be in no doubt that manned space flight is a use it or loose it technology.
No, historically it's been a "develop from scratch in 8 years" technology. Not all that long, considering. And the cost to reconstruct the ability would be less than the first time, because so much documentation has been stored, and the unknowns have already been solved.
There is no prospect for launching any useful manned space flight in the next 10 years, just like there hasn't been one for 17. Sitting it out until we actually have an important objective is the best strategy (especially since many of the core skills have already been lost). An example of such a good objective would be to follow up a permanent (solar or nuculear powered) Mars rover with a manned crew.
Hey, in 15-25 years you could even get China or the EU to handle the occasional manned launch- either by multinational cooperation, technology-transfer-treaties, or simply poaching engineering staff. But the proud US isn't willing to depend on others for something.
The only reason manned spaceflights have continued for the past 10 years is a matter of advertising. "Use it, or admit you forgot how to use it, or that it wasn't a good idea in the first place" technology.
technical documentation is so often written to be written, not to be read
Then that is a failure of the technical organization, and they should be held accountable. Creating proper documents should occupy only about 1% of the NASA budget for one year. (Or, just double the cost of a single shuttle launch by assigning every workers a "shadow" to record his movements). Hey, these guys don't have any flying to do right now- lets get them to document!
Keeping the shuttles running does depend on having the same guys who built it always on hand to make sure it's working. But that's a product of the shuttle program's over-complexification. Titan IV rockets weren't so difficult to explain to areospace experts.
To people who actually use the software it's simply a tool.
The people who pay for software writing think it's highly valuable. When a CEO of some large, non-software company gets 3-5 programmers working on modifying their database frontends, he tends to get an inflated idea of the value of that software. Even though it's probably nothing special, he'll see it as an advantage over his competitors, and get a smidgen of Microsoft-feeling his possessiveness of it.
If people strongly believed in supporting Microsoft's business model there wouldn't be so much "piracy" of their software..
If people believed less strongly in supporting it, then "pirates" wouldn't risk arrest.
Apparently, when voting on copyright law, the majority of US citizens allow themselves to be convinced that (software) copyrights are a good idea. But in the privacy of offices and especially homes, many feel they don't really need to obey it, even though they were unable to present a rational argument against the law.
It's a little like the Probibition amendment. A big majority decided ingesting alcohol was evil, but yet, a majority continued to ingest alcohol.
Major misalignments between the popular and legal acceptability of an action are always destructive. Widespread disregarding of a law promotes other kinds of lawlessness, sets the stage for capricious show-trials, and worst of all, prevents the disagreeable law from being changed.
If software copyright violation was punished more universally, then the public would either change those laws, or embrace GPL'd programs. (Either way, accomplishing the FSF goal)
That's why Microsoft doesn't want all "pirates" to be caught and punished yet- as long as there is still an alternative development model out there, they don't want their customers to comprehend what it would mean to pay the price on every, single copy.
To most of these end-users, having or not having access to the source code is moot so why push something on them that they probably don't even want?
That is a fundamental, fundamental aspect of the GPL which RMS has expounded on many times.
In short, every end-user should have the power to select a programmer on the free market to improve the software he uses. If they don't want it, they simply aren't aware of the possibilities.
(The specific example of a military Tank Commander is one of the most extreme cases where the user wouldn't want to modify the code, because it'll be tied into painfully over-engineered software. Almost every other situation, including military users in R&D or training, will have greater utility from source code access. The X-Box in particular is someplace where users would love the source code.)
But the IT department has that right as agents for the corporation
According to the text of relevant laws and licenses, they have rights as individuals as well.
Corporations don't like that idea, though, and that is one reason why Bill Gates has been so successful in convincing them that the GPL is viral and evil. Gates often spreads misdirection (the word "viral" is inaccurate), but he can speak the truth too: when he tells CEOs that if their programmers modify GPLed code, the corporation might not be able to keep the changes from escaping to the public, he's right.
What he doesn't tell them, of course, is that public release will usually be non-harmful or even beneficial. (Letting that slip undermines his business model)
But that's just it, my employer does not "give" me the binaries.
They do "give" it to you (English definition), just like they "give" you many tools, although neither employee nor boss expects him to keep permanent possession- but that's not the point. Office furniture, computer hardware- they're physical goods. Whoever bought them can transfer them wherever he wants, but there's no law against copying them, since that's impossible.
When giving software to employees, however, the employer will have committed an act of copyright violation, unless he abides by the license. And if the license is the GPL, that means that anyone with access to the binaries can freely spread them far and wide.
