Shared media Ethernet (Thin ethernet over coax, or 10baseT with a hub instead of a switch) has a similar requirement for CSMA/CD, and with the accompanying reduction in usable bandwidth. 10 Mbps -> 3 Mbps is common.
At an average of 50-60 bytes apiece, that's a total of a whopping 47 kbps, or 0.0047% of capacity
The effective capacity consumed could be quite a bit higher than that due to CSMA/CD overhead and the like. If someone else is transmitting a station has to wait a random amount of time before transmitting, for example. That is a non trivial factor that can easily take a busy 10 Mbps network down to 3 Mbps of usable capacity, for example.
There is long standing precedent that anyone can listen to virtually anything broadcast over the radio, as long as he or she does not disclose private transmissions to third parties or use the contents for his own benefit.
Several years ago Congress enacted restrictions on the interception of private "oral communication", including prohibiting the manufacture of ham radio equipment that could tune into certain bands (such as the original analog cell phone spectrum).
With those caveats, anything that is broadcast that is "readily accessible to the general public" appears to be fair game.
18 USC 2511 (g): It shall not be unlawful under this chapter or chapter 121 of this title for any person (i) to intercept or access an electronic communication made through an electronic communication system that is configured so that such electronic communication is readily accessible to the general public;
Unless intercepting unencrypted wireless traffic on an open network is a violation of some other law, it appears the Electronic Communications Privacy Act doesn't restrict it at all. (I posted this on the linked blog as well).
In other words, if you want to have an expectation of privacy on a wireless network using a readily available technology, you better use encryption and exercise some modicum of discretion about who you share the keys with.
Strictly speaking, such filesystems use a 4K "block" size or a 4K "cluster" size by default, where a block or cluster is mapped to a group of typically smaller sectors. In most filesystems the block size is adjustable in powers of two, although it is rare for modern filesystem to support block sizes smaller than 1K or the underlying sector size, which ever is larger. And with 4K physical sector sizes about to become predominant, there won't be a lot of point in setting it smaller than that.
Of course, if you have a large number of really small files, you can always use a filesystem that does "tail packing", allowing it to store more than one (small) file per filesystem block.
This is all made somewhat more confusing by the fact that new 4K physical sector drives are going to keep pretending they have 512 byte logical sectors for several years for compatibility reasons. Then you have the filesystem block or cluster size that is logically independent of that.
As I understand it he was ousted primarily over his support for TARP
That is what the Tea Party folks claim, the real motives of the two thousand some odd delegates who voted in favor of other candidates are more complex. I suspect, for example, that sponsoring a federal healthcare reform bill with stronger individual mandates than those enacted with Obamacare was a bigger issue than TARP. The opposing candidates, at convention, hardly mentioned either. Bennett's voting record on a number of controversial points was extremely well publicized prior to the convention, if the results are any indication a substantial majority of delegates weren't impressed.
Worse, they didn't want to engage in rational discussion of the issues and were unwilling to consider compromise. They also seem to misunderstand what exactly one senator, or every Republican senator could do with respect to HCR. All you got was just a drumbeat about the "loving" the Constitution. This will get them nowhere in Congress.
That is a poor generalization based on a few anecdotes. There were ~3500 state delegates and the idiosyncrasies of a small number of them does not a proper analysis make. The thing about Senator Bennett is he doesn't actually seem to get it.
Conservative Republicans in the state haven't liked him very much for years. This year there were just a lot more conservative delegates - in large part due to horror at the prospect of trillion dollar deficits as far as the eye can see.
Perhaps Bennett thinks earmarks for such projects as the "bridge to nowhere" are fine. Perhaps he thinks that sponsoring health care reform bills that have stronger individual mandates (at the federal level) than the bill that actually passed is legitimate. Perhaps he thinks that granting $800 billion to the Treasury department to spend as they saw fit was a good idea. Perhaps he thought that opposing reasonable reforms of Fannie Mae back in 2003 was a good idea.
The merits of those positions notwithstanding, most of the delegates disagreed, and they chose two candidates (excellent candidates I might add) that more closely reflected their views. It has nothing to do with "anti incumbent" feeling. It has to do with the rise of movement conservatism, something that Senator Bennett doesn't seem to have much sympathy for. So the serious conservatives voted him out, and by a rather overwhelming margin at that. He couldn't command more than 27% of the vote.
You can when they're also pushing for an end to earmarks, reduced government spending, and a generalized "the government can't do anything right" attitude.
