the check is payable to Boies, Schiller, & Flexner, LLP
Don't be so sure... SCO and Novell still have an outstanding dispute over who owns new royalties on the UNIX code base. Since this is a new customer, it's probably SCO's (or whoever they give it to), but that's not certain at all.
So, you're wondering where the paycheck comes from...
First off, the paycheck is the paycheck. Don't blame the software for not BEING the paycheck or for not magically producing one. If you had a solid understanding of proprietary software products like Oracle, Java, C# and MS visual developer bondange 3.0, you'd still have just as much trouble finding work in this economy. The question is, are there people out there that consider open source contribution to be a valuable bullet on someone's resume.
For our part, I can say that my company has hired at least 3 people PURELY on the basis of their open source contributions and for many others it was a major consideration. When you see that someone can produce good code, and they manage a project well, you hire them.
There's a dark side of this, though. Companies like mine are no longer in the dark, having to look at canned code samples and trust our rushed interview process. We're looking at the body of work that you produced, and if it's junk... well, you can imagine the result. Good people will profit from open source... the rest will have to content themselves with hiding in the relative obscurity of proprietary code.
No one is saying that this is a terrible imposition and we can't bear the weight of editing a text file.
What we're saying is that you're not ALLOWED by the GPL to derive from a GPLed work in such a way that would impose this restriction. What that means is that you cannot let GPLed code get anywhere near XFree86 4.4 in any way that could be considered derivation. This is a tragic departure from the X Window System's MIT-originated history of BSD-like open source licensing that dates back to the mid-80s.
Why? Because XFree felt they needed more credit? Frankly that's as bad as Stallman going around insisting that people call operating systems that use his tool-chain "GNU/whatever", and even Stallman has never stooped to enforcing that in a license.
"My position is that if you write/own the code you get to say how it's used"
That can be your position if you like, but it's not the case.
If you write/own the code you get to say how it's NOT used. How it IS used is entirely up to distributions and individuals who choose to use it on their own, and in this case, I applaud those who choose to steer clear. It's not that I'm a rabid GPL vs whatever proponent, it's just that I look at the landscape of licensing and I don't see any good reason to invent yet another variant in XFree's move. The MPL was a good example of when you have to roll your own. They had certain contractual reasons that neither the GPL nor the BSD license was appropriate.
XFree just seems to want to "make it their own", and while they are free to do that, we're also free to take the old source base and walk away.
Like or dislike the GPL, but you really have to face the fact that it's the most commonly used open source-compatible license (pre-dating that term, of course), and being fundamentally incompatible with it makes you a very questionable choice for an open source platform.
Moore is VERY prolific, and just about anything he has done will have a certain depth of field that makes it worth exploring. I recommend:
Supreme -- It's the disilation of everything "Superman", but with a distinctly self-aware narative and art.
Top 10 -- Comics of done the multiverse thing to death, and it's the standard explanation for a world "just like ours", but with super-heroes. Moore speculates on what the extreme endpoint of this sort of fantasy story-telling is by describing a world where EVERYONE (and their pets) are supers. It then follows the pain that this causes for one particular metropolitan police dept.
Tom Strong -- The 50s all over again. This is the sense-of-wonder storytelling you thought was dead.
V for Vendetta -- Not to everyone's liking, but if you have never had anarchy as a political point of view explained to you, it suits as a starter. Don't try to take it as the ONLY viewpoint on anarchy however.
What he does best is re-interpret various aspects of the comics genre to write his own stories that feel new and interesting, even though you know all the players (in Watchmen, for example, he was simply pulling from golden age heroes that DC had just acquired rights for).
Ellis does this same sort of thing in Authority and Planetary, and to a lesser extent in the non-super-heroic Transmetropolitan which is brilliant, and you should read it ASAP!
The problem with security through obsucurity is fairly simple: wide-scale kiddie attacks are just noise in the system. Granted, a lot of noise is annoying, potentially even crippling, but it's not the primary concern.
What you should REALLY worry about is the folks that say, "if I steal corporate secrets from [pick a large tech company] I can make $10 million selling them to the highest bidder in [pick a country that has emerging tech], thus it is worth my time to spend $1 million on security bug discovery." Now you have a whole other ball-game. Here, Microsoft cannot hide behind the veil of publicly reported attacks because these sorts of intrusions will be as stealthy as possbile and if the work, no one will ever know.
Thus, you have to look at how many vulnerabilities there were, say, last year and extrapolate how many people will have available to them to perform such attacks.