Imagine if someone from the corporate IT department is called to the witness stand:
"Fred, did you on the date of March 1st, distribute or cause to be distributed to your co-employees Carol, Betty, and Alice, the software labelled exhibit A, which you have already admitted contains substantial portions of the plaintiff's source code?"
Fred cannot honestly deny he "distributed" the code. Then the only question is, was he violating copyright when he did so? Or, in doing so, had he consented to the GPL? In the first case, "Go directly to punitive damages". In the second case, Carol, Betty, and Alice are free to pass out further GPLed copies to the 4 corners of the earth.
The software installed on the computer remains the property of my employer
It was never the property of the employer. It belongs to the author and copyright holder the entire time.
The employer had permission to use it under certain conditions. If the software was GPL, those conditions included never restricting anyone from sharing copies of the software.
And I'm envisioning a kiosk that would be set up in a public place to say give directions to anyone who wishes to use it. By the definition of user you are attempting to assert, I would.
No, because the kiosk user doesn't have access to the filesystem. An employee of a corporation does. Maybe not the end-user who runs the program each day, but the IT dept who installed it certainly does.
And anyway, the FSF is considering extending the GPL to ensure that the user is entitled to the source code even if he's using a program running on someone else's computer. That won't matter for a while yet, though. (An example of a license that moves towards this goal is here)
but I think the same case law that allows a software publisher to go after a business that has illegal copies of software installed on the businesses systems governs here.
That is a more general thing, and relates to the definition of incorporation. If a crime is committed by a person acting on behalf of a corporation, liablity will be assessed against the corp. as a whole, rather than against any person within it.
But, laws like that only shift around who gets punished for any particular offense. It doesn't change what things are illegal for a corporation to do, which includes everything a private citizen can't do, and more.
So in my example of Fred above, if he turns out to be guilty, the corporation will pay the damages, if he was acting under their orders.
Oh, how can you tell? Some lawyers work for my corporation, although I don't like to duplicate their speaking style exactly.
As contracted employees, they are not independent entities.
Yes they are. The idea that members of a corporation somehow become one big person is a legal fiction. Certain laws treat a corporation as a single person. Copyright law does not.
A license between an author and a customer can treat corporations specially. The GPL does not.
I think the Adobe comparison is bogus; with the GPL you can make multiple copies for your own use; with a commercial licence typically you can't.
In both cases, the only thing stopping you from making copies is copyright law. If the author permits you to make copies, you can. Adobe will give you this permission if you agree to give them some amount of dollars. GPL authors give you this permission if you agree to allow the recipient to make unlimited further copies.
Copyright law is the threat with which authors force you to agree to a license before duplicating their work. Some authors ask for money, others ask that you permit your changes to be shared freely. Whatever the price is, if you don't agree, you can't copy the software.
If a corporation isn't willing to allow it's employees to pass out GPLed software they've modified, then it must tell them not to modify GPL software.
Most of the stuff I worked on had little or no demand outside of the governement/DoD world.
That's still true today, I'm sure that a major reason that more DoD code doesn't get leaked out to the public is that it's unfriendly, difficult to use (especially on consumer hardware), and plain-old boring.
from a licensing perspective, the end user is the acquiring agency;
That is how many software licenses are written, but it's not how the GPL is written. The GPL doesn't make any specific mention of corporations having special status. Nor does US copyright law give corporations (or agencies, or other kinds of organizations) special rights as a user of copyright (they have a few differences as a holder of copyright, though).
When the government wants software from Microsoft, they can negotiate a Volume License. There is no equivalent to a "Volume GPL", though. The GPL makes no mention of "groups", "companies", or "sites", so each individual person is the same as any other.
So, assuming a government agency recieved a modified GPL program as a deliverable. It'll have the GPL still attached, and each time they distribute it (to one of their military end-users), they'll have to abide by the GPL, or be in violation of copyright law.
That is why, I believe, contractors so far do not use GPLed code as the basis for deliverables- the government wouldn't like abiding by that license once they'd recieved the end product.
That would be like every cash register in a Linux based point of sale system coming with its own set of source CDs.
That's an interesting question, and one I don't see as 100% resolved yet. It comes down to the meaning of "give the binaries to" (because anyone with the binaries can demand the source). Does a person "have" the binaries, if they're embedded in a device whose filesystem he can't access?
Then, what about the related case of a consumer-product (like an MP3 player) having a GPL program embedded in the firmware? Is that end-user entitled to the source code? From watching RMS, it seems he wants the answer to be yes. (Of course, this is a little different from a cash register or battle tank, as the user owns that hardware)
This answer which you cited last week seems pretty clear to me.
Unfortunately, that answer solves nothing. It is the contents of a FAQ on GNU's webpage, and it doesn't matter, legally. The only things which really matter are the text of the GPL, and the laws of the jurisdiction.
