Unfortunately, just counting downloads isn't a very good way to do this, since those numbers are relatively easy to manipulate in various ways. Far better would be to survey x thousand people whose computers have been carefully checked for malware. Or to use "trusted computing" hardware to report reliable statistics from people's media players (how's that for an ironic use of technology).
What about Creative Commons? Now the license is completely pointless - just pay your tax and you are free to copy all you want!
Creative Commons licenses are just attempts to reduce the harm inflicted by a ludicrously inefficient and burdensome digital copyright system. It would be far better to fix the law so that it works well for every work, not just those that are CC-licensed. Most of the people involved in creating the Creative Commons licenses in the first place would agree with this analysis.
What if you're an unsigned band? Will you get paid for the copies that can be legally distributed under this law?
If for some silly reason they can't, well, every unsigned band and all of their fans should raise hell until the system is implemented properly. They'd have a pretty persuasive case in the French parliament.
What if the government decided to enact a blanket "graffiti tax" and legalize graffiti? Why should we pay for damage that we have not caused?
How do you propose to draw a cogent analogy between copying MP3 files and drawing on the side of buildings?
There is one part of your post that I do kind of agree with though: "You, sir, are a troll, plain and simple":)
For people who currently observe the law and do not download at all (or only download stuff the copyright owner has given away), this is a tax with no return.
What about the now-legal option of commencing to download free music? That's a pretty nice return! There are others too, like lower CD prices. The only losers would be people who want a net connection but have no interest in any cultural works whatsoever. The needs of the many outweight the needs of the few.
It weakens the rights of authors and hands tax money to the publishers.
Why? As an artist, you can self-publish on the 'net, and if the system is designed right you'll get payed in proportion to your popularity.
Actually, I have a pet theory about why publishers have been so opposed to these sorts of levy schemes. I think they're afraid that it would only be a short step to a mandated revenue split, saying something like "artists must get at least 50% of this money". Such splits exist in some countries' private copying and public lending rights schemes, so the precedent is clear. It would be a radical change from the less than 5% they get from CD sales at present!
But follow me further, if you will: What happens if something like GPL'd software gets included in the definition of content that right now we think will only include songs and music? Would a French company be allowed to re-distribute GPL'd software in violation of the terms of the GPL by claiming this law frees them of the constraints of copyright?
Well, assuming that the system is applied to software (which it probably wouldn't be)... what incentives would a firm have to keep their source code secret if their users didn't have to pay for binaries? Remember also that if they distributed the code in any country where the law enforced the GPL, they'd have to comply.
If for some strange reason this did turn into a problem, it would be easy to add a clause exluding already free/open source software from the blanket license.
I agree that collecting societies have a poor track record on distributing levies. But that's because they were forced to rely on crazy methodologies like monitoring what got played by radio stations. These days they could mix in data from sources like last.fm, provided nobody knew what they were up to. If people knew their methodology, a lot more security would need to be in place between the play button and the final royalty cheque.
The proportion of comments here criticising this woman is a disgrace. Yes, the most cynical analyses of how law firms are self-serving may apply. If so, why are slashdotters so keen to defend this behaviour?
Copyright and patent law are complicated systems with (many would argue) important social functions that need to be finely balanced. A competent IP firm should be representing both plaintiffs and defendants in these systems. Furthermore, it is not clear that DRM is good, even for an IP firm's copyright holding clients, or for the copyright system as a whole. Employees who think critically about these matters are to be commended.
Perhaps an argument can be made that IP law firms benefit from having the worst and most tangled laws possible, and that regardless of the implications for cultural industries or for good policy, the firm should be encouraging chaos and DRM. It's not good to see slashdotters being so keen on this kind of free commerce.
Personally, I find the idea of paying a levy on every piece of media I *could* use to pirate music repugnant.
Personally, I find the idea of paying taxes so that Australia can purchase weapons -- overpriced weapons, in fact -- from wealthy US arms dealers really repugnant. But I don't advocate the abolition of the tax system, because I think that on balance, we'd probably be worse off without it. A better government would be nice though:) If enough people listen to and enjoy your Dharma recordings, the center should get a slice of the levy revenue.