You know not of whom you speak. Senator Bennett recently lost out to two more conservative candidates for the opportunity to run in the Republican primary. Among other reasons, because he never saw an earmark he didn't like.
Wait, am I violating their copyright on a code they keep in violation of license they got that code under?
"Keeping" code is not a violation of a license like the GPL. Distributing a derived work without providing source code and similar license terms is. If you distribute that same derived work prior to the grant of such a license, now you are violating multiple copyrights - those of the original authors and those of the new ones. Both could sue you and win.
Which idea? What claim? All I said was that a collection of unmodified source files is not a derived work.
Code that requires GPLed libraries to compile are derived works
On the contrary, a derived work is created when such code is compiled and linked with code under another license. Prior to compilation and linking, not in the slightest. And compilation only matters if the header file contains substantial code in its own right (inline functions, exotic macros, C++ templates, etc).
The only way new source files would be a derived work would be if it was substantially similar to (i.e. "copied from") another work, and that similarity was not a matter of functional necessity.
The idea that compatibility creates derivation is one of the most irrational and legally unsupportable ideas ever conceived. If it were made the law of the land, the software industry (among others) would be dropped dead in its tracks.
in Canada, if you're hired as a contractor to produce a work, you retain the copyright on that work... However, I don't believe this is true in the US.
It is true in the US as well. Contractors automatically retain copyright unless otherwise provided for by contract. That doesn't mean the customer won't have an implied license, of course, just not the copyright.
you'll probably need to demonstrate that you weren't aware of that at the time
If it is in written, signed contract, it doesn't matter whether you were aware of it at the time, unless the terms agreed to are illegal, unconscionable, lacking consideration, signed under duress, explained fraudulently, etc.
Lacking a written agreement, the University may not have a leg to stand on as it wouldn't be a work-for-hire regardless
It depends on whether the developer was an employee or a contractor. If an employee, the University doesn't need any sort of agreement. Everything done by an employee in that capacity is automatically a work for hire unless otherwise provided for by contract.
If the developer was a contractor, the presumption is reversed. A contractor retains copyright to all of his work unless otherwise provided for by contract. Most contracts for hourly software development work specify intellectual property transfer of the rights with regard to that work. Although only rarely do they do so with regard to any pre-existing work, for obvious reasons.
You should read the summary. The question he is asking is whether the university owns the final combined version. Even if they have clear license to his original code, they do not own it unless there was a written transfer of copyright.
You have made some serious accusations in your comment, without any sort of evidence. In other contexts, that could get you sued for slander. Which promise or obligation is he breaking?
He wrote original software. Did they acquire the copyright from him? Do they have a clear license so that they do not need to? If anyone is violating the law here, it is employer by making the legally unjustifiable claim that they now own the derived work en toto without a written transfer of copyright from him.
Without such a transfer, he retains a copyright over his original contributions, they are his, he was never paid for them, and anything they do with that code is subject to some sort of license from him, implied or otherwise. If they want clear title, they should pay for it.
If he didn't provide them with some other license, didn't assign copyright, and if the code had GPL notices, etc., in the initial distribution, he may be able to argue that he gave it to them under the GPL
Unless there was a written transfer of copyright, he didn't "give" the original code to them at all. If there wasn't a GPL notice or verbal license to similar effect, at best they have some sort of implied license to his original code. If he quits, they may not be able to legally continue to prepare non-trivial derived works, even internal ones.
If anything it is best for his employer if there is a written software license to his original code, because otherwise they could be seriously limited in what they can do.
because legally you did nothing but work on an open source project
Sorry, No. Everything an employee does is a "work for hire" unless otherwise provided for by contract.
whose license FORCES you to release the code
Not true. You could sue them and win, and they still wouldn't have to license their additions under the GPL. They could just pay monetary damages and refrain from further outside distribution.
no written or verbal agreement was ever made to transfer copyright over to my employer
As I understand it, it is impossible to transfer a copyright with a verbal agreement:
"A transfer of copyright ownership, other than by operation of law, is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed or such owner's duly authorized agent." (17 USC 204)
But if even one copy ever "leaks", it's out in the wild and free for all.
Not true. There is no such thing as a viral license. The only remedy against violation of the GPL is to sue the violator. Then if they remedy the problem by licensing their additions under the GPL, you can use it. Until then, you are just violating their copyright as well as the copyright of the original contributors. Nothing is ever automatically placed under the GPL.