Open source on the other hand, contends that not only are the fewer exploits on the whole, but YOU have the source code, and can analyze it yourself and/or fix it if you find problems. When you're a huge corporation that can be a life-or-death difference because you are a very juicy target.
I may be woefully behind the times here, but last I had heard from the mailing list was that CMYK color-space management and color-matching could be done in nieve ways that were not encumbered, but if you wanted to, for example, match the same CMYK handling that print shops will use, you were SOL.
The only problem is finding the "why didn't I think of that" type obviousness and distinguishing it from the "everyone and his brother knows that" type of obviousness, but that's the heart of patent examination, and I don't think that the PTO is incopetent at doing that core job.
Yeah, I have to agree with Australia. A patents that makes 3 claims of which all three have prior art can't really be considered patentable just because no one has put those 3 particular beans in the same bag before.
The GUI is in fact, "all new" in the very real sense that the code is wildly different. If you use it for a while, you'll see that while the new Gimp retains much of the initial impression that the old Gimp did, this is clearly a Wilbur of a different color.
Much of the UI was re-written and much of the internals were made into a library. Gimp 2.0 also boasts a much improved plugin mechanism and real support for many of the structural things that people had been asking for, and without which modern features of photo-editing software could never be added. Does this mean that Gimp 2.0 is the be-all, end-all? Not even close, but it does mean that Gimp 2.0 is on a much more viable development path than 1.anything was.
Well yes, business model patents and others (though business models are the worst offenders) where the claim is, "X is an established thing or process, but I propose to do it in conjunction with Y!" are absurd.
It reminds me of something that an executive producer once said (I think it was JMS talking about his time on Murder She Wrote, where he was not producing, but had some authority as a writer, perhaps story editor or some such). Someone came to him with a "story idea". He asked for details and the guy said, "Amnisia". He replied, "ok, what's your idea?" They guy was confused and said again, "Amnisia" as if that were, in and of itself, a story idea. The pitch, needless to say, was over.
Patents are sort of the same. People are used to throwing noise at the patent office like "Widget on the Internet" and they really seem to expect that this noise is worth someone's valuable time to review. Is it going to become necessary to establish a reputation system? Should people be charged more to submit based on how likely it is that their patent will be a joke?
As I understand it (and I guess I should point out that I *don't* understand it), the claims can fall into different sorts of categories. Some are central to the patent, and some are more background for it. I would say that if any claim that your patent could be used to lock a technology has substantial, well published prior art that has been around for 10 or more years, then it's just a bad patent.
The problem seems to me to be this idea that patent attorneys have that the goal of a patent submission is to test the waters of what you can get, and then revise downward. Why not put the burden on the submitter to make sure that they claims are reasonable and make it worth their while to be overly narrow in order to ensure that it doesn't get rejected?
The claims presented in this application will likely be signifigantly different if it becomes allowed.
While that's true, I wonder if the easiest way to fix the patent system or at least to significantly band-aid it would be to permanently reject any patent for which initial submission has obvious prior art that dates back more than 10 years prior to the claim. At the very least that would eliminate a ton of the claims like this without having to waste everyone's time in cycling processes of reivew an revision.
You're looking at the entry point to solving the problem, not the whole solution. Once you can trust the sender's identity, you can START to establish trust relationships with senders.
As senders gain reputation, spam filtering begins to approach 100% effectiveness....
Actually, that would be a viable second step, but in order to OCR you're going to have to start with some sense of where the characters are on the document, so assembling and scale/perspecitive correcting the document as an image is a valid first step.
Why do you assume that a statement about the present indicates anything about the future?
I was responding to a post about how tyrannical the US is, and my point was that if you look at how we behave in terms of our own people and our neighbors, we're not what I would call tyrannical relatively speaking, and I don't think that its fair to say that such weapons as are discussed in the paper from the Air Force are any worse off in our hands than in many others'.
Would I rather see them under the control of the Swiss or Canada? Perhaps, but the US has shown that over the long haul they can remain in control of devestating military advantage and a) not destroy the world and b) not destroy ourselves and c) not expand our territory as a result. Compared to what many countries would do with such power, that's not too bad.
Now that said, our campaign of supression of all other countries on our hemisphere has been rather brutal (even true for Canada, though there we've stuck to mostly economic brutality). So, don't think I'm saying we're saints, just that the post I replied to was way off base.