"Distribution" has a meaning in the English language. "To pass out". Giving the software to other members of a company means "passing it out". And before it can be passed out, someone's got to make some copies. "Reproduce it".
"Reproducing" or "distributing" a copyrighted work (which all GPLed programs are), is illegal. Unless the copyright holder has given you permission. It is one of the actions that are "prohibited by law if you do not accept this License." (GPLv2 section 5)
So, if one employee gives the program to another, he must do so under the GPL. Which means granting the reciever full permission to give further copies to whoever he wants.
A comparison: Can an employee of a company make a copy of Adobe Photoshop for another member of the company? Then why could he make a copy of Gimp, without special permission from the author.
Now, if the FSF had intended to permit corporations to alter GPL code, and give it to all their employees, but under the threat of termination and lawsuit if they passed it out to anyone else, then they should've stated this in the license.
The vague FAQ entry might amount to a promise from the FSF that they won't procescute a corporation for doing that, but other users of the GPL (like Linus Torvalds) will not necessarily agree with that interpretation.
(Note that the following discussion assumes the military contractor will be modifying GPLed source code. The fact that they will be "writing software on Linux" doesn't necessarily mean they'll do that)
I'm not sure where one draws the line between the rights of the programmer and the rights of the corporation (s)he works for.
Well, an employee should feel free to do anything his employer gave him permission to do.
When the corporation (or another employee, acting on the corporation's behalf) gives him a copy of a modified, GPL program, it gives him permission to redistribute it to whoever he wants.
reasonable grounds for dismissal
If that ever happened, it would be proof that the corporation is violating the GPL. It would violate this line of section 6: "You may not impose any further restrictions on the recipients' exercise of the rights granted herein."
Once they impose a restriction on redistribution to any user of the program, they no longer have rights to modify or distribute that program.
These two scenarios really don't sit very happily together.
No, they don't sit well. It brings to mind a giant GPL loophole:
Modify a GPL program for commercial sale
Require each customer to "join" your company before ordering the product. (The workload is nothing, and the salary is $1/year, much less than the product cost)
Profit
Why can't this happen? Because the assumption that "one corporation is a single entity for purposes of the GPL" is incorrect. A company cannot modify a GPL program for "internal use only", unless they're willing to take the risk that a random employee will post it to USENET's comp.source, and they'll have no legal recourse.
I discussed this in some detail several times, like just last week (when, for a strange reason, I forgot to log in).
The short version is "Microsoft doesn't consider a corporation to be one person for software licensing purposes, so why should the GPL?"
When you develop software under governement contract, the governement owns the code.
It all depends on the contract, of course. But today, the default seems to be that both the customer and the contractor own the code. Either can use it.
I've seen this in a few big Pentagon software projects- the contract is fed out to multiple development companies. Each contributes only a part of the whole. The government owns the result, and can do whatever it likes with it- but, each contractor can keep the sections they individually wrote, and re-sell them if possible.
(Supposedly, it's cheaper to hire the developers if they think there's a chance they can find another customer later. Of course, the nature of the product, and how dangerous it is to share, should dictate the government's choice.)
Funny, I was speaking to a military Chemical/Biological weapons specialist back in 2001, and he claimed that never in history had any military knowingly released CB weapons on it's own troops.
(Of course, maybe there's a trick of words in here. Like, "When they started to lose, they were no longer my troops")
Source and data are different things. For example, Google.com has released the patent ("source code") to their page-ranking system. But the finely-tuned numeric constants it uses to produce good, weighted results are still secret. So competitors don't really knnow the full capabilities.
In terms of the military, you can break software into 2 categories for purposes of your concern:
Field deployed firmware: Stuff that runs a radar or a missile. This is hand-crafted at a low, assembly level, and is utterly dependent on the hardware. Getting the source without having the hardware (or knowing it's exact specifications), is of little use. So far, code like this has been so tiny, and has so much effort put into it, that there's no incentive to reuse existing things very much (at least, without completely purchasing it first)
Research, training, and planning: More like traditional desktop programs, but often running on a big SGI or something. Like simulators that predict how tight a missile can track, or how far a radar can see with a certain scan pattern. For code like this, the data files which describe the actual, secret performance values ("Exactly how far can an AMRAM shoot?") are kept separate from the source code. This way, programmers and analysts who aren't cleared to see the secret data can still work on/with the software. And then once the bugs are worked out and the UI is pretty, the Generals can plug in their real data and go to work planning their wars.