From my understanding of the GPL, you were fine creating software that relied on GPLed libraries and not GPLing it, provided you didn't release it. The place where it kicks in is when you distribute (say, to the university). At that point, you were obligated to put your code under the GPL.
Unless he distributed binaries to the university that were compiled and linked before being hired, or modified versions of the GPL libraries themselves, no such obligation applies. The GPL kicks in when you distribute a derived work. A collection of unmodified source files is no such thing.
On the contrary: The code he wrote, on his own, before getting hired is almost certainly his. The code he wrote while employed is his employers. The code third parties wrote is of course the property of neither. All else being equal.
It is worth noting that while he does not have the right to distribute the work he did for his employer, unless the original code he wrote was clearly licensed to his employer under the GPL or comparable license, his employer may not have the right to prepare derived works of anything that contains his original code either.
If not, then there's simply no record that he ever did any work before he went to work for them
Legally, he doesn't need a record of any kind. If he did any work before he was hired, and his employment contract did not specify a transfer of existing copyrights, he owns the copyright to all of his previous work.
Of course, records would help. Dated backups, documentation files, comments in source code, original copyright statements, emails, etc. He doesn't actually need to have distributed the software to anybody.
take away the regulators ability to maintain order and you have a recipe for disaster
No one has taken away the regulators ability to maintain order. The FCC decided several years ago they didn't want it. Now they want it back. No problem, all they have to do is apply the right law ("Title II" of the 1934 Communications Act as amended). No one is stopping them except themselves.
Boo hoo. The FCC exists to implement the laws that Congress passes, not execute its own whims. And if the laws that Congress passes makes it a burden to regulate common carriers, well then perhaps Congress should look into that rather than expecting the FCC to make a politically opportunistic end run around the law.
In this case Congress has already given the FCC extensive authority to regulate Internet access providers. The FCC simply has to use Title II to do it.
You would think that since the trouble started with the FCC, they could just change their minds and put things back the way they were
They have changed their minds, they have ample reason to do so, and that is what they are planning to do. They just have to be able to convince the courts that they are not acting in an arbitrary and capricious manner. I am frankly amazed that the courts deferred to their original (mis)classification in the first place, because it was obviously wrong by any reasonable standard.
Slashdot Mirror
Shared media Ethernet (Thin ethernet over coax, or 10baseT with a hub instead of a switch) has a similar requirement for CSMA/CD, and with the accompanying reduction in usable bandwidth. 10 Mbps -> 3 Mbps is common.
At an average of 50-60 bytes apiece, that's a total of a whopping 47 kbps, or 0.0047% of capacity
The effective capacity consumed could be quite a bit higher than that due to CSMA/CD overhead and the like. If someone else is transmitting a station has to wait a random amount of time before transmitting, for example. That is a non trivial factor that can easily take a busy 10 Mbps network down to 3 Mbps of usable capacity, for example.
There is long standing precedent that anyone can listen to virtually anything broadcast over the radio, as long as he or she does not disclose private transmissions to third parties or use the contents for his own benefit.
Several years ago Congress enacted restrictions on the interception of private "oral communication", including prohibiting the manufacture of ham radio equipment that could tune into certain bands (such as the original analog cell phone spectrum).
With those caveats, anything that is broadcast that is "readily accessible to the general public" appears to be fair game.
18 USC 2511 (g): It shall not be unlawful under this chapter or chapter 121 of this title for any person (i) to intercept or access an electronic communication made through an electronic communication system that is configured so that such electronic communication is readily accessible to the general public;
Unless intercepting unencrypted wireless traffic on an open network is a violation of some other law, it appears the Electronic Communications Privacy Act doesn't restrict it at all. (I posted this on the linked blog as well).
In other words, if you want to have an expectation of privacy on a wireless network using a readily available technology, you better use encryption and exercise some modicum of discretion about who you share the keys with.
most file systems already use a 4k sector.
Strictly speaking, such filesystems use a 4K "block" size or a 4K "cluster" size by default, where a block or cluster is mapped to a group of typically smaller sectors. In most filesystems the block size is adjustable in powers of two, although it is rare for modern filesystem to support block sizes smaller than 1K or the underlying sector size, which ever is larger. And with 4K physical sector sizes about to become predominant, there won't be a lot of point in setting it smaller than that.
Of course, if you have a large number of really small files, you can always use a filesystem that does "tail packing", allowing it to store more than one (small) file per filesystem block.