"All it does it remove the mask of civility and democracy from what is ***IMHO*** an increasingly tyrannical power."
I hate to pick on a specific foreign power, but I'm going to here. My appologies to those it offends.
China is going to space. What scenario would you prefer: that China have a sole lock on military power in Earth orbit or that we share that (potential) battlefield?
I'd much rather that there be a stand-off. Heck, I'd prefer that over having the US there alone.
And as for the US being a tyrannical power... heh, you clearly have never seen tyranny. Yeah, the US isn't exectly the good guy. What we've done to Central and South America are pretty awful, for example, but compare that to the rest of the world, and I would say we're better neighbors than about 1/3 of the world, and better to our own citizens than about 2/3 of the world.
Should we be better? Hell yeah, but that's not the definition of a tyranny.
Yeah, the article is a very well-crafted case against this guy's credibility. I would be shocked if they wanted to actually take this one to court. The hope on the part of Infinium must have been that this pissant little website would bow under legal threat and retract, regardless.
I consider them to be poorly written documentation, and as such should be corrected by a tech writer who can read them, and then handed to a developer to code actual drivers.
This is also known as clean-room reverse engineering.
Open source drivers would be the best way to go. They provide the company with recognition and support from the rest of the community, but if they don't want to go that way, then we can do what we've always done: write the support ourselves. The binary only driver just makes our job a bit easier.
"Like it or not, the Bush Administration is responsible for those agencies and is therefore the most legitimate target for their complaints."
And complaints against an administration are inherently political.
Had this been an objective analysis of the trends in accuracty across several government bodies, it could have been apolitical, but as it is, it's an assessment of the administration.
Again, that's not to say it's a BAD report or that it should not have been aimed at Bush, but the fact that it IS that sort of report means that you should be looking at it very closely for bias. What's more it cannot possibly be comprehensive since the scope is so astoundingly large. That means there was a process of selection as to which facts to present... was it random? Based on likely impact? Based on level of innacuracy (that would be damning)?
If you apply the same selection criteria to other administrations, how would they fare? Why is that? Did the scope change? Are the agencies being more or less open about those facts now?
This is not cut-and-dry as both the publishers of the report and the administration would like you to think.
"Your reasoning - the scientists are releasing a scientific paper. Their conclusions have political ramifications. Therefore they are making a political statement"
That's not at all what he said. What he said was that it was released as a report on "The Bush Administration" rather than on the activities of the agencies that have outlasted and will outlast the Bush Administration like the DoD, the CDC, etc. This was a political statement aimed squarely at a particular administration.
That doesn't make it wrong, but it certainly does turn on my bias-detection filters...
In other words, this would work ok as an element of a larger system that could deal with the other 50%.
Interestingly, this is EXACLTY how SpamAssassin works, running a variety of testing engines from simple header-regexps to basian analysis to DNS-blacklist checks. SA is essentially just a framework for pulling diverse testing schemes together and using all of their input to evaluate mail. This technique seems like it would be a good addition.
Let's be clear (not that I disagree with you overall): BSD licensing gives your USERS absolute freedom within the constraints allowed by copyright law (they cannot, for example, put the work into the public domain or re-assign copyright to themselves).
The GPL on the other hand, is aimed at giving freedom to everyone at once and keeping it that way.
Stallman got the idea for the GPL because he couldn't get access to some printer drivers. So, you see, if those printer drivers had been BSD-licensed it would not have helped him, the printer company still might have withheld them. I'm not saying Stallman is right or wrong, just pointing out what the GPL is a tool for: making sure that your code remains available to everyone who wants to use it, regardless of who they may have gotten it from.
I've been thinking of coming up with a new sort of license based on the GPL, but aimed at a compromise position.
The idea would be that you can do whatever you like with the code, as long as you distribute the original code that you started with and a set of patches for any changes you made TO THAT CODE (not additions). Thus, if you modify main.c and add a new function, you have to share the diffs, and the diffs must be functional. If you link against a new library or include a new header file, you're under no duress to release it.
This means that you might add in a call to some new spiffy encryption module, but only provide a diff with the new function call and a stub library that does nothing. That would be valid.
Then again, the big advantage of the GPL is that it's got a lot of legal eyes on it, and it's most likely to be the one tested in court, not my variant.
Slashdot Mirror
the check is payable to Boies, Schiller, & Flexner, LLP
Don't be so sure... SCO and Novell still have an outstanding dispute over who owns new royalties on the UNIX code base. Since this is a new customer, it's probably SCO's (or whoever they give it to), but that's not certain at all.