Even though the Pentagon hasn't so far made any major software Open Source, they have been acting as if large numbers of un-cleared eyeballs would see their source code. (Programmers are expensive enough, without putting them through FBI background checks. Why, how do you expect a Pakistani H1B to pass those kids of loyalty tests?)
Also, I don't think they would be making major modifications to the Kernel.
So far, the DoD guys working on FCS (which is still only a raw, raw prototype) use unmodified kernel source code, with no new modules added. They do however recompile their kernels with a specific config file, to improve some kind of networking support.
The distributions they use are Red Hat or Familiar, depending on the hardware (although, so far PDAs only look like a gimmick). For a while they used Progeny (debian) linux too.
The DOD fork of Linux won't ever make it back onto the main 'branch.'
That will be up to Linus, because one of the software engineers working for Boeing will offer to send him the patches.
Why? Because they'll have a big team of programmers working on Linux, at least a few of them will feel favorable towards the public effort and want to send their changes out.
And since those programmers are sharing code amoung themselves, all of them have it under the GPL- so their managers have no legal means to prevent them from redistributing it.
On the other hand, it is unlikely that the DoD will truely fund Boeing or someone to fork linux. They'll pick RTLinux's or some other company's fork to use, possibly giving them a few additional requirements. That company will continue to release it's code the way it always has.
This is very difficult for most users- as most people are using one of the "pre-configured" X11 desktops like KDE/Gnome, rather than "rolling their own". And the behavior of "shifting a running window from one terminal to another" is so rare, that none of the major desktops have felt a need to support it in their GUIs.
To accomplish this task, you can use either X11 or VNC. Using VNC, it works the same as with any other VNC server (on Windows, Mac, Linux, whatever)- a VNC client can connect and reconnect multiple times, with the desktop programs still running.
To do it in X11... well, I've never done this. But I can give you a clue: xmove will make an intermediary Xserver between the application and the "real" Xserver, allowing the application's windows to be sent to other terminals while it's still running. (Debian people can apt-get xmove)
You can't grab the video memory if the OS won't let you.
If major OSes start to ban printing, screencap, and clipboarding, and they make it impossible for the OS to be hacked into, then serious information-workers will demand a workaround.
Fast, high-quality Firewire cameras are common and inexpensive. They are capable of capturing text from a CRT in quite a legible form (especially if larger than 16 point). An OCR on another computer (or a different process on the same one) can change that image into editable text in real time.
Of course, it will be many years before there's any need for such a thing. By that time, everyone will have small PDAs with cameras and fast processors. The user just installs an OCR program, then waves the PDA over a desk or monitor and suddenly has PDFs of everything it saw.
Of course, then there will be a push to make this software illegal, which is in the end all you can do.
(A connection to the VGA cable would be much more reliable in the near future, although eventually display hardware may use an encrypted signal. But the emissions from the monitor will always be there)
Theoretically, a person could buy Microsoft Windows XP in a store, and bring that cdrom with him as he goes to examine MS's code. After getting the source compiled, he can start comparing the binaries with those that are publically distributed.
(Of course, there are so many service packs and things that even if the comparison shows a mismatch, MS can claim it doesn't indicate anything worrisome)
That is very, very intentional. It's more important when it happens with file extensions than actual programs, though.
Today, it is no longer possible to tell someone "Send me a bitmap file from the document", without him assuming you mean the output of MS Paint and MS Word, respectively.
through diligent study you may be able to determine whether or not you are in violation of some particular software patent
Moreover, if you do attempt to keep aware of software patents, then in the event you are found guilty of violating one, the punishment will be worse, as it was willful, knowing infringement.
how to 'map' cyberspace. In the paper we suggested that it should be mapped like a giant Venn diagram
That's an obvious idea, people have been envisioning such things since back with Vannevar Bush. As you discovered, there are many more important factors than simply having a good idea. 99% perspiration indeed.
You don't always need a working implementation for prior art (especially in the field of software, where the distinction between the design and the actual product is vague). The number of joke patents issued for science-fiction ideas attests to this.
But you do need to have released the description to the public.
Submitting a paper to a teacher for grading is specifically not adequate. Court cases on that exact basis have already been lost.
About the only saving grace in the patent system is the fact that they still expire in less than a human lifetime. Funny that Congress has seen fit to quintuple the length of copyrights, while leaving patents unextended. (Hope I don't give anyone ideas)
behind in aviation development at the start of world war I (as documented by none other than the US federal government itself)
That's one of my favorite funny stories. The irony of the heroic status given to Orville and Wilbur, in comparison to the money they were denied, calls the "capitalist" nature of the US into question. (US Aviation development wasn't degraded as much as it might've been, because the government took away the Wrights' patent while it still had more than 5 years left.)