PS - have you tried recording with laptops? Or with iaudios?
Doesn't that provide a slam-dunk defense for anyone accused of infringing a software patent? It seems that if you were sued for infringement you could just point out to the court that the patent was erroneously issued. After a couple such cases, the precedent would be firmly established and future defendants would hardly have to do more than show up.
That would be hard, MSH leverages.NET quite extensively. You might see a Mono Shell, or a Python Shell using the same concepts though.
I tried writing a object-oriented python shell a few years back. There are certainly a great many improvements to be had over bash or zsh. Nice tab completion was difficult to implement though, because parsing incomplete python commands is tricky and because other strategies (such as building hypothetical constructions and prodding them) are made difficult by python's highly dynamic nature...
Unlike most other nations' legal systems, human dignity and therefore privacy is central to the German constitution (this was a result of its being drafted in the wake of second world war). It follows that German copyright law does not trump privacy concerns; this was one of the reasons why Germany invented the levy-funded private copying system.
In some cases, the most efficient way to deploy networks is to build just one with universal and ungated coverage.
Your concern about the behvaiour of state-run telcos of course has a basis in history. But it's less likley to be a problem on a locality-by-locality basis. Each local government has to compete with its neighbours to provide better value for tax.
Otherwise businesses, and eventually residents will vote -- first with their ballot papers, or if that fails, with their feet.
Well, this time I think I can stand by my comment a little more firmly. I glanced at a few Oppedahl & Larson patents whose titles looked software related, and about half of them (#6826714, #6816022) are "unimpressive" even on reading the claims. This is not your fault. You have a business to run and clients to serve. The system doesn't give you much incentive to on only work to patent significant breakthroughs, or to do authoritative searches of non-patent literature.
Of course I understand that if I actually have prior art (rather than the feeling that I have learnt nothing from the filing), I could send it to the US patent office during examination. Or here in Australia, I could initiate an opposition. This is not at all the point.
Plenty of developers are threatened with patents that are probably invalid. Even when they are valid, the applicable standards of novelty and non-obviousness provide no guarantee that society is benefiting from the exchange of invention effort and monopoly. I claim that this is very rarely the case for single software patents. And that on top of that, the overheads involved in living with the system as a whole are large.
As an occasional free software developer, do you think I want to have to spend time searching for combinations of words that might somehow describe what I'm working on (the chances that the disclosure would ever help me solve a problem are very slim indeed)? Participate in the application process, to notify the patent holder of my existence and risk willful infringement?
Even if it turns out that your claim is correct and sw patents are a comparative aid to SMEs as opposed to large firms, I think most of my concerns about their economic effects remain valid.
Your bald and unsupported statement is that a small player who has invented something will trip over other patents. (Note the "'ll" in your posting.)
I agree that this shouldn't have been universal (I'd say in my defence that it is a common rhetorical mode of English expression).
I doubt you can point to even two real-life examples of this happening.
A fair challenge, which I found much harder to answer than I had expected. Aside from the rather large and hairy case of the GSM licensing regime, Google didn't provide any quick answers. If I get a spare moment, I'll see if I can find some examples by more rigorous research.
Carl, you appear to have a common misconception about the effect of software patents.
The situation you are describing is applicable only for small players who come up with a very nifty algorithm or solution to something, but don't intend to ever produce a product. If they try to actually do anything, they'll trip over other patents and be forced to cross-licence by the major players.
Most software innovation isn't a single "breakthrough" step like RSA, LZ77/LZ78, DocuComp or whatever (the last of which is not really such a breakthrough), which could be the basis of licensing business strategy. It's loads of things put together into a useful program.
It's extremely unlikely that the negative economic effects of software patents (especially in relation to control of standards, because the ratio of patentable innovations to products is so high compared to other industry sectors, and because of the effect on free/open source software) are countered by the increase in incentives for small players to come up with important new algorithms. That scenario is vanishingly rare compared to the thousands of software patents that are issuing. And let's face it, Rivest, Shamir and Adelman would have been working hard on asymmetric crypto even if there hadn't been the extra carrot of a patent for MIT.
