The part that denies you complete control over your own work by requiring distributed modifications and derivations also be licensed under the GPL.
If you don't like the GPL conditions for modifying and redistributing GPL'ed software, just don't modify GPL'ed software. After all, that's the same approach you take with any commercial software (where the conditions are: you modify and redistribute it and you have to pay stiff penalties).
Besides, your understanding of the GPL itself is incorrect. You can modify GPL'ed software to your heart's content. You can even distribute your own additions to GPL'ed software in whatever form you like under whatever license you like (see Section 5 of the GPL2).
Only in the case that you patch or link your modifications into existing GPL'ed software and redistribute the result are you obligated to put your code under the GPL. What triggers that is that you are redistributing someone else's code. Don't redistribute other people's code and you don't have any obligations or legal problems. That's the same deal you get from Microsoft or Sun.
In addition, while companies like Microsoft or Sun claim ownership of your code merely because you looked at theirs, GPL'ed software never makes such claims.
I'm tired of theses "this can save lives" arguments. Fascism "can save lives", too, but most people seem to agree that it just isn't worth it. Well, actually, they agree in the abstract, at least, but each individual step towards it gets justified with your kind of argument.
I think being able to track one's own location via GPS or cell phones is really swell. But when the police or employers can do it as a matter of course, then it fundamentally changes the kind of society we have.
Microsoft, a business, gives you their software under the following conditions:
You pay us money.
You can use the software on one machine.
GPL'ed software dispenses with the "you pay us money" part. Then, it gives you the following additional rights:
You can alter the software.
You can recompile the software.
You can redistribute the software under the same license that you got it.
What's "anti-business" about distributing a better product at a lower cost, giving more functionality to the end user? Is Macintosh "anti-business" because it's better than Windows? Is "Snap Cola" "anti-business" because it's cheaper than "Coca Cola"?
If you don't like the GPL or GPL'ed software, you have the same choice as with any other product: just don't use it. Well, actually, you have more of a choice, because while you may not be able to avoid using Microsoft software, you can definitely get through life without ever touching GPL'ed software.
One of the reasons they're doing this is that students often tend to use multiple extensions on a single outlet, which is the second leading cause of fire deaths, according to this.
Extension cords and power strips are a fire hazard for two reasons. First, people walk over them, where they get frayed and create shorts. Second, they make it easier to overload the circuit, which means that you are then depending on the circuit breaker to work properly.
But by themselves, extension cords and power strips don't cause fires. They are not substantially different from the wiring that runs in your walls. If you only plug your cell phone, laptop, PDA charger, and stereo into them and don't stress them mechanically, they should be no more likely to cause a fire than if you plugged all that into an outlet. In fact, many power strips have their own built-in circuit breakers, which give you extra protection. And if you are really worried, you can also get a circuit breaker that plugs into the wall socket, giving you extra protection against shorts in the cord itself.
Holographic memory turns out not to be a good match to today's computers: the data comes out too fast to be useful with current systems and writing takes too long.
Optical computers have been used in some specialty applications (e.g., optical correlators), but they are not usually cost effective.
As for a flip of the earth's magnetic field, that's probably not strong enough to damage any media, but it might result in higher radiation levels, with all sorts of unpleasant consequences.
It's a nice idea to encourage people to claim their piece of the settlement, and as much as I like seeing it happen to Microsoft, it really can't be done in a fair and legal way.
The settlement is whatever the parties agree on. They could have agreed to give the money to chain smokers, or to donate to endangered ducks, or to burn a billion dollars to heat orphanages. It doesn't matter matter whether you consider the distribution fair, what matters is what they agreed to.
And what they agreed to is that they owe a billion dollars, a light penalty anyway compared to the harm they have caused. But now they are trying to delay and wriggle out of that, too. That is definitely not fair.
Hostile, Yes indeed! But the major problem and the reason it's still so very very hostile is that we've never commited to being there.
We have had a pretty much unbroken presence in space since the 1950's. Not, it hasn't been a manned presence, but so what? Putting people into space is unnecessarily risky and has little reward.
Reliable, we are back to people risking their hides.
No, we are not. Unlike the age of exploration, we don't have to go ourselves anymore. We can simply lean back while ever cheaper robotic probes explore our solar system (and possibly beyond).