Just try to imagine the "fair market value" for a concept like "the airplane". Tack on the price of victory in WW1 and WW2, and of beating the Soviets to the moon. Then scale all that up from 1903 to 2003 currency. We're looking at one trillion dollars.
(Of course, back then situations were different. Patents didn't apply internationally, so there were a different set of problems. An inventor in Britain would create something, and then US companies would market it, while other Britons are barred)
Slashdot Mirror
be in no doubt that manned space flight is a use it or loose it technology.
No, historically it's been a "develop from scratch in 8 years" technology. Not all that long, considering. And the cost to reconstruct the ability would be less than the first time, because so much documentation has been stored, and the unknowns have already been solved.
There is no prospect for launching any useful manned space flight in the next 10 years, just like there hasn't been one for 17. Sitting it out until we actually have an important objective is the best strategy (especially since many of the core skills have already been lost). An example of such a good objective would be to follow up a permanent (solar or nuculear powered) Mars rover with a manned crew.
Hey, in 15-25 years you could even get China or the EU to handle the occasional manned launch- either by multinational cooperation, technology-transfer-treaties, or simply poaching engineering staff. But the proud US isn't willing to depend on others for something.
The only reason manned spaceflights have continued for the past 10 years is a matter of advertising. "Use it, or admit you forgot how to use it, or that it wasn't a good idea in the first place" technology.
technical documentation is so often written to be written, not to be read
Then that is a failure of the technical organization, and they should be held accountable. Creating proper documents should occupy only about 1% of the NASA budget for one year. (Or, just double the cost of a single shuttle launch by assigning every workers a "shadow" to record his movements). Hey, these guys don't have any flying to do right now- lets get them to document!
Keeping the shuttles running does depend on having the same guys who built it always on hand to make sure it's working. But that's a product of the shuttle program's over-complexification. Titan IV rockets weren't so difficult to explain to areospace experts.
To people who actually use the software it's simply a tool.
The people who pay for software writing think it's highly valuable. When a CEO of some large, non-software company gets 3-5 programmers working on modifying their database frontends, he tends to get an inflated idea of the value of that software. Even though it's probably nothing special, he'll see it as an advantage over his competitors, and get a smidgen of Microsoft-feeling his possessiveness of it.
If people strongly believed in supporting Microsoft's business model there wouldn't be so much "piracy" of their software..
If people believed less strongly in supporting it, then "pirates" wouldn't risk arrest.
Apparently, when voting on copyright law, the majority of US citizens allow themselves to be convinced that (software) copyrights are a good idea. But in the privacy of offices and especially homes, many feel they don't really need to obey it, even though they were unable to present a rational argument against the law.
It's a little like the Probibition amendment. A big majority decided ingesting alcohol was evil, but yet, a majority continued to ingest alcohol.
Major misalignments between the popular and legal acceptability of an action are always destructive. Widespread disregarding of a law promotes other kinds of lawlessness, sets the stage for capricious show-trials, and worst of all, prevents the disagreeable law from being changed.
If software copyright violation was punished more universally, then the public would either change those laws, or embrace GPL'd programs. (Either way, accomplishing the FSF goal)
That's why Microsoft doesn't want all "pirates" to be caught and punished yet- as long as there is still an alternative development model out there, they don't want their customers to comprehend what it would mean to pay the price on every, single copy.
To most of these end-users, having or not having access to the source code is moot so why push something on them that they probably don't even want?
That is a fundamental, fundamental aspect of the GPL which RMS has expounded on many times.
In short, every end-user should have the power to select a programmer on the free market to improve the software he uses. If they don't want it, they simply aren't aware of the possibilities.
(The specific example of a military Tank Commander is one of the most extreme cases where the user wouldn't want to modify the code, because it'll be tied into painfully over-engineered software. Almost every other situation, including military users in R&D or training, will have greater utility from source code access. The X-Box in particular is someplace where users would love the source code.)
But the IT department has that right as agents for the corporation
According to the text of relevant laws and licenses, they have rights as individuals as well.
Corporations don't like that idea, though, and that is one reason why Bill Gates has been so successful in convincing them that the GPL is viral and evil. Gates often spreads misdirection (the word "viral" is inaccurate), but he can speak the truth too: when he tells CEOs that if their programmers modify GPLed code, the corporation might not be able to keep the changes from escaping to the public, he's right.
What he doesn't tell them, of course, is that public release will usually be non-harmful or even beneficial. (Letting that slip undermines his business model)
But that's just it, my employer does not "give" me the binaries.
They do "give" it to you (English definition), just like they "give" you many tools, although neither employee nor boss expects him to keep permanent possession- but that's not the point. Office furniture, computer hardware- they're physical goods. Whoever bought them can transfer them wherever he wants, but there's no law against copying them, since that's impossible.