This is all made somewhat more confusing by the fact that new 4K physical sector drives are going to keep pretending they have 512 byte logical sectors for several years for compatibility reasons. Then you have the filesystem block or cluster size that is logically independent of that.
As I understand it he was ousted primarily over his support for TARP
That is what the Tea Party folks claim, the real motives of the two thousand some odd delegates who voted in favor of other candidates are more complex. I suspect, for example, that sponsoring a federal healthcare reform bill with stronger individual mandates than those enacted with Obamacare was a bigger issue than TARP. The opposing candidates, at convention, hardly mentioned either. Bennett's voting record on a number of controversial points was extremely well publicized prior to the convention, if the results are any indication a substantial majority of delegates weren't impressed.
Worse, they didn't want to engage in rational discussion of the issues and were unwilling to consider compromise. They also seem to misunderstand what exactly one senator, or every Republican senator could do with respect to HCR. All you got was just a drumbeat about the "loving" the Constitution. This will get them nowhere in Congress.
That is a poor generalization based on a few anecdotes. There were ~3500 state delegates and the idiosyncrasies of a small number of them does not a proper analysis make. The thing about Senator Bennett is he doesn't actually seem to get it.
Conservative Republicans in the state haven't liked him very much for years. This year there were just a lot more conservative delegates - in large part due to horror at the prospect of trillion dollar deficits as far as the eye can see.
Perhaps Bennett thinks earmarks for such projects as the "bridge to nowhere" are fine. Perhaps he thinks that sponsoring health care reform bills that have stronger individual mandates (at the federal level) than the bill that actually passed is legitimate. Perhaps he thinks that granting $800 billion to the Treasury department to spend as they saw fit was a good idea. Perhaps he thought that opposing reasonable reforms of Fannie Mae back in 2003 was a good idea.
The merits of those positions notwithstanding, most of the delegates disagreed, and they chose two candidates (excellent candidates I might add) that more closely reflected their views. It has nothing to do with "anti incumbent" feeling. It has to do with the rise of movement conservatism, something that Senator Bennett doesn't seem to have much sympathy for. So the serious conservatives voted him out, and by a rather overwhelming margin at that. He couldn't command more than 27% of the vote.
You can when they're also pushing for an end to earmarks, reduced government spending, and a generalized "the government can't do anything right" attitude.
You know not of whom you speak. Senator Bennett recently lost out to two more conservative candidates for the opportunity to run in the Republican primary. Among other reasons, because he never saw an earmark he didn't like.
Wait, am I violating their copyright on a code they keep in violation of license they got that code under?
"Keeping" code is not a violation of a license like the GPL. Distributing a derived work without providing source code and similar license terms is. If you distribute that same derived work prior to the grant of such a license, now you are violating multiple copyrights - those of the original authors and those of the new ones. Both could sue you and win.
Which idea? What claim? All I said was that a collection of unmodified source files is not a derived work.
Code that requires GPLed libraries to compile are derived works
On the contrary, a derived work is created when such code is compiled and linked with code under another license. Prior to compilation and linking, not in the slightest. And compilation only matters if the header file contains substantial code in its own right (inline functions, exotic macros, C++ templates, etc).
The only way new source files would be a derived work would be if it was substantially similar to (i.e. "copied from") another work, and that similarity was not a matter of functional necessity.
The idea that compatibility creates derivation is one of the most irrational and legally unsupportable ideas ever conceived. If it were made the law of the land, the software industry (among others) would be dropped dead in its tracks.
in Canada, if you're hired as a contractor to produce a work, you retain the copyright on that work... However, I don't believe this is true in the US.
It is true in the US as well. Contractors automatically retain copyright unless otherwise provided for by contract. That doesn't mean the customer won't have an implied license, of course, just not the copyright.
you'll probably need to demonstrate that you weren't aware of that at the time
If it is in written, signed contract, it doesn't matter whether you were aware of it at the time, unless the terms agreed to are illegal, unconscionable, lacking consideration, signed under duress, explained fraudulently, etc.
Lacking a written agreement, the University may not have a leg to stand on as it wouldn't be a work-for-hire regardless
It depends on whether the developer was an employee or a contractor. If an employee, the University doesn't need any sort of agreement. Everything done by an employee in that capacity is automatically a work for hire unless otherwise provided for by contract.
If the developer was a contractor, the presumption is reversed. A contractor retains copyright to all of his work unless otherwise provided for by contract. Most contracts for hourly software development work specify intellectual property transfer of the rights with regard to that work. Although only rarely do they do so with regard to any pre-existing work, for obvious reasons.