This is one of the strangest and most perverse arguments I've ever heard in favour of patents. "Obstructions are good because they force us to overcome them". Can't we think of less destructive ways to encourage innovation?
I'd be curious to know whether you would support the kind of interoperability exception contained in article 6a of the EU parliament's (amended, anti-swpat) version of the software patent directive? If so, you might want to support or donate to the FFII anyway, because they played a key role in getting the parliament to vote for that.
I spent a bit of time worrying about those problems in the course of my own research on public funding systems as alternatives to copyright.
The conclusion I came to was that there just isn't a fair way to split the credit for large collaborative projects. If you think the linux kernel is hard, what about a nifty embedded gadget or something which relies on the kernel for 90% of its functionality, but was made by completely independent developers?
The fact that you got a threat letter doesn't mean that you were infringing copyright. The CRIA calims it was infringement, but that doesn't mean a court would agree.
In most countries, sharing even one copy of a file over a p2p network is not fair dealing/fair use. Over IRC, however, if you know the person at the other end, it might be.
The story in Canada is different. The law at this precise moment appears to be that almost all p2p sharing is legal because of the levies you pay on blank CDs, iPods, etc.
Your government is about to change that and make file sharing illegal, unless you do something about it.
There is always a workaround. It may be "chipping" the motherboard - possibly will be illegal, but who cares. It may be even running a pair of computers, using the TCPA one as an access device for the non-TCPA one.
Not necessarily. At some point, they're going to start using tamper resistant hardware. Good luck mod-chipping that. The only hole you can actually count on, is digital-to-analogue-to-digital conversion, to get data out of your "trusted" box and into an ordinary computer. That might be fine for ebook pirates, but it's going to be one hell of a PITA for daily life.
Rather than planning to spend tens or hundreds of hours on work-arounds when this happens, it would make sense to donate tens or hundreds of hours of your wages to the EFForsimilarorganisations which are working to prevent this problem from prevailing in the first place.
I am really glad you have given extensive thought to this conversation, instead of just flaming my original post as I expected would happen. I do have a license that I am currently writing all of my code against, and can provide you the URL via email (don't want it slashdotted) if you are interested in providing feedback regarding it.
Sure, I can have a look, although I can't guarantee that I'll be able to make any genuinely helpful observations:).
Perhaps I am misunderstanding the GPL and its FAQ (which in itself is a problem if people can't agree on what it says)
Yes, there is some misunderstanding, but I think it stems from the underlying copyright law, rather than the GPL. A license is a grant of permission; there's nothing in the GPL itself which stops you from granting other licenses (further permission) if you own copyright to a work.
The situation is somewhat complicated, because the law says that if you make derivative work (a modified version of someone else's code), they own most of the copyright (only "most" of it because you retain the exclusive right of derivation). But in the case of a java program, it's pretty hard to argue that it's a derivative of object.java, unless you're distributing it as one big linked-together system. So the GPL doesn't have power over your code when it's on its own. You're required to grant GPL permissions for the whole system; but if you're using some permissive license like X11 or BSD, you've already done that.
This section explains that you can ONLY dual license if it is done at the beginning. You can not later choose to add another license to the already-GPL'd program.
I can't find any claim like that on the other end of that link! It's always possible to move from GPL-only to dual licensing, but you need permission from all of the contributors, so if you leave it too late it can be very hard.
And, it seems, you got the licensing wrong with Cygwin. It's perfectly allowable to link your LGPL app to a GPL library (the GPL is satisfied because any developer could take your code, and the cygwin library, and run off to build some larger GPLed app with it).
If your employers are scared of the GPL like that, it's probably because they don't understand it. Many of the biggest players in the industry are quite happy to use the GPL for some of their code -- IBM, Sun, SGI, Novell, and RedHat all spring to mind.