When giving software to employees, however, the employer will have committed an act of copyright violation, unless he abides by the license. And if the license is the GPL, that means that anyone with access to the binaries can freely spread them far and wide.
Imagine if someone from the corporate IT department is called to the witness stand:
"Fred, did you on the date of March 1st, distribute or cause to be distributed to your co-employees Carol, Betty, and Alice, the software labelled exhibit A, which you have already admitted contains substantial portions of the plaintiff's source code?"
Fred cannot honestly deny he "distributed" the code. Then the only question is, was he violating copyright when he did so? Or, in doing so, had he consented to the GPL? In the first case, "Go directly to punitive damages". In the second case, Carol, Betty, and Alice are free to pass out further GPLed copies to the 4 corners of the earth.
The software installed on the computer remains the property of my employer
It was never the property of the employer. It belongs to the author and copyright holder the entire time.
The employer had permission to use it under certain conditions. If the software was GPL, those conditions included never restricting anyone from sharing copies of the software.
And I'm envisioning a kiosk that would be set up in a public place to say give directions to anyone who wishes to use it. By the definition of user you are attempting to assert, I would.
No, because the kiosk user doesn't have access to the filesystem. An employee of a corporation does. Maybe not the end-user who runs the program each day, but the IT dept who installed it certainly does.
And anyway, the FSF is considering extending the GPL to ensure that the user is entitled to the source code even if he's using a program running on someone else's computer. That won't matter for a while yet, though. (An example of a license that moves towards this goal is here)
but I think the same case law that allows a software publisher to go after a business that has illegal copies of software installed on the businesses systems governs here.
That is a more general thing, and relates to the definition of incorporation. If a crime is committed by a person acting on behalf of a corporation, liablity will be assessed against the corp. as a whole, rather than against any person within it.
But, laws like that only shift around who gets punished for any particular offense. It doesn't change what things are illegal for a corporation to do, which includes everything a private citizen can't do, and more.
So in my example of Fred above, if he turns out to be guilty, the corporation will pay the damages, if he was acting under their orders.
you are not a lawyer
Oh, how can you tell? Some lawyers work for my corporation, although I don't like to duplicate their speaking style exactly.
As contracted employees, they are not independent entities.
Yes they are. The idea that members of a corporation somehow become one big person is a legal fiction. Certain laws treat a corporation as a single person. Copyright law does not.
A license between an author and a customer can treat corporations specially. The GPL does not.
I think the Adobe comparison is bogus; with the GPL you can make multiple copies for your own use; with a commercial licence typically you can't.
In both cases, the only thing stopping you from making copies is copyright law. If the author permits you to make copies, you can. Adobe will give you this permission if you agree to give them some amount of dollars. GPL authors give you this permission if you agree to allow the recipient to make unlimited further copies.
Copyright law is the threat with which authors force you to agree to a license before duplicating their work. Some authors ask for money, others ask that you permit your changes to be shared freely. Whatever the price is, if you don't agree, you can't copy the software.
If a corporation isn't willing to allow it's employees to pass out GPLed software they've modified, then it must tell them not to modify GPL software.
Only the governement or the store that owns the cash register has legal standing to request the source code.
And any person they've given the binaries to. Which in the case of any software that the government uses heavily, is a large number of people.
The GPL may not recognize entities like businesses and the governement but the law does.
Copyright law does not recognize businesses, government agencies, or any other aggregation of people.
Most of the stuff I worked on had little or no demand outside of the governement/DoD world.
That's still true today, I'm sure that a major reason that more DoD code doesn't get leaked out to the public is that it's unfriendly, difficult to use (especially on consumer hardware), and plain-old boring.
from a licensing perspective, the end user is the acquiring agency;
That is how many software licenses are written, but it's not how the GPL is written. The GPL doesn't make any specific mention of corporations having special status. Nor does US copyright law give corporations (or agencies, or other kinds of organizations) special rights as a user of copyright (they have a few differences as a holder of copyright, though).
When the government wants software from Microsoft, they can negotiate a Volume License. There is no equivalent to a "Volume GPL", though. The GPL makes no mention of "groups", "companies", or "sites", so each individual person is the same as any other.
So, assuming a government agency recieved a modified GPL program as a deliverable. It'll have the GPL still attached, and each time they distribute it (to one of their military end-users), they'll have to abide by the GPL, or be in violation of copyright law.
That is why, I believe, contractors so far do not use GPLed code as the basis for deliverables- the government wouldn't like abiding by that license once they'd recieved the end product.
That would be like every cash register in a Linux based point of sale system coming with its own set of source CDs.