So, you're wondering where the paycheck comes from...
First off, the paycheck is the paycheck. Don't blame the software for not BEING the paycheck or for not magically producing one. If you had a solid understanding of proprietary software products like Oracle, Java, C# and MS visual developer bondange 3.0, you'd still have just as much trouble finding work in this economy. The question is, are there people out there that consider open source contribution to be a valuable bullet on someone's resume.
For our part, I can say that my company has hired at least 3 people PURELY on the basis of their open source contributions and for many others it was a major consideration. When you see that someone can produce good code, and they manage a project well, you hire them.
There's a dark side of this, though. Companies like mine are no longer in the dark, having to look at canned code samples and trust our rushed interview process. We're looking at the body of work that you produced, and if it's junk... well, you can imagine the result. Good people will profit from open source... the rest will have to content themselves with hiding in the relative obscurity of proprietary code.
You have missed the point entirely.
No one is saying that this is a terrible imposition and we can't bear the weight of editing a text file.
What we're saying is that you're not ALLOWED by the GPL to derive from a GPLed work in such a way that would impose this restriction. What that means is that you cannot let GPLed code get anywhere near XFree86 4.4 in any way that could be considered derivation. This is a tragic departure from the X Window System's MIT-originated history of BSD-like open source licensing that dates back to the mid-80s.
Why? Because XFree felt they needed more credit? Frankly that's as bad as Stallman going around insisting that people call operating systems that use his tool-chain "GNU/whatever", and even Stallman has never stooped to enforcing that in a license.
"My position is that if you write/own the code you get to say how it's used"
That can be your position if you like, but it's not the case.
If you write/own the code you get to say how it's NOT used. How it IS used is entirely up to distributions and individuals who choose to use it on their own, and in this case, I applaud those who choose to steer clear. It's not that I'm a rabid GPL vs whatever proponent, it's just that I look at the landscape of licensing and I don't see any good reason to invent yet another variant in XFree's move. The MPL was a good example of when you have to roll your own. They had certain contractual reasons that neither the GPL nor the BSD license was appropriate.
XFree just seems to want to "make it their own", and while they are free to do that, we're also free to take the old source base and walk away.
Like or dislike the GPL, but you really have to face the fact that it's the most commonly used open source-compatible license (pre-dating that term, of course), and being fundamentally incompatible with it makes you a very questionable choice for an open source platform.
I wasn't mistaking any word for have, it was a simple typo, and get over yourself unless you're signing up to proof-read my Slashdot postings, gratis.
- Supreme -- It's the disilation of everything "Superman", but with a distinctly self-aware narative and art.
- Top 10 -- Comics of done the multiverse thing to death, and it's the standard explanation for a world "just like ours", but with super-heroes. Moore speculates on what the extreme endpoint of this sort of fantasy story-telling is by describing a world where EVERYONE (and their pets) are supers. It then follows the pain that this causes for one particular metropolitan police dept.
- Tom Strong -- The 50s all over again. This is the sense-of-wonder storytelling you thought was dead.
- V for Vendetta -- Not to everyone's liking, but if you have never had anarchy as a political point of view explained to you, it suits as a starter. Don't try to take it as the ONLY viewpoint on anarchy however.
What he does best is re-interpret various aspects of the comics genre to write his own stories that feel new and interesting, even though you know all the players (in Watchmen, for example, he was simply pulling from golden age heroes that DC had just acquired rights for).Ellis does this same sort of thing in Authority and Planetary, and to a lesser extent in the non-super-heroic Transmetropolitan which is brilliant, and you should read it ASAP!
The problem with security through obsucurity is fairly simple: wide-scale kiddie attacks are just noise in the system. Granted, a lot of noise is annoying, potentially even crippling, but it's not the primary concern.
What you should REALLY worry about is the folks that say, "if I steal corporate secrets from [pick a large tech company] I can make $10 million selling them to the highest bidder in [pick a country that has emerging tech], thus it is worth my time to spend $1 million on security bug discovery." Now you have a whole other ball-game. Here, Microsoft cannot hide behind the veil of publicly reported attacks because these sorts of intrusions will be as stealthy as possbile and if the work, no one will ever know.
Thus, you have to look at how many vulnerabilities there were, say, last year and extrapolate how many people will have available to them to perform such attacks.
