they break implied commitments, they get criticized. I consider changing a license from the GPL to the SPL breaking an implied commitment.
The SPL removes most of the incentive to contribute. Why would I want to contribute software to a company, just to be forced to pay for my own contribution forever more? Sure, it may give me the functionality I need in the short term, but in the long term, I would have been better off at contributing to an open source project.
It wouldn't have been a bad licensing scheme if they had used it from the beginning. But changing terms after benefitting from bug reports and suggestions for enhancements from end users who thought they were contributing to a GPL'ed project is wrong.
People suggested bug fixes and suggested enhancements assuming they were working on an GPL'ed product and then Sistina changed the terms. That is the part that isn't fair.
Of course, it was inevitable that any successful medium causes advertising, PR, and corporate interests to move in. But while they may dominate percentage-wise, the fact is that in absolute numbers, we have access to much more diverse content than ever before.
Internationalization is coming--driven by technology and mobility. The choice we have is to try and make it work for the people. Of course, that's an uphill struggle: wealthy private institutions have a leg up. But as long as voter participation in the US is still somewhere around 50%, I don't think anybody in the US has a right to complain about the politicians or policies they are getting.
TrollTech now allows redistribution under the GPL, but ISVs still need to pay a steep licensing fee. What possible interest would Sun have to commit their customers to writing to a toolkit that costs them a lot of money and that Sun has no control over? The KDE/Qt licensing is worse than the CDE/Motif licensing from the point of view of ISVs.
You see, in the real world, the license issues involved in Gtk+ vs. Qt aren't about "freedom", they are about economics. Qt just doesn't cut it there--if people had wanted to pay that kind of money for a toolkit, they didn't have to wait for Troll Tech or KDE to come around.
Don't shift blame. This isn't a "failure", the kind of standardization you are talking about was never promised for C++. The KDE project should have known that. In fact, overall, ANSI C++ has probably turned out to be a much better language design than one might have thought a few years ago, but as the basis for a GUI toolkit, it is still a third rate choice.
The KDE project made two seriously bad choices at the start (choosing Qt and C++), and it doesn't matter how nice the KDE desktop becomes (and it is pretty nice) those issues will just not go away.
AMD can fix this easily: add a 1:2 divider circuit on their clock lines, and voila, instant 2.8GHz Athlon. And if that's too blatant, they can just call it their "hyperthreading NOP engine"; that even doubles the MIPS.
GNU, Debian, RedHat, Gnome, KDE, BSD, Linux, and lots of other open source software were created without a SuSE-like business-model. Open source software predates SuSE by many decades. So, evidently, open source software doesn't need SuSE-like attempts at generating income for its creation. In fact, I would say that if open source software could only succeed by becoming proprietary software, as your argument suggests, open source would have failed by definition.
But, hey, why stop there? If we have such unbounded trust in our legal officials, why do we bother with laws at all? Why not give police complete freedom to pick up people that are engaged in unsavory activities according to their judgement, give prosecutors complete freedom to craft punishments, and give judges complete freedom to impose whatever they see fit for whatever action they see as illegal or detrimental to society? Given the penalties currently on the books, we might as well.
What this comes down to is that the US is increasingly not a country of laws but a country run by the law enforcement and justice system. The distinction is profound and it is very important to keep it in mind. In fact, we have a name for the latter situation: it's called a "police state".
This will make countries even more trigger happy, since they have even less time to react to an incoming missile.
Using the kinds of workarounds CodeWeavers is making available only perpetuates such formats. You should instead ask web sites and content providers to use open, documented formats. Even if the open formats are encumbered by patents (like MPEG), that is still a better deal than using something proprietary; patents eventually expire, but undocumented proprietary formats never become open, they become obsolete and forgotten.
Besides, don't fool yourself for a moment: Microsoft and Apple will only allow this sort of thing to go on if they see it as being either useless or in their advantage. Otherwise, they have plenty of legal and technical means for stopping it.
You got your history mixed up. Sun, the predominant workstation manufacturer when X11 came out, was fighting X11 tooth and nail, first with SunView (a direct-to-framebuffer system like Windows), then with NeWS (a Postscript-based system). Both failed miserably, as customers overwhelmingly preferred to install MIT X11 themselves, and Sun eventually gave up and started shipping X11. Several other vendors tried their own window systems as well and failed.
X11's network transparency and separation of functionality and policy made it a far more versatile and useful system than any of the alternatives; while these design choices clearly do not result in the most efficient rendering (although X11 is still very good) and clearly do not promote a consistent user experience, the overall advantage is still on the X11 side. X11 has turned out to be a reliable foundation on which to build a wide variety of applications and systems.
and point them to Berlin
You can clearly improve on X11 in many ways. It's too bad that Berlin is looking to the past and repeating the errors of previous systems, rather than coming up with a new, innovative design.
SuSE seems like a nice distro, but I got the feeling that they are trying much harder than, say, RedHat, to force people to buy the CDs. I think it would be problematic for Linux if SuSE became the predominant distribution.