Buyers don't lack information with respect to music: if word of mouth and reviews weren't enough, you can now preview CDs on sites like Amazon. With that, most of the rest of the argument seems to collapse, and there are other holes in it as well.
Just install a copy of VNC and/or x0rfbserver on their desktop. You can connect to their machines and help them out while you are on the phone. You can also go web browsing together, etc.
(Be sure to use security correctly: definitely set a password. You can also use stunnel or ssh tunneling for additional security.)
This isn't so much of a root vulnerability as a default configuration that trusts the integrity of the local network services.
That is a root vulnerability. You could perhaps trust LANs 20 years ago, you absolutely cannot trust them today, and any vendor that ships software that, by default, trusts the LAN is shipping software with severe security problems.
Running software on untrusted hardware can never really be secure. If the school wants to do this sort of thing, they need to provide the machines.
They could either buy a set of laptops specifically for exams, or they could buy some low-end machine whose primary function is word processing. Examples are the Dana AlphaSmart and the LaserPC. A simple cold boot will bring them back into a known configuration. Buying a few dozen of those may even be cheaper than a site license for the "Extegrity" software.
On the one hand, current DVDs really don't quite have the quality of some of the original film material, on the other hand, the technology to create HD-DVDs cost effectively simply wasn't there when DVDs first came out. So, this upgrade is quite justifiable, IMO. And an HD movie is still too large to be downloaded conveniently.
Take a look at your DVDs, though, and freeze some frames: for a lot of movies, it probably doesn't make much sense to get HD-DVDs because the quality of the original isn't all that high to begin with. You can already see plenty of film grain, fuzziness, and other film-related artifacts even with regular DVD resolutions and compression. Motion picture film just isn't all that great.
Attempts to create new audio disc formats, on the other hand, don't make much sense for the user: audio CDs really capture audio better than most stereos can reproduce it and most people can perceive it. Furthermore, the next "format" for audio is likely going to be the Internet. So, new audio formats are just an attempt at making lots of money with no real benefit to the customer.
Nothing particularly sinister. The "hooks" they have is probably caller ID on the modems. If he connected via DSL, they have the port he connected from (via the IP or MAC address). AOL is probably completely free to give out that information, with no need for wiretap authorization.
Normal 1U rack mounts or rack-mounted blade servers will work for consolidation: they can share power and they can be dynamically reassigned to different tasks without any need for virtualization.
NFS/NIS was an incompetent implementation of a feature useful to lots of people. That's why lots of people ended up using it even though it had no security and would corrupt their data with regularity. (NFSv4 may, finally, address these issues.)
Solaris 10 sounds like a competent implementation of features useful to only a tiny fraction of the market, if they are useful to anybody at all. Sun can go head to head with IBM with that, wooing a market consisting of maybe a few thousand customers. For the rest of us, I just don't think it matters.
I suspect RedHat doesn't support it (assuming they don't) because there isn't much demand for it. If you want multiple Linux instances, you just buy multiple machines.
Sun virtualization is rather bizarre from an economic point of view: first you pay a premium to get all your hardware in one huge "Enterprise Server", and then you virtualize it out again.
If that's the case, then why not run Windows NT on everything?
Because Windows NT is very expensive, because its APIs keep changing, and because it requires way too much manual effort to install and maintain.
At least an order of magnitude more, I'd say. Or maybe they have some other method of making sure their code is stable, such as special compilers or automated code verification... Whatever it is that they do, there is no denying that Solaris is far more stable than Linux.
Sun's systems have had really serious bugs in the past (memory leaks, data corruption, buggy code generation in the compilers, etc.), it's just that after years of bug reports from the fields, many of them have been fixed.
Whatever stability Solaris has, it's simply due to age, not any kind of special magical powers Sun has.
Don't tell me that nobody will ever need a ZB of storage. Maybe not tomorrow, but in 20 years? Yea.
You are absolutely right that reliability, performance, storage capacity, and service matter. But monolithic systems like Solaris running on monolithic hardware like Sun Enterprise servers won't be able to scale to those sizes.
Your problem [1] is that you're too short sighted; You don't see the big picture.
