The problem is that the MS platform is technically lousy - too high coupling. Most of the people that work on alternative platforms do not WANT to work on the MS platform, as they find it inferior to the platforms they choose to work on.
As a such, "ceding the desktop" involves a direct sacrifice: Getting a desktop experience the people that develop the code *do not want*.
Your suggestion could be of merit if the developer pool was managed and constant; it is not, so the entire idea is moot.
Microsoft didn't get to a monopoly position by government assistance, The threat of men in blue uniforms with guns coming to your house, pulling you away and putting you in a small room for a long period of time - that's supplied by the government, and was supplied as assistance to Microsoft, in the form of copyright law.
In a free market, anything Microsoft does could be freely copied by anybody, for free.
So, Microsoft got where they are through a grant of limited monopoly by the government, enforced by men with guns.
They used this limited monopoly to get other monopolies, again supported by the government through contract enforcement over "per CPU" contracts - again, messing with the free market, with government men with guns supporting this.
So, to put things in balance again, it's fairly appropriate that the government comes and takes things away from them, as the government has been supporting them - and still are supporting them in the form of copyright grants. Of course, a simple remedy would be for government to remove their monopoly grants due to abuse; however, that might actually be too harsh a remedy. I like the simple one proposed - it just make the market need to do a full evaluation in more cases, which increase the information availability and thus brings us closer to an ideal market.
I'd guess that without Linux, we'd have a bunch of BSDs and a lot less use of the GPL license, all around. In my opinion, the reason free operating systems took off at that time is just sufficient developer mass on the Internet combined with hardware reaching the level where things could be implemented reasonably; something would have filled the free OS niche no matter what. Linux happened to take over for BSD due to being better at low end hardware and having a development model that made people feel more that it was "theirs" than BSD, so it was better in competition - but I believe BSD would have filled the niche well if Linux hadn't been there to take it. Better, worse - who can say? But well, sure.
You're welcome, and I apologize for the tone of my reply - it's just that this is a common "counter argument" that's put forth on creationist websites, and I'm extremely tired of getting those arguments thrown up all the time, after countering them literally hundreds of times. Yes, the complexity issue is real, and the consequence of it is that we believe there must have been some precursor to DNA that later led to DNA.
The clay hypothesis is one - it is based on clay crystals reproducing and then over time starting to use RNA as supporting material, and then RNA taking off from that. Another is based on autocatalytic reactions starting to be selected, and leading to RNA/DNA.
So what you're basically saying is that it is statistically impossible for a guy to throw the sun from the surface of the earth into the sky, therefore there is no sun in the sky, right?
Fortunately, there's effectively nobody that claim that the sun was thrown into the sky, so there's no problem. For a couple of examples of non-sun-throwing variants, read up on Spiegelman's Monster and how it occurs from non-RNA basis in the presence of Q-beta replicase (origin of self-replicating RNA sequences from amino acid soup *in the laboratory*) or Graham Cairns-Smith's clay hypothesis.
Science knows the challenges present, and are looking at them. The argument from personal incredulity ("I can't think of a way, therefore there is no way") is horribly arrogant - especially when you haven't bothered to look at the work done by the more knowledgeable and more numerous people than you, more numerous and more knowledgeable with more experience, more time and more communication.
It's not clear that it is unprofitable. There's only a few boilerplate letters per "lawsuit", and most are settled at $3000 or so. That's a heck of a lot of money for a mailing few letters.
In the article he says he launched the GNU OS in 1984 and seven years later a kid from Finland blows right past him. What was Stallman doing during those seven years? What's he been doing in the sixteen years since? While I am no great fan of Stallman, I still want to give him credit where due: He's been working on the operating system and orchestrating that effort, as opposed to the kernel. This includes GCC and Emacs. And he's been working on orchestrating the FSF and the political side of things.
Whether Stallman's variant of free software is a good thing is an open question - personally, I think he may have done more harm than good (as I think a more free license like the BSD license would have been better for everybody), but he's definitively been doing something...