Ok, so what is the problem here? That means that if Object.java were GPL, every single Object in the Java Programming Language would also have to be GPL, every program that uses Java, etc. Therefore, is Object.java were made GPL, you could ONLY make GPL software. Most companies are not willing to do that, because it imposes restrictions on how they use THEIR code. Many software developers (BSD, Apache, etc) are not willing to do that either, for the same reason. If I **ever** want to make my code Public Domain, say, it can NEVER be GPL first. Public Domain is the ONLY true Public license, with the various Open Source licenses being derivatives.
It's not that you can "ONLY make GPL software" -- it's just that you have to allow a GPLed version (this is called being "GPL compatible"). The GPL never stops you from allowing others to do things with code you own the copyright to. So in our example, if you write a public domain library on top of object.java, you can let 3rd parties write proprietary apps with it; they just have to make sure that they have met the licensing requirements for any java stuff they use (perhaps by paying Sun a license fee, or using someone else's version of Java).
...What is the best possible scenario FOR THE COMMUNITY? The best scenario is if it is used EVERYWHERE....
I think you have an oversimplified notion of benefit to the community. Community benefit is a complicated function of what code gets written, how much money developers are able to earn to support their coding, and how code is openly available to be improved or incorporated in new projects. Dual GPL/proprietary licensing (as used by, for example MySQL or QT) allows some developers to simultaneously earn revenue and produce free software. GPLed libraries encourage people who are tossing up between open or closed source development, to take the open route, because it's cheaper -- you can use more of the existing tools. If your management doesn't like the GPL, it shouldn't be too hard to get them to cough up the fee for the alternative license.
...And, let's take it a step further. Let's say **I** wrote it. And let's say I get hired on to write it for some BeOS version 8. Well, that's great! Except, I have to be careful NOT to use the code I already wrote, rewrite it from scratch and make sure it doesn't seem similar in code....
If you still own the copyright to your original version, you can still use it. The fact that there's a GPL licensed version floating around on the net does not prevent you from licensing it to BeOS on different terms.
If I want to write a proprietary Java program, right now, it would be free. Why impose a paid license?
To earn some cash to fund development of the language, perhaps? The idea is something like: those who write proprietary apps contribute to the common good in cash; those who write free software contribute in kind.
If just the Object class was GPL'd, no one could write truly free open-source (say, BSD) java programs.
complain about how Java is closed-source, though much of it is under Apache (or Apache-like) licenses
"much of it"... except the JVM, perhaps?
GPLing the JVM would be a very wise choice for Sun, for exactly the "business unfriendly" reasons you cite. Java would get the benefits of widespread adoption by the Free Software community; at the same time, if firms want to write proprietary apps against it, they'd have to pay for a license, thus preserving healthy revenue streams for Sun.
Slashdot Mirror
Unfortunately, just counting downloads isn't a very good way to do this, since those numbers are relatively easy to manipulate in various ways. Far better would be to survey x thousand people whose computers have been carefully checked for malware. Or to use "trusted computing" hardware to report reliable statistics from people's media players (how's that for an ironic use of technology).
What about Creative Commons? Now the license is completely pointless - just pay your tax and you are free to copy all you want!
:)
Creative Commons licenses are just attempts to reduce the harm inflicted by a ludicrously inefficient and burdensome digital copyright system. It would be far better to fix the law so that it works well for every work, not just those that are CC-licensed. Most of the people involved in creating the Creative Commons licenses in the first place would agree with this analysis.
What if you're an unsigned band? Will you get paid for the copies that can be legally distributed under this law?
If for some silly reason they can't, well, every unsigned band and all of their fans should raise hell until the system is implemented properly. They'd have a pretty persuasive case in the French parliament.
What if the government decided to enact a blanket "graffiti tax" and legalize graffiti? Why should we pay for damage that we have not caused?
How do you propose to draw a cogent analogy between copying MP3 files and drawing on the side of buildings?
There is one part of your post that I do kind of agree with though: "You, sir, are a troll, plain and simple"
For people who currently observe the law and do not download at all (or only download stuff the copyright owner has given away), this is a tax with no return.