No, your problem (and Sun's problem) is that you are firmly stuck in a 1960's mindset.
[2] Not every Linux user or advocate is a fanatic. I use the term to refer to the more rabid zealots.
And I'm not a Linux fanatic either. The Linux kernel is pretty mediocre as far as kernels go. I think BSD or Mach or any of a number of other mediocre kernels would unseat Sun as well.
You see, it's not about whether Linux is "better" than Solaris, it's about the fact that Sun is just barking up the wrong tree: it just doesn't matter much what operating system you use to run your individual nodes, as long as it's cheap and simple. What matters is the application and storage software that runs on top of it and that can deal with both the distributed nature of the system and inevitable node failures.
Linux is a good OS and I am no stranger to it whatsoever, but it has a long way to go to catch up to Solaris.
Some races are not worth running in. This is one of them. Let Sun pursue pipe dreams of "Zettabyte disk drives", useless sequences-of-nines uptimes, and gigantic, overpriced shared memory machines, while humble little Linux forms the backbone of the real supercomputers and mainframes of the future: compute clusters built from unreliable, commodity machines.
but it isn't remotely close to being stable and reliable enough to play on the enterprise level. There's just too many features going in, unaudited, unverified.
And you think that Solaris is better audited and verified? I don't think so. Sun doesn't have the resources.
And what do you want anyway? If your enterprise setup is so flaky that an errant kernel or kernel security flaw compromises your operation, you have a problem, and it's not fixable by spending millions on gold-plated hardware and software.
You misunderstood the analogy. For Solaris vs. Linux, it's (unnecessary) scalability and robustness that is the key property, for the cars it's speed.
What I am saying is Linux is a nice car for daily commute, but Solaris is a better investment.
Well, let's try another analogy: do you keep your wallet in a bankvault or even a safe at home? Probably not, because it's too much hassle and not worth it.
Sun provides great support, hardware, compatibilty with past versions of it's software,
The SunOS/Solaris transition has been anything but smooth, and Sun sites have traditionally dreaded upgrading because it has been so disruptive. You do get decent support, but at a huge cost. And hardware compatibiliity? Well, it's a single-vendor solution--buy from a Linux hardware vendor and you'll get the same or a better experience with Linux than that.
Java, and more.
Yes, Solaris does have better support for Java than Linux. That tells you more about the fact that Java is a Sun-proprietary that Sun deploys according to their business interests than anything about Linux.
Features like this won't even make their way into Linux for another 5 years or so.
Let's hope so. If we are really lucky, most of them will never make it into Linux because they are useless bloat for almost all users.
Let Sun go for the really, really high end of million dollar hardware running specialized operating systems. I'm happy with Linux being the operating system for huge compute clusters made out of commodity hardware, and for that, Linux is more than enough already.
Solaris's problems are not in robustness or scalability--it already handles those very well. Trouble is, so do Linux and BSD and a lot of other systems. Arguably, not as well as Solaris, but well enough.
By analogy, sure, a Ferrari is a nice car, but for a daily commute, a Honda Civic is both cheaper and more practical, and it really doesn't matter that it doesn't go as fast as the Ferrari in theory. With software like Solaris 10, Sun is creating ever more expensive Ferraris.
I think it's a little more complicated than that. There are different kinds of innovation and different kinds of open source projects.
To me, X11 was quite innovative. Keep in mind that it came out a publicly financed research project as well.
Python and Perl, on the other hand, don't strike me as "innovative" at all: we have had scripting languages and interpreted languages like that since the 1960's, and both of them are architecturally still stuck in the dark ages. But, innovative or not, from a practical point of view, Python and Perl still beat any of the scripting solutions Apple or Microsoft offer.
Overall, OSS is a really mixed bag on innovation, but that's perfectly fine; it means that overall, there is a lot of innovation in OSS (often financed through public research grants). For Apple and Microsoft's product range, on the other hand, things seem simpler to me: for the most part, they aren't very innovative. What Apple and Microsoft do fairly well is graphic design, marketing, branding, developer relations, and sales. Good for building a business, but that's all.
That feature was originally present in a number of X11 window managers. But someone (HP, I believe) got a patent on it and people decided it wasn't worth fighting.
Maybe nowadays, HP doesn't care anymore or the patent has expired.