NOTE: The below is based on some assumptions that things are about the same as they were 4-5 years ago when I last checked the respective architectural issues etc. I've not seen any sign of GCC doing massive leaps in architecture in the meantime, so I believe the analysis is still valid - but take it with the caveat that my knowledge may be dated.
GCC today would be monstrous to re-implement, and the benefit of doing so would be essentially nil. A pure reimplementation from scratch of the GCC architecture would have essentially zero benefits. Upgrading some of the alternative architectures - e.g. TenDRA to support enough to be able to compile the BSDs and then switching the compiler (or making the source tree possible to compile with several compilers) may be reasonable. There's benefits in different compilers than GCC; other compilers often have better error messages, other compilers can often compile faster (the GCC architecture result in fairly slow compilation, and TenDRA is usually way faster), and e.g. the Intel compiler will usually generate better code than GCC. That may be "would usually", I'm not up to date in the area.
Also, supporting different compilers at the source tree level will also usually make it easier to support source checkers (lints), which makes it easier to create correct code.
Having done part of this work some years ago (as I was doing linting of the FreeBSD kernel), I would guess the amount of work for fixing the FreeBSD kernel to be compilable with TenDRA at about one month's work. That assumes that TenDRA has gotten the asm extensions that was promised about 5-6 years ago; supposedly, these were really little work - less than a day, as I remember my understanding from the TenDRA developers.
Remember, MS at least has the policy of never shipping with known bugs.
Do they? When was this introduced, and do you have a reference? I remember Windows 2000 (?) shipping with some high-number-of-thousands bugs, just wondering when they changed and how they define this, assuming you're right.)
Actually, I didn't discourage the use of the GPL; I said the GPLv2 was technically incompatible with more or less any other license, which it seems we agree on:)
Overall, I both encourage and discourage the use of the GPL - whether the GPL is appropriate depends on context. I encourage people to think carefully about what license they choose, what effects that license has, and whether these effects are appropriate for their goals.
I'll admit that I often find that people seems to choose the GPL based on incorrect beliefs, and that I discourage those beliefs. The most prominent example is the belief that the BSD licensed codebases do not get any changes back from proprietary derivates - proprietary derivates give back significant technology to the BSD projects - and that if they had been GPLed there would be changes flowing back - there wouldn't, as the proprietary derivates are made to make money, money that they couldn't make if the codebase had been under the GPL.
As a such, you could say I argue against the GPL - but that's distinct from the arguments around the GPLv2 and BSD licenses. The point in that case is to make people be careful with this, as getting free software projects suddenly yanked over license issues would be lousy for free software overall. I see it as much better if people handle licenses right in the first place.
IMO, it is outside the spirit of the GPL as it adds significant extra restrictions. The only reason that it is accepted by the GPL community is that they don't have experience with the situations where those restrictions show up.
I find it hard to say what would be more of a miscarriage of justice; blocking distribution/modification as per the GPL, to remedy the situation by restoring the rights of the original author per the GPL (s)he released under (and removing the ability of derivates to distribute), or "keeping the rights the community expected", when the community was behaving negligently. I see any remedy beyond blocking distribution/modification as almost certainly being unreasonable.
The problem, as I see it, is that the ability to shoot down distribution/modification of the derivate works is a quite large threat. Effectively, this means killing the derivate work - the one where BSDed code was introduced. So, introducing BSD licensed code in a GPLed codebase risk a very difficult situation later on, unless the person introducing it has control of the GPL licensing side of the codebase.
As for the psychology part: I recommend reading the first three chapters of Myers' "Social Psychology" - it covers in detail many of the ways we distort recall and attitudes, with a bunch of caveats and conditions as to when it happens. It's much better if you get it from there than from me. While I've read different sources in the area, Myers has organized them carefully and has deeper knowledge, so you'll get better information. My personal experience is that straight thinking is extremely tricky - there's a bunch of forces that try to push us off it in the direction of "the kind of thinking that is socially most useful".