What about the now-legal option of commencing to download free music? That's a pretty nice return! There are others too, like lower CD prices. The only losers would be people who want a net connection but have no interest in any cultural works whatsoever. The needs of the many outweight the needs of the few.
It weakens the rights of authors and hands tax money to the publishers.
Why? As an artist, you can self-publish on the 'net, and if the system is designed right you'll get payed in proportion to your popularity.
Actually, I have a pet theory about why publishers have been so opposed to these sorts of levy schemes. I think they're afraid that it would only be a short step to a mandated revenue split, saying something like "artists must get at least 50% of this money". Such splits exist in some countries' private copying and public lending rights schemes, so the precedent is clear. It would be a radical change from the less than 5% they get from CD sales at present!
But follow me further, if you will: What happens if something like GPL'd software gets included in the definition of content that right now we think will only include songs and music? Would a French company be allowed to re-distribute GPL'd software in violation of the terms of the GPL by claiming this law frees them of the constraints of copyright?
Well, assuming that the system is applied to software (which it probably wouldn't be)... what incentives would a firm have to keep their source code secret if their users didn't have to pay for binaries? Remember also that if they distributed the code in any country where the law enforced the GPL, they'd have to comply.
If for some strange reason this did turn into a problem, it would be easy to add a clause exluding already free/open source software from the blanket license.
I agree that collecting societies have a poor track record on distributing levies. But that's because they were forced to rely on crazy methodologies like monitoring what got played by radio stations. These days they could mix in data from sources like last.fm, provided nobody knew what they were up to. If people knew their methodology, a lot more security would need to be in place between the play button and the final royalty cheque.
The proportion of comments here criticising this woman is a disgrace. Yes, the most cynical analyses of how law firms are self-serving may apply. If so, why are slashdotters so keen to defend this behaviour?
Copyright and patent law are complicated systems with (many would argue) important social functions that need to be finely balanced. A competent IP firm should be representing both plaintiffs and defendants in these systems. Furthermore, it is not clear that DRM is good, even for an IP firm's copyright holding clients, or for the copyright system as a whole. Employees who think critically about these matters are to be commended.
Perhaps an argument can be made that IP law firms benefit from having the worst and most tangled laws possible, and that regardless of the implications for cultural industries or for good policy, the firm should be encouraging chaos and DRM. It's not good to see slashdotters being so keen on this kind of free commerce.
Personally, I find the idea of paying a levy on every piece of media I *could* use to pirate music repugnant.
:) If enough people listen to and enjoy your Dharma recordings, the center should get a slice of the levy revenue.
Personally, I find the idea of paying taxes so that Australia can purchase weapons -- overpriced weapons, in fact -- from wealthy US arms dealers really repugnant. But I don't advocate the abolition of the tax system, because I think that on balance, we'd probably be worse off without it. A better government would be nice though
PS - have you tried recording with laptops? Or with iaudios?
Doesn't that provide a slam-dunk defense for anyone accused of infringing a software patent? It seems that if you were sued for infringement you could just point out to the court that the patent was erroneously issued. After a couple such cases, the precedent would be firmly established and future defendants would hardly have to do more than show up.
:)
No
That would be hard, MSH leverages .NET quite extensively. You might see a Mono Shell, or a Python Shell using the same concepts though.
I tried writing a object-oriented python shell a few years back. There are certainly a great many improvements to be had over bash or zsh. Nice tab completion was difficult to implement though, because parsing incomplete python commands is tricky and because other strategies (such as building hypothetical constructions and prodding them) are made difficult by python's highly dynamic nature...
Unlike most other nations' legal systems, human dignity and therefore privacy is central to the German constitution (this was a result of its being drafted in the wake of second world war). It follows that German copyright law does not trump privacy concerns; this was one of the reasons why Germany invented the levy-funded private copying system.
It's amazing where the actual attempts at peer review end up burried on slashdot :)
In some cases, the most efficient way to deploy networks is to build just one with universal and ungated coverage.