Open source on the other hand, contends that not only are the fewer exploits on the whole, but YOU have the source code, and can analyze it yourself and/or fix it if you find problems. When you're a huge corporation that can be a life-or-death difference because you are a very juicy target.
I may be woefully behind the times here, but last I had heard from the mailing list was that CMYK color-space management and color-matching could be done in nieve ways that were not encumbered, but if you wanted to, for example, match the same CMYK handling that print shops will use, you were SOL.
Feel free to correct me.
That's smart.
The only problem is finding the "why didn't I think of that" type obviousness and distinguishing it from the "everyone and his brother knows that" type of obviousness, but that's the heart of patent examination, and I don't think that the PTO is incopetent at doing that core job.
Yeah, I have to agree with Australia. A patents that makes 3 claims of which all three have prior art can't really be considered patentable just because no one has put those 3 particular beans in the same bag before.
Correct, and until the patents expire on CMYK handling, that's all you're going to get in free software.
The GUI is in fact, "all new" in the very real sense that the code is wildly different. If you use it for a while, you'll see that while the new Gimp retains much of the initial impression that the old Gimp did, this is clearly a Wilbur of a different color.
Much of the UI was re-written and much of the internals were made into a library. Gimp 2.0 also boasts a much improved plugin mechanism and real support for many of the structural things that people had been asking for, and without which modern features of photo-editing software could never be added. Does this mean that Gimp 2.0 is the be-all, end-all? Not even close, but it does mean that Gimp 2.0 is on a much more viable development path than 1.anything was.
Well yes, business model patents and others (though business models are the worst offenders) where the claim is, "X is an established thing or process, but I propose to do it in conjunction with Y!" are absurd.
It reminds me of something that an executive producer once said (I think it was JMS talking about his time on Murder She Wrote, where he was not producing, but had some authority as a writer, perhaps story editor or some such). Someone came to him with a "story idea". He asked for details and the guy said, "Amnisia". He replied, "ok, what's your idea?" They guy was confused and said again, "Amnisia" as if that were, in and of itself, a story idea. The pitch, needless to say, was over.
Patents are sort of the same. People are used to throwing noise at the patent office like "Widget on the Internet" and they really seem to expect that this noise is worth someone's valuable time to review. Is it going to become necessary to establish a reputation system? Should people be charged more to submit based on how likely it is that their patent will be a joke?
As I understand it (and I guess I should point out that I *don't* understand it), the claims can fall into different sorts of categories. Some are central to the patent, and some are more background for it. I would say that if any claim that your patent could be used to lock a technology has substantial, well published prior art that has been around for 10 or more years, then it's just a bad patent.
The problem seems to me to be this idea that patent attorneys have that the goal of a patent submission is to test the waters of what you can get, and then revise downward. Why not put the burden on the submitter to make sure that they claims are reasonable and make it worth their while to be overly narrow in order to ensure that it doesn't get rejected?
The claims presented in this application will likely be signifigantly different if it becomes allowed.
While that's true, I wonder if the easiest way to fix the patent system or at least to significantly band-aid it would be to permanently reject any patent for which initial submission has obvious prior art that dates back more than 10 years prior to the claim. At the very least that would eliminate a ton of the claims like this without having to waste everyone's time in cycling processes of reivew an revision.
You're looking at the entry point to solving the problem, not the whole solution. Once you can trust the sender's identity, you can START to establish trust relationships with senders.
As senders gain reputation, spam filtering begins to approach 100% effectiveness....
Actually, that would be a viable second step, but in order to OCR you're going to have to start with some sense of where the characters are on the document, so assembling and scale/perspecitive correcting the document as an image is a valid first step.
Your suggestion had already been made. In my original post: "Should we be better? Hell yeah, but that's not the definition of a tyranny."
You were making my point for me.
Why do you assume that a statement about the present indicates anything about the future?
I was responding to a post about how tyrannical the US is, and my point was that if you look at how we behave in terms of our own people and our neighbors, we're not what I would call tyrannical relatively speaking, and I don't think that its fair to say that such weapons as are discussed in the paper from the Air Force are any worse off in our hands than in many others'.
Would I rather see them under the control of the Swiss or Canada? Perhaps, but the US has shown that over the long haul they can remain in control of devestating military advantage and a) not destroy the world and b) not destroy ourselves and c) not expand our territory as a result. Compared to what many countries would do with such power, that's not too bad.
Now that said, our campaign of supression of all other countries on our hemisphere has been rather brutal (even true for Canada, though there we've stuck to mostly economic brutality). So, don't think I'm saying we're saints, just that the post I replied to was way off base.