The further we get into this discussion the more I think I agree with the FSF's position
That's basic social psychology: You are arguing that position, therefore you are believing more strongly in it. When used actively, it is known as "inoculation theory" - use weak attacks on a position to strengthen the belief. I don't remember the name when it "just happens"; e.g. Tory Higgins has done research on this, and there's coverage in "Social Psychology" by Myers (chapter 4, Behaviour and Attitudes). You'll find this form of attitude adjustment is a core result of social psychology.
Back to discussion: You're missing the GPL vs BSD license point. The point is that the GPL prohibit further restrictions, and the 3-clause BSD license introduce restrictions. So, the restrictions in the 3-clause BSD that make it unsuitable for embedded development in some cases *make it GPL incompatible*. Of course you would have to do that reproduction when using 3-clause BSDed code even without the GPL - that's the point.
WRT your claims:
You're also assuming that the examples I mentioned are the only ways of getting at this. They were just examples. Other ways of getting at this could be to pay off the original author enough, when others have picked up the code and done the mixin - more or less pay off killing the entire downstream from the point where non-GPLed code was mixed in.
"Writing out" contributions is a remedy that isn't strongly tested, and that isn't anywhere near trivial when refactoring etc has been done. "Lack of damages" is also a proposition that hasn't been tested; there's been several IMO credible directions for how to show damages. Anyway, these apply to ALL cases of the GPL being incompatible with other licenses - you're arguing that "Putting the GPL on anything doesn't really matter, as we can fix it when we get hit with it". That is a total tangent to whether it is inside or outside the license.
Think "Embedded system". For instance, I once programmed for a wireless handsfree. The final product came in a box about 10x10x10cm (that's 4" by 4" by 4"), including printed documentation, charger, and the handsfree itself. There's nowhere to put a file, unless we added an extra CD just for this.
A derivate that want to use GPLed software for this kind of work is then encumbered with an extra burden, compared to plain GPL. The GPL goes not allow that.
And there's one more case than a "GPL author got pissed off" - it is if the GPL author is specifically paid for getting pissed off, maybe even doing the development just to be able to sue. This could happen either because of some company doing an attack - e.g, Microsoft - or because some BSD developer got pissed off by the behaviour of GPL authors and did the pure GPL contributions to be able to sue and require separation.
Eivind.
Re:Biggest myths of all have been around for ages.
on
Why Myths Persist
·
· Score: 1
I hope you include Hinduism in that "religion" basket you see as having many arguments in its favor. And the old Norse pantheon. And so on.
Otherwise, I'm just disbelieving one more religion than you, putting you at 99.9%+ disbelieving in religion.
No, I don't see myself as "understanding the basics" - I see myself as having a deep understanding acquired over the almost twenty years since I first started sniffing at free software.
I am a (presently inactive) committer for FreeBSD and have been for over 10 years. I had considered myself part of the free software movement for years before that, though not very active. My way of thinking is that I want more free software, not less proprietary software. I see the increase in free software and free software use as increasing freedom; I see blocking the creation of proprietary software for the niches that free software can't serve as decreasing freedom. I include increasing expense until it is prohibitive in "blocking", as that is the only kind of blocking that can happen.
You are, as far as I can tell, arguing that this blocking is "increasing freedom". That's what I see as "black is white".
And no, I would not have to follow the style rules for FreeBSD to deliver to customers; that is the point. From the point of view of the customers, the patch is of good quality. From the point of view of FreeBSD, however, something that doesn't follow the style rules is of bad quality, and I can't commit it.
Slashdot Mirror
As a such, "ceding the desktop" involves a direct sacrifice: Getting a desktop experience the people that develop the code *do not want*.
Your suggestion could be of merit if the developer pool was managed and constant; it is not, so the entire idea is moot.
Eivind.
In a free market, anything Microsoft does could be freely copied by anybody, for free.
So, Microsoft got where they are through a grant of limited monopoly by the government, enforced by men with guns.