It's not clear to me that any of those window management hacks are all that useful anyway...
Slashdot Mirror
The part that denies you complete control over your own work by requiring distributed modifications and derivations also be licensed under the GPL.
If you don't like the GPL conditions for modifying and redistributing GPL'ed software, just don't modify GPL'ed software. After all, that's the same approach you take with any commercial software (where the conditions are: you modify and redistribute it and you have to pay stiff penalties).
Besides, your understanding of the GPL itself is incorrect. You can modify GPL'ed software to your heart's content. You can even distribute your own additions to GPL'ed software in whatever form you like under whatever license you like (see Section 5 of the GPL2).
Only in the case that you patch or link your modifications into existing GPL'ed software and redistribute the result are you obligated to put your code under the GPL. What triggers that is that you are redistributing someone else's code. Don't redistribute other people's code and you don't have any obligations or legal problems. That's the same deal you get from Microsoft or Sun.
In addition, while companies like Microsoft or Sun claim ownership of your code merely because you looked at theirs, GPL'ed software never makes such claims.
I'm tired of theses "this can save lives" arguments. Fascism "can save lives", too, but most people seem to agree that it just isn't worth it. Well, actually, they agree in the abstract, at least, but each individual step towards it gets justified with your kind of argument.
I think being able to track one's own location via GPS or cell phones is really swell. But when the police or employers can do it as a matter of course, then it fundamentally changes the kind of society we have.
GPL'ed software dispenses with the "you pay us money" part. Then, it gives you the following additional rights:
What's "anti-business" about distributing a better product at a lower cost, giving more functionality to the end user? Is Macintosh "anti-business" because it's better than Windows? Is "Snap Cola" "anti-business" because it's cheaper than "Coca Cola"?
If you don't like the GPL or GPL'ed software, you have the same choice as with any other product: just don't use it. Well, actually, you have more of a choice, because while you may not be able to avoid using Microsoft software, you can definitely get through life without ever touching GPL'ed software.
One of the reasons they're doing this is that students often tend to use multiple extensions on a single outlet, which is the second leading cause of fire deaths, according to this.
Extension cords and power strips are a fire hazard for two reasons. First, people walk over them, where they get frayed and create shorts. Second, they make it easier to overload the circuit, which means that you are then depending on the circuit breaker to work properly.
But by themselves, extension cords and power strips don't cause fires. They are not substantially different from the wiring that runs in your walls. If you only plug your cell phone, laptop, PDA charger, and stereo into them and don't stress them mechanically, they should be no more likely to cause a fire than if you plugged all that into an outlet. In fact, many power strips have their own built-in circuit breakers, which give you extra protection. And if you are really worried, you can also get a circuit breaker that plugs into the wall socket, giving you extra protection against shorts in the cord itself.
Holographic memory turns out not to be a good match to today's computers: the data comes out too fast to be useful with current systems and writing takes too long.
Optical computers have been used in some specialty applications (e.g., optical correlators), but they are not usually cost effective.
As for a flip of the earth's magnetic field, that's probably not strong enough to damage any media, but it might result in higher radiation levels, with all sorts of unpleasant consequences.
It's a nice idea to encourage people to claim their piece of the settlement, and as much as I like seeing it happen to Microsoft, it really can't be done in a fair and legal way.
The settlement is whatever the parties agree on. They could have agreed to give the money to chain smokers, or to donate to endangered ducks, or to burn a billion dollars to heat orphanages. It doesn't matter matter whether you consider the distribution fair, what matters is what they agreed to.
And what they agreed to is that they owe a billion dollars, a light penalty anyway compared to the harm they have caused. But now they are trying to delay and wriggle out of that, too. That is definitely not fair.
Hostile, Yes indeed! But the major problem and the reason it's still so very very hostile is that we've never commited to being there.
We have had a pretty much unbroken presence in space since the 1950's. Not, it hasn't been a manned presence, but so what? Putting people into space is unnecessarily risky and has little reward.
Reliable, we are back to people risking their hides.
No, we are not. Unlike the age of exploration, we don't have to go ourselves anymore. We can simply lean back while ever cheaper robotic probes explore our solar system (and possibly beyond).