Your concern about the behvaiour of state-run telcos of course has a basis in history. But it's less likley to be a problem on a locality-by-locality basis. Each local government has to compete with its neighbours to provide better value for tax.
Otherwise businesses, and eventually residents will vote -- first with their ballot papers, or if that fails, with their feet.
Well, this time I think I can stand by my comment a little more firmly. I glanced at a few Oppedahl & Larson patents whose titles looked software related, and about half of them (#6826714, #6816022) are "unimpressive" even on reading the claims. This is not your fault. You have a business to run and clients to serve. The system doesn't give you much incentive to on only work to patent significant breakthroughs, or to do authoritative searches of non-patent literature.
Of course I understand that if I actually have prior art (rather than the feeling that I have learnt nothing from the filing), I could send it to the US patent office during examination. Or here in Australia, I could initiate an opposition. This is not at all the point.
Plenty of developers are threatened with patents that are probably invalid. Even when they are valid, the applicable standards of novelty and non-obviousness provide no guarantee that society is benefiting from the exchange of invention effort and monopoly. I claim that this is very rarely the case for single software patents. And that on top of that, the overheads involved in living with the system as a whole are large.
As an occasional free software developer, do you think I want to have to spend time searching for combinations of words that might somehow describe what I'm working on (the chances that the disclosure would ever help me solve a problem are very slim indeed)? Participate in the application process, to notify the patent holder of my existence and risk willful infringement?
Even if it turns out that your claim is correct and sw patents are a comparative aid to SMEs as opposed to large firms, I think most of my concerns about their economic effects remain valid.
Your bald and unsupported statement is that a small player who has invented something will trip over other patents. (Note the "'ll" in your posting.)
I agree that this shouldn't have been universal (I'd say in my defence that it is a common rhetorical mode of English expression).
I doubt you can point to even two real-life examples of this happening.
A fair challenge, which I found much harder to answer than I had expected. Aside from the rather large and hairy case of the GSM licensing regime, Google didn't provide any quick answers. If I get a spare moment, I'll see if I can find some examples by more rigorous research.
PS - might be convinceable on this issue, but not impressed by some of the applications your firm is administering.
Passphrases are just long passwords with (usually) low entropy.
In fact, the entropy of passphrases might be so low (perhaps even 1 bit per character) that they can be inferred from keystroke timing analysis.
Carl, you appear to have a common misconception about the effect of software patents.
The situation you are describing is applicable only for small players who come up with a very nifty algorithm or solution to something, but don't intend to ever produce a product. If they try to actually do anything, they'll trip over other patents and be forced to cross-licence by the major players.
Most software innovation isn't a single "breakthrough" step like RSA, LZ77/LZ78, DocuComp or whatever (the last of which is not really such a breakthrough), which could be the basis of licensing business strategy. It's loads of things put together into a useful program.
It's extremely unlikely that the negative economic effects of software patents (especially in relation to control of standards, because the ratio of patentable innovations to products is so high compared to other industry sectors, and because of the effect on free/open source software) are countered by the increase in incentives for small players to come up with important new algorithms. That scenario is vanishingly rare compared to the thousands of software patents that are issuing. And let's face it, Rivest, Shamir and Adelman would have been working hard on asymmetric crypto even if there hadn't been the extra carrot of a patent for MIT.
However, I hope I'll reach my destination tomorrow in time: 170 km in 8 hours on a mountain bike :/
I hope you have slick tires.
I'd be curious to know whether you would support the kind of interoperability exception contained in article 6a of the EU parliament's (amended, anti-swpat) version of the software patent directive? If so, you might want to support or donate to the FFII anyway, because they played a key role in getting the parliament to vote for that.
The conclusion I came to was that there just isn't a fair way to split the credit for large collaborative projects. If you think the linux kernel is hard, what about a nifty embedded gadget or something which relies on the kernel for 90% of its functionality, but was made by completely independent developers?
The situation is a bit better if you offer payments in advance for specific features, as per the Wall Street Performer Protocol or IBM's patent.
In most countries, sharing even one copy of a file over a p2p network is not fair dealing/fair use. Over IRC, however, if you know the person at the other end, it might be.