You should read the summary. The question he is asking is whether the university owns the final combined version. Even if they have clear license to his original code, they do not own it unless there was a written transfer of copyright.
You have made some serious accusations in your comment, without any sort of evidence. In other contexts, that could get you sued for slander. Which promise or obligation is he breaking?
He wrote original software. Did they acquire the copyright from him? Do they have a clear license so that they do not need to? If anyone is violating the law here, it is employer by making the legally unjustifiable claim that they now own the derived work en toto without a written transfer of copyright from him.
Without such a transfer, he retains a copyright over his original contributions, they are his, he was never paid for them, and anything they do with that code is subject to some sort of license from him, implied or otherwise. If they want clear title, they should pay for it.
If he didn't provide them with some other license, didn't assign copyright, and if the code had GPL notices, etc., in the initial distribution, he may be able to argue that he gave it to them under the GPL
Unless there was a written transfer of copyright, he didn't "give" the original code to them at all. If there wasn't a GPL notice or verbal license to similar effect, at best they have some sort of implied license to his original code. If he quits, they may not be able to legally continue to prepare non-trivial derived works, even internal ones.
If anything it is best for his employer if there is a written software license to his original code, because otherwise they could be seriously limited in what they can do.
because legally you did nothing but work on an open source project
Sorry, No. Everything an employee does is a "work for hire" unless otherwise provided for by contract.
whose license FORCES you to release the code
Not true. You could sue them and win, and they still wouldn't have to license their additions under the GPL. They could just pay monetary damages and refrain from further outside distribution.
no written or verbal agreement was ever made to transfer copyright over to my employer
As I understand it, it is impossible to transfer a copyright with a verbal agreement:
"A transfer of copyright ownership, other than by operation of law, is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed or such owner's duly authorized agent." (17 USC 204)
But if even one copy ever "leaks", it's out in the wild and free for all.
Not true. There is no such thing as a viral license. The only remedy against violation of the GPL is to sue the violator. Then if they remedy the problem by licensing their additions under the GPL, you can use it. Until then, you are just violating their copyright as well as the copyright of the original contributors. Nothing is ever automatically placed under the GPL.
From my understanding of the GPL, you were fine creating software that relied on GPLed libraries and not GPLing it, provided you didn't release it. The place where it kicks in is when you distribute (say, to the university). At that point, you were obligated to put your code under the GPL.
Unless he distributed binaries to the university that were compiled and linked before being hired, or modified versions of the GPL libraries themselves, no such obligation applies. The GPL kicks in when you distribute a derived work. A collection of unmodified source files is no such thing.
On the contrary: The code he wrote, on his own, before getting hired is almost certainly his. The code he wrote while employed is his employers. The code third parties wrote is of course the property of neither. All else being equal.
It is worth noting that while he does not have the right to distribute the work he did for his employer, unless the original code he wrote was clearly licensed to his employer under the GPL or comparable license, his employer may not have the right to prepare derived works of anything that contains his original code either.
If not, then there's simply no record that he ever did any work before he went to work for them
Legally, he doesn't need a record of any kind. If he did any work before he was hired, and his employment contract did not specify a transfer of existing copyrights, he owns the copyright to all of his previous work.
Of course, records would help. Dated backups, documentation files, comments in source code, original copyright statements, emails, etc. He doesn't actually need to have distributed the software to anybody.
take away the regulators ability to maintain order and you have a recipe for disaster
No one has taken away the regulators ability to maintain order. The FCC decided several years ago they didn't want it. Now they want it back. No problem, all they have to do is apply the right law ("Title II" of the 1934 Communications Act as amended). No one is stopping them except themselves.
they do not WANT to classify ISPs using Title II
Boo hoo. The FCC exists to implement the laws that Congress passes, not execute its own whims. And if the laws that Congress passes makes it a burden to regulate common carriers, well then perhaps Congress should look into that rather than expecting the FCC to make a politically opportunistic end run around the law.
In this case Congress has already given the FCC extensive authority to regulate Internet access providers. The FCC simply has to use Title II to do it.
You would think that since the trouble started with the FCC, they could just change their minds and put things back the way they were
They have changed their minds, they have ample reason to do so, and that is what they are planning to do. They just have to be able to convince the courts that they are not acting in an arbitrary and capricious manner. I am frankly amazed that the courts deferred to their original (mis)classification in the first place, because it was obviously wrong by any reasonable standard.