Buyers don't lack information with respect to music: if word of mouth and reviews weren't enough, you can now preview CDs on sites like Amazon. With that, most of the rest of the argument seems to collapse, and there are other holes in it as well.
Just install a copy of VNC and/or x0rfbserver on their desktop. You can connect to their machines and help them out while you are on the phone. You can also go web browsing together, etc.
(Be sure to use security correctly: definitely set a password. You can also use stunnel or ssh tunneling for additional security.)
This isn't so much of a root vulnerability as a default configuration that trusts the integrity of the local network services.
That is a root vulnerability. You could perhaps trust LANs 20 years ago, you absolutely cannot trust them today, and any vendor that ships software that, by default, trusts the LAN is shipping software with severe security problems.
Running software on untrusted hardware can never really be secure. If the school wants to do this sort of thing, they need to provide the machines.
They could either buy a set of laptops specifically for exams, or they could buy some low-end machine whose primary function is word processing. Examples are the Dana AlphaSmart and the LaserPC. A simple cold boot will bring them back into a known configuration. Buying a few dozen of those may even be cheaper than a site license for the "Extegrity" software.
On the one hand, current DVDs really don't quite have the quality of some of the original film material, on the other hand, the technology to create HD-DVDs cost effectively simply wasn't there when DVDs first came out. So, this upgrade is quite justifiable, IMO. And an HD movie is still too large to be downloaded conveniently.
Take a look at your DVDs, though, and freeze some frames: for a lot of movies, it probably doesn't make much sense to get HD-DVDs because the quality of the original isn't all that high to begin with. You can already see plenty of film grain, fuzziness, and other film-related artifacts even with regular DVD resolutions and compression. Motion picture film just isn't all that great.
Attempts to create new audio disc formats, on the other hand, don't make much sense for the user: audio CDs really capture audio better than most stereos can reproduce it and most people can perceive it. Furthermore, the next "format" for audio is likely going to be the Internet. So, new audio formats are just an attempt at making lots of money with no real benefit to the customer.
Nothing particularly sinister. The "hooks" they have is probably caller ID on the modems. If he connected via DSL, they have the port he connected from (via the IP or MAC address). AOL is probably completely free to give out that information, with no need for wiretap authorization.
Normal 1U rack mounts or rack-mounted blade servers will work for consolidation: they can share power and they can be dynamically reassigned to different tasks without any need for virtualization.
NFS/NIS was an incompetent implementation of a feature useful to lots of people. That's why lots of people ended up using it even though it had no security and would corrupt their data with regularity. (NFSv4 may, finally, address these issues.)
Solaris 10 sounds like a competent implementation of features useful to only a tiny fraction of the market, if they are useful to anybody at all. Sun can go head to head with IBM with that, wooing a market consisting of maybe a few thousand customers. For the rest of us, I just don't think it matters.
I suspect RedHat doesn't support it (assuming they don't) because there isn't much demand for it. If you want multiple Linux instances, you just buy multiple machines.
Sun virtualization is rather bizarre from an economic point of view: first you pay a premium to get all your hardware in one huge "Enterprise Server", and then you virtualize it out again.
If that's the case, then why not run Windows NT on everything?
Because Windows NT is very expensive, because its APIs keep changing, and because it requires way too much manual effort to install and maintain.
At least an order of magnitude more, I'd say. Or maybe they have some other method of making sure their code is stable, such as special compilers or automated code verification... Whatever it is that they do, there is no denying that Solaris is far more stable than Linux.
Sun's systems have had really serious bugs in the past (memory leaks, data corruption, buggy code generation in the compilers, etc.), it's just that after years of bug reports from the fields, many of them have been fixed.
Whatever stability Solaris has, it's simply due to age, not any kind of special magical powers Sun has.
Don't tell me that nobody will ever need a ZB of storage. Maybe not tomorrow, but in 20 years? Yea.
You are absolutely right that reliability, performance, storage capacity, and service matter. But monolithic systems like Solaris running on monolithic hardware like Sun Enterprise servers won't be able to scale to those sizes.
Your problem [1] is that you're too short sighted; You don't see the big picture.
No, your problem (and Sun's problem) is that you are firmly stuck in a 1960's mindset.