The story in Canada is different. The law at this precise moment appears to be that almost all p2p sharing is legal because of the levies you pay on blank CDs, iPods, etc.
Your government is about to change that and make file sharing illegal, unless you do something about it.
Not necessarily. At some point, they're going to start using tamper resistant hardware. Good luck mod-chipping that. The only hole you can actually count on, is digital-to-analogue-to-digital conversion, to get data out of your "trusted" box and into an ordinary computer. That might be fine for ebook pirates, but it's going to be one hell of a PITA for daily life.
Rather than planning to spend tens or hundreds of hours on work-arounds when this happens, it would make sense to donate tens or hundreds of hours of your wages to the EFF or similar organisations which are working to prevent this problem from prevailing in the first place.
Sure, I can have a look, although I can't guarantee that I'll be able to make any genuinely helpful observations :).
Yes, there is some misunderstanding, but I think it stems from the underlying copyright law, rather than the GPL. A license is a grant of permission; there's nothing in the GPL itself which stops you from granting other licenses (further permission) if you own copyright to a work.
The situation is somewhat complicated, because the law says that if you make derivative work (a modified version of someone else's code), they own most of the copyright (only "most" of it because you retain the exclusive right of derivation). But in the case of a java program, it's pretty hard to argue that it's a derivative of object.java, unless you're distributing it as one big linked-together system. So the GPL doesn't have power over your code when it's on its own. You're required to grant GPL permissions for the whole system; but if you're using some permissive license like X11 or BSD, you've already done that.
This section explains that you can ONLY dual license if it is done at the beginning. You can not later choose to add another license to the already-GPL'd program.
I can't find any claim like that on the other end of that link! It's always possible to move from GPL-only to dual licensing, but you need permission from all of the contributors, so if you leave it too late it can be very hard.
And, it seems, you got the licensing wrong with Cygwin. It's perfectly allowable to link your LGPL app to a GPL library (the GPL is satisfied because any developer could take your code, and the cygwin library, and run off to build some larger GPLed app with it).
If your employers are scared of the GPL like that, it's probably because they don't understand it. Many of the biggest players in the industry are quite happy to use the GPL for some of their code -- IBM, Sun, SGI, Novell, and RedHat all spring to mind.
It's not that you can "ONLY make GPL software" -- it's just that you have to allow a GPLed version (this is called being "GPL compatible"). The GPL never stops you from allowing others to do things with code you own the copyright to. So in our example, if you write a public domain library on top of object.java, you can let 3rd parties write proprietary apps with it; they just have to make sure that they have met the licensing requirements for any java stuff they use (perhaps by paying Sun a license fee, or using someone else's version of Java).
I think you have an oversimplified notion of benefit to the community. Community benefit is a complicated function of what code gets written, how much money developers are able to earn to support their coding, and how code is openly available to be improved or incorporated in new projects. Dual GPL/proprietary licensing (as used by, for example MySQL or QT) allows some developers to simultaneously earn revenue and produce free software. GPLed libraries encourage people who are tossing up between open or closed source development, to take the open route, because it's cheaper -- you can use more of the existing tools. If your management doesn't like the GPL, it shouldn't be too hard to get them to cough up the fee for the alternative license.
If you still own the copyright to your original version, you can still use it. The fact that there's a GPL licensed version floating around on the net does not prevent you from licensing it to BeOS on different terms.
To earn some cash to fund development of the language, perhaps? The idea is something like: those who write proprietary apps contribute to the common good in cash; those who write free software contribute in kind.
If just the Object class was GPL'd, no one could write truly free open-source (say, BSD) java programs.
Why not? The modified BSD license is GPL compatible.
"much of it"... except the JVM, perhaps?
GPLing the JVM would be a very wise choice for Sun, for exactly the "business unfriendly" reasons you cite. Java would get the benefits of widespread adoption by the Free Software community; at the same time, if firms want to write proprietary apps against it, they'd have to pay for a license, thus preserving healthy revenue streams for Sun.