They used this limited monopoly to get other monopolies, again supported by the government through contract enforcement over "per CPU" contracts - again, messing with the free market, with government men with guns supporting this.
So, to put things in balance again, it's fairly appropriate that the government comes and takes things away from them, as the government has been supporting them - and still are supporting them in the form of copyright grants. Of course, a simple remedy would be for government to remove their monopoly grants due to abuse; however, that might actually be too harsh a remedy. I like the simple one proposed - it just make the market need to do a full evaluation in more cases, which increase the information availability and thus brings us closer to an ideal market.
Eivind.
Eivind.
I'll stick with insightful, based on experience - but interesting will do much of the same :)
Eivind.
The BSD license introduce the additional restriction of reproduction of the exact words of the BSD license.
Eivind.
Eivind.
The clay hypothesis is one - it is based on clay crystals reproducing and then over time starting to use RNA as supporting material, and then RNA taking off from that. Another is based on autocatalytic reactions starting to be selected, and leading to RNA/DNA.
Eivind.
Fortunately, there's effectively nobody that claim that the sun was thrown into the sky, so there's no problem. For a couple of examples of non-sun-throwing variants, read up on Spiegelman's Monster and how it occurs from non-RNA basis in the presence of Q-beta replicase (origin of self-replicating RNA sequences from amino acid soup *in the laboratory*) or Graham Cairns-Smith's clay hypothesis.
Science knows the challenges present, and are looking at them. The argument from personal incredulity ("I can't think of a way, therefore there is no way") is horribly arrogant - especially when you haven't bothered to look at the work done by the more knowledgeable and more numerous people than you, more numerous and more knowledgeable with more experience, more time and more communication.
Eivind.
All BSD licenses have one more requirement that the GPL does not: Reproduction of the exact words of the BSD licenses. Which there can be many of.
Ah, cool, I'm glad to hear that. Thanks! (And thanks for the work you do around this; even though it isn't in my country, it is appreciated.)
Yours truly,
God.
Eivind.
Whether Stallman's variant of free software is a good thing is an open question - personally, I think he may have done more harm than good (as I think a more free license like the BSD license would have been better for everybody), but he's definitively been doing something...
Eivind.
Eivind.
Also, supporting different compilers at the source tree level will also usually make it easier to support source checkers (lints), which makes it easier to create correct code.
Having done part of this work some years ago (as I was doing linting of the FreeBSD kernel), I would guess the amount of work for fixing the FreeBSD kernel to be compilable with TenDRA at about one month's work. That assumes that TenDRA has gotten the asm extensions that was promised about 5-6 years ago; supposedly, these were really little work - less than a day, as I remember my understanding from the TenDRA developers.
Eivind.
SSIA.
Do they? When was this introduced, and do you have a reference? I remember Windows 2000 (?) shipping with some high-number-of-thousands bugs, just wondering when they changed and how they define this, assuming you're right.)
Eivind.
Overall, I both encourage and discourage the use of the GPL - whether the GPL is appropriate depends on context. I encourage people to think carefully about what license they choose, what effects that license has, and whether these effects are appropriate for their goals.
I'll admit that I often find that people seems to choose the GPL based on incorrect beliefs, and that I discourage those beliefs. The most prominent example is the belief that the BSD licensed codebases do not get any changes back from proprietary derivates - proprietary derivates give back significant technology to the BSD projects - and that if they had been GPLed there would be changes flowing back - there wouldn't, as the proprietary derivates are made to make money, money that they couldn't make if the codebase had been under the GPL.
As a such, you could say I argue against the GPL - but that's distinct from the arguments around the GPLv2 and BSD licenses. The point in that case is to make people be careful with this, as getting free software projects suddenly yanked over license issues would be lousy for free software overall. I see it as much better if people handle licenses right in the first place.
Eivind.