[2] Not every Linux user or advocate is a fanatic. I use the term to refer to the more rabid zealots.
And I'm not a Linux fanatic either. The Linux kernel is pretty mediocre as far as kernels go. I think BSD or Mach or any of a number of other mediocre kernels would unseat Sun as well.
You see, it's not about whether Linux is "better" than Solaris, it's about the fact that Sun is just barking up the wrong tree: it just doesn't matter much what operating system you use to run your individual nodes, as long as it's cheap and simple. What matters is the application and storage software that runs on top of it and that can deal with both the distributed nature of the system and inevitable node failures.
Linux is a good OS and I am no stranger to it whatsoever, but it has a long way to go to catch up to Solaris.
Some races are not worth running in. This is one of them. Let Sun pursue pipe dreams of "Zettabyte disk drives", useless sequences-of-nines uptimes, and gigantic, overpriced shared memory machines, while humble little Linux forms the backbone of the real supercomputers and mainframes of the future: compute clusters built from unreliable, commodity machines.
but it isn't remotely close to being stable and reliable enough to play on the enterprise level. There's just too many features going in, unaudited, unverified.
And you think that Solaris is better audited and verified? I don't think so. Sun doesn't have the resources.
And what do you want anyway? If your enterprise setup is so flaky that an errant kernel or kernel security flaw compromises your operation, you have a problem, and it's not fixable by spending millions on gold-plated hardware and software.
You misunderstood the analogy. For Solaris vs. Linux, it's (unnecessary) scalability and robustness that is the key property, for the cars it's speed.
What I am saying is Linux is a nice car for daily commute, but Solaris is a better investment.
Well, let's try another analogy: do you keep your wallet in a bankvault or even a safe at home? Probably not, because it's too much hassle and not worth it.
Sun provides great support, hardware, compatibilty with past versions of it's software,
The SunOS/Solaris transition has been anything but smooth, and Sun sites have traditionally dreaded upgrading because it has been so disruptive. You do get decent support, but at a huge cost. And hardware compatibiliity? Well, it's a single-vendor solution--buy from a Linux hardware vendor and you'll get the same or a better experience with Linux than that.
Java, and more.
Yes, Solaris does have better support for Java than Linux. That tells you more about the fact that Java is a Sun-proprietary that Sun deploys according to their business interests than anything about Linux.
Features like this won't even make their way into Linux for another 5 years or so.
Let's hope so. If we are really lucky, most of them will never make it into Linux because they are useless bloat for almost all users.
Let Sun go for the really, really high end of million dollar hardware running specialized operating systems. I'm happy with Linux being the operating system for huge compute clusters made out of commodity hardware, and for that, Linux is more than enough already.
Solaris's problems are not in robustness or scalability--it already handles those very well. Trouble is, so do Linux and BSD and a lot of other systems. Arguably, not as well as Solaris, but well enough.
By analogy, sure, a Ferrari is a nice car, but for a daily commute, a Honda Civic is both cheaper and more practical, and it really doesn't matter that it doesn't go as fast as the Ferrari in theory. With software like Solaris 10, Sun is creating ever more expensive Ferraris.
I think it's a little more complicated than that. There are different kinds of innovation and different kinds of open source projects.
To me, X11 was quite innovative. Keep in mind that it came out a publicly financed research project as well.
Python and Perl, on the other hand, don't strike me as "innovative" at all: we have had scripting languages and interpreted languages like that since the 1960's, and both of them are architecturally still stuck in the dark ages. But, innovative or not, from a practical point of view, Python and Perl still beat any of the scripting solutions Apple or Microsoft offer.
Overall, OSS is a really mixed bag on innovation, but that's perfectly fine; it means that overall, there is a lot of innovation in OSS (often financed through public research grants). For Apple and Microsoft's product range, on the other hand, things seem simpler to me: for the most part, they aren't very innovative. What Apple and Microsoft do fairly well is graphic design, marketing, branding, developer relations, and sales. Good for building a business, but that's all.
That feature was originally present in a number of X11 window managers. But someone (HP, I believe) got a patent on it and people decided it wasn't worth fighting.
Maybe nowadays, HP doesn't care anymore or the patent has expired.
It's not clear to me that any of those window management hacks are all that useful anyway...