That's an interesting question, and one I don't see as 100% resolved yet. It comes down to the meaning of "give the binaries to" (because anyone with the binaries can demand the source). Does a person "have" the binaries, if they're embedded in a device whose filesystem he can't access?
Then, what about the related case of a consumer-product (like an MP3 player) having a GPL program embedded in the firmware? Is that end-user entitled to the source code? From watching RMS, it seems he wants the answer to be yes. (Of course, this is a little different from a cash register or battle tank, as the user owns that hardware)
This answer which you cited last week seems pretty clear to me.
Unfortunately, that answer solves nothing. It is the contents of a FAQ on GNU's webpage, and it doesn't matter, legally. The only things which really matter are the text of the GPL, and the laws of the jurisdiction.
"Distribution" has a meaning in the English language. "To pass out". Giving the software to other members of a company means "passing it out". And before it can be passed out, someone's got to make some copies. "Reproduce it".
"Reproducing" or "distributing" a copyrighted work (which all GPLed programs are), is illegal. Unless the copyright holder has given you permission. It is one of the actions that are "prohibited by law if you do not accept this License." (GPLv2 section 5)
So, if one employee gives the program to another, he must do so under the GPL. Which means granting the reciever full permission to give further copies to whoever he wants.
A comparison:
Can an employee of a company make a copy of Adobe Photoshop for another member of the company? Then why could he make a copy of Gimp, without special permission from the author.
Now, if the FSF had intended to permit corporations to alter GPL code, and give it to all their employees, but under the threat of termination and lawsuit if they passed it out to anyone else, then they should've stated this in the license.
The vague FAQ entry might amount to a promise from the FSF that they won't procescute a corporation for doing that, but other users of the GPL (like Linus Torvalds) will not necessarily agree with that interpretation.
(Note that the following discussion assumes the military contractor will be modifying GPLed source code. The fact that they will be "writing software on Linux" doesn't necessarily mean they'll do that)
I'm not sure where one draws the line between the rights of the programmer and the rights of the corporation (s)he works for.
Well, an employee should feel free to do anything his employer gave him permission to do.
When the corporation (or another employee, acting on the corporation's behalf) gives him a copy of a modified, GPL program, it gives him permission to redistribute it to whoever he wants.
reasonable grounds for dismissal
If that ever happened, it would be proof that the corporation is violating the GPL. It would violate this line of section 6:
"You may not impose any further restrictions on the recipients' exercise of the rights granted herein."
Once they impose a restriction on redistribution to any user of the program, they no longer have rights to modify or distribute that program.
No, they don't sit well. It brings to mind a giant GPL loophole:
Why can't this happen? Because the assumption that "one corporation is a single entity for purposes of the GPL" is incorrect. A company cannot modify a GPL program for "internal use only", unless they're willing to take the risk that a random employee will post it to USENET's comp.source, and they'll have no legal recourse.
I discussed this in some detail several times, like just last week (when, for a strange reason, I forgot to log in).
The short version is "Microsoft doesn't consider a corporation to be one person for software licensing purposes, so why should the GPL?"
When you develop software under governement contract, the governement owns the code.
It all depends on the contract, of course. But today, the default seems to be that both the customer and the contractor own the code. Either can use it.
I've seen this in a few big Pentagon software projects- the contract is fed out to multiple development companies. Each contributes only a part of the whole. The government owns the result, and can do whatever it likes with it- but, each contractor can keep the sections they individually wrote, and re-sell them if possible.
(Supposedly, it's cheaper to hire the developers if they think there's a chance they can find another customer later. Of course, the nature of the product, and how dangerous it is to share, should dictate the government's choice.)
Funny, I was speaking to a military Chemical/Biological weapons specialist back in 2001, and he claimed that never in history had any military knowingly released CB weapons on it's own troops.
(Of course, maybe there's a trick of words in here. Like, "When they started to lose, they were no longer my troops")
In terms of the military, you can break software into 2 categories for purposes of your concern:
Even though the Pentagon hasn't so far made any major software Open Source, they have been acting as if large numbers of un-cleared eyeballs would see their source code. (Programmers are expensive enough, without putting them through FBI background checks. Why, how do you expect a Pakistani H1B to pass those kids of loyalty tests?)
Also, I don't think they would be making major modifications to the Kernel.
So far, the DoD guys working on FCS (which is still only a raw, raw prototype) use unmodified kernel source code, with no new modules added. They do however recompile their kernels with a specific config file, to improve some kind of networking support.
The distributions they use are Red Hat or Familiar, depending on the hardware (although, so far PDAs only look like a gimmick). For a while they used Progeny (debian) linux too.