I find it hard to say what would be more of a miscarriage of justice; blocking distribution/modification as per the GPL, to remedy the situation by restoring the rights of the original author per the GPL (s)he released under (and removing the ability of derivates to distribute), or "keeping the rights the community expected", when the community was behaving negligently. I see any remedy beyond blocking distribution/modification as almost certainly being unreasonable.
The problem, as I see it, is that the ability to shoot down distribution/modification of the derivate works is a quite large threat. Effectively, this means killing the derivate work - the one where BSDed code was introduced. So, introducing BSD licensed code in a GPLed codebase risk a very difficult situation later on, unless the person introducing it has control of the GPL licensing side of the codebase.
As for the psychology part: I recommend reading the first three chapters of Myers' "Social Psychology" - it covers in detail many of the ways we distort recall and attitudes, with a bunch of caveats and conditions as to when it happens. It's much better if you get it from there than from me. While I've read different sources in the area, Myers has organized them carefully and has deeper knowledge, so you'll get better information. My personal experience is that straight thinking is extremely tricky - there's a bunch of forces that try to push us off it in the direction of "the kind of thinking that is socially most useful".
That's basic social psychology: You are arguing that position, therefore you are believing more strongly in it. When used actively, it is known as "inoculation theory" - use weak attacks on a position to strengthen the belief. I don't remember the name when it "just happens"; e.g. Tory Higgins has done research on this, and there's coverage in "Social Psychology" by Myers (chapter 4, Behaviour and Attitudes). You'll find this form of attitude adjustment is a core result of social psychology.
Back to discussion: You're missing the GPL vs BSD license point. The point is that the GPL prohibit further restrictions, and the 3-clause BSD license introduce restrictions. So, the restrictions in the 3-clause BSD that make it unsuitable for embedded development in some cases *make it GPL incompatible*. Of course you would have to do that reproduction when using 3-clause BSDed code even without the GPL - that's the point.
WRT your claims: You're also assuming that the examples I mentioned are the only ways of getting at this. They were just examples. Other ways of getting at this could be to pay off the original author enough, when others have picked up the code and done the mixin - more or less pay off killing the entire downstream from the point where non-GPLed code was mixed in.
"Writing out" contributions is a remedy that isn't strongly tested, and that isn't anywhere near trivial when refactoring etc has been done. "Lack of damages" is also a proposition that hasn't been tested; there's been several IMO credible directions for how to show damages. Anyway, these apply to ALL cases of the GPL being incompatible with other licenses - you're arguing that "Putting the GPL on anything doesn't really matter, as we can fix it when we get hit with it". That is a total tangent to whether it is inside or outside the license.
Eivind.
A derivate that want to use GPLed software for this kind of work is then encumbered with an extra burden, compared to plain GPL. The GPL goes not allow that.
And there's one more case than a "GPL author got pissed off" - it is if the GPL author is specifically paid for getting pissed off, maybe even doing the development just to be able to sue. This could happen either because of some company doing an attack - e.g, Microsoft - or because some BSD developer got pissed off by the behaviour of GPL authors and did the pure GPL contributions to be able to sue and require separation.
Eivind.
Otherwise, I'm just disbelieving one more religion than you, putting you at 99.9%+ disbelieving in religion.
Eivind.
Eivind.
I am a (presently inactive) committer for FreeBSD and have been for over 10 years. I had considered myself part of the free software movement for years before that, though not very active. My way of thinking is that I want more free software, not less proprietary software. I see the increase in free software and free software use as increasing freedom; I see blocking the creation of proprietary software for the niches that free software can't serve as decreasing freedom. I include increasing expense until it is prohibitive in "blocking", as that is the only kind of blocking that can happen.
You are, as far as I can tell, arguing that this blocking is "increasing freedom". That's what I see as "black is white".
And no, I would not have to follow the style rules for FreeBSD to deliver to customers; that is the point. From the point of view of the customers, the patch is of good quality. From the point of view of FreeBSD, however, something that doesn't follow the style rules is of bad quality, and I can't commit it.
Eivind.