"All it does it remove the mask of civility and democracy from what is ***IMHO*** an increasingly tyrannical power."
I hate to pick on a specific foreign power, but I'm going to here. My appologies to those it offends.
China is going to space. What scenario would you prefer: that China have a sole lock on military power in Earth orbit or that we share that (potential) battlefield?
I'd much rather that there be a stand-off. Heck, I'd prefer that over having the US there alone.
And as for the US being a tyrannical power... heh, you clearly have never seen tyranny. Yeah, the US isn't exectly the good guy. What we've done to Central and South America are pretty awful, for example, but compare that to the rest of the world, and I would say we're better neighbors than about 1/3 of the world, and better to our own citizens than about 2/3 of the world.
Should we be better? Hell yeah, but that's not the definition of a tyranny.
Yeah, the article is a very well-crafted case against this guy's credibility. I would be shocked if they wanted to actually take this one to court. The hope on the part of Infinium must have been that this pissant little website would bow under legal threat and retract, regardless.
;-)
Oops
Binanry-only drivers are, IMHO, an ok thing.
I consider them to be poorly written documentation, and as such should be corrected by a tech writer who can read them, and then handed to a developer to code actual drivers.
This is also known as clean-room reverse engineering.
Open source drivers would be the best way to go. They provide the company with recognition and support from the rest of the community, but if they don't want to go that way, then we can do what we've always done: write the support ourselves. The binary only driver just makes our job a bit easier.
"Like it or not, the Bush Administration is responsible for those agencies and is therefore the most legitimate target for their complaints."
And complaints against an administration are inherently political.
Had this been an objective analysis of the trends in accuracty across several government bodies, it could have been apolitical, but as it is, it's an assessment of the administration.
Again, that's not to say it's a BAD report or that it should not have been aimed at Bush, but the fact that it IS that sort of report means that you should be looking at it very closely for bias. What's more it cannot possibly be comprehensive since the scope is so astoundingly large. That means there was a process of selection as to which facts to present... was it random? Based on likely impact? Based on level of innacuracy (that would be damning)?
If you apply the same selection criteria to other administrations, how would they fare? Why is that? Did the scope change? Are the agencies being more or less open about those facts now?
This is not cut-and-dry as both the publishers of the report and the administration would like you to think.
"Your reasoning - the scientists are releasing a scientific paper. Their conclusions have political ramifications. Therefore they are making a political statement"
That's not at all what he said. What he said was that it was released as a report on "The Bush Administration" rather than on the activities of the agencies that have outlasted and will outlast the Bush Administration like the DoD, the CDC, etc. This was a political statement aimed squarely at a particular administration.
That doesn't make it wrong, but it certainly does turn on my bias-detection filters...
In other words, this would work ok as an element of a larger system that could deal with the other 50%.
Interestingly, this is EXACLTY how SpamAssassin works, running a variety of testing engines from simple header-regexps to basian analysis to DNS-blacklist checks. SA is essentially just a framework for pulling diverse testing schemes together and using all of their input to evaluate mail. This technique seems like it would be a good addition.
"BSD licensing gives you absolute freedom"
Let's be clear (not that I disagree with you overall): BSD licensing gives your USERS absolute freedom within the constraints allowed by copyright law (they cannot, for example, put the work into the public domain or re-assign copyright to themselves).
The GPL on the other hand, is aimed at giving freedom to everyone at once and keeping it that way.
Stallman got the idea for the GPL because he couldn't get access to some printer drivers. So, you see, if those printer drivers had been BSD-licensed it would not have helped him, the printer company still might have withheld them. I'm not saying Stallman is right or wrong, just pointing out what the GPL is a tool for: making sure that your code remains available to everyone who wants to use it, regardless of who they may have gotten it from.
I've been thinking of coming up with a new sort of license based on the GPL, but aimed at a compromise position.
The idea would be that you can do whatever you like with the code, as long as you distribute the original code that you started with and a set of patches for any changes you made TO THAT CODE (not additions). Thus, if you modify main.c and add a new function, you have to share the diffs, and the diffs must be functional. If you link against a new library or include a new header file, you're under no duress to release it.
This means that you might add in a call to some new spiffy encryption module, but only provide a diff with the new function call and a stub library that does nothing. That would be valid.
Then again, the big advantage of the GPL is that it's got a lot of legal eyes on it, and it's most likely to be the one tested in court, not my variant.