The DOD fork of Linux won't ever make it back onto the main 'branch.'
That will be up to Linus, because one of the software engineers working for Boeing will offer to send him the patches.
Why? Because they'll have a big team of programmers working on Linux, at least a few of them will feel favorable towards the public effort and want to send their changes out.
And since those programmers are sharing code amoung themselves, all of them have it under the GPL- so their managers have no legal means to prevent them from redistributing it.
On the other hand, it is unlikely that the DoD will truely fund Boeing or someone to fork linux. They'll pick RTLinux's or some other company's fork to use, possibly giving them a few additional requirements. That company will continue to release it's code the way it always has.
1,1000 attendees? That's nothing compared to 13,500 in the USA last year. And it lasted 3 weeks!
This is very difficult for most users- as most people are using one of the "pre-configured" X11 desktops like KDE/Gnome, rather than "rolling their own". And the behavior of "shifting a running window from one terminal to another" is so rare, that none of the major desktops have felt a need to support it in their GUIs.
To accomplish this task, you can use either X11 or VNC. Using VNC, it works the same as with any other VNC server (on Windows, Mac, Linux, whatever)- a VNC client can connect and reconnect multiple times, with the desktop programs still running.
To do it in X11... well, I've never done this. But I can give you a clue: xmove will make an intermediary Xserver between the application and the "real" Xserver, allowing the application's windows to be sent to other terminals while it's still running.
(Debian people can apt-get xmove)
You can't grab the video memory if the OS won't let you.
If major OSes start to ban printing, screencap, and clipboarding, and they make it impossible for the OS to be hacked into, then serious information-workers will demand a workaround.
Fast, high-quality Firewire cameras are common and inexpensive. They are capable of capturing text from a CRT in quite a legible form (especially if larger than 16 point). An OCR on another computer (or a different process on the same one) can change that image into editable text in real time.
Of course, it will be many years before there's any need for such a thing. By that time, everyone will have small PDAs with cameras and fast processors. The user just installs an OCR program, then waves the PDA over a desk or monitor and suddenly has PDFs of everything it saw.
Of course, then there will be a push to make this software illegal, which is in the end all you can do.
(A connection to the VGA cable would be much more reliable in the near future, although eventually display hardware may use an encrypted signal. But the emissions from the monitor will always be there)
Theoretically, a person could buy Microsoft Windows XP in a store, and bring that cdrom with him as he goes to examine MS's code. After getting the source compiled, he can start comparing the binaries with those that are publically distributed.
(Of course, there are so many service packs and things that even if the comparison shows a mismatch, MS can claim it doesn't indicate anything worrisome)
About $25 million.
So, only a few minutes of income for Microsoft.
He must be an optimist.
That is very, very intentional. It's more important when it happens with file extensions than actual programs, though.
Today, it is no longer possible to tell someone "Send me a bitmap file from the document", without him assuming you mean the output of MS Paint and MS Word, respectively.
through diligent study you may be able to determine whether or not you are in violation of some particular software patent
Moreover, if you do attempt to keep aware of software patents, then in the event you are found guilty of violating one, the punishment will be worse, as it was willful, knowing infringement.
how to 'map' cyberspace. In the paper we suggested that it should be mapped like a giant Venn diagram
That's an obvious idea, people have been envisioning such things since back with Vannevar Bush. As you discovered, there are many more important factors than simply having a good idea.
99% perspiration indeed.
You don't always need a working implementation for prior art (especially in the field of software, where the distinction between the design and the actual product is vague). The number of joke patents issued for science-fiction ideas attests to this.
But you do need to have released the description to the public.
Submitting a paper to a teacher for grading is specifically not adequate. Court cases on that exact basis have already been lost.
About the only saving grace in the patent system is the fact that they still expire in less than a human lifetime. Funny that Congress has seen fit to quintuple the length of copyrights, while leaving patents unextended. (Hope I don't give anyone ideas)
behind in aviation development at the start of world war I (as documented by none other than the US federal government itself)
That's one of my favorite funny stories. The irony of the heroic status given to Orville and Wilbur, in comparison to the money they were denied, calls the "capitalist" nature of the US into question. (US Aviation development wasn't degraded as much as it might've been, because the government took away the Wrights' patent while it still had more than 5 years left.)
Just try to imagine the "fair market value" for a concept like "the airplane". Tack on the price of victory in WW1 and WW2, and of beating the Soviets to the moon. Then scale all that up from 1903 to 2003 currency. We're looking at one trillion dollars.
(Of course, back then situations were different. Patents didn't apply internationally, so there were a different set of problems. An inventor in Britain would create something, and then US companies would market it, while other Britons are barred)