Of course, the firmware should actually be in a different file that's not even part of the module, with as restrictive a license as you want, but whatever.
Thus also makes easier such things as upgrading the firmware or allowing the save driver to operate with different hardware (which requires different firmware.)
That being said, I can see why companies would not want to release driver source code. I once worked on a project at another company that had a Xilinx chip onboard. On load, the driver would program the chip through it's jtag port. Dunno if I would want people poking around in that. Yeah, it could be reverse engineered...but you'd have to know what you were looking for. With the source it would be a no-brainer.
This might make some sense with custom hardware, if you are using off the shelf chips there are typically data sheets/books available.
Betcha NVidia and their ilk have firmware in their drivers. That's probably why they have a bug in their shorts about releasing code.
Only if they are using off the shelf hardware. Otherwise all you can see is that you have a string of bytes uploaded at the start and after it is uploaded the interface behaves in such and such a way.
The other protection offered by binary only is legal protection. If release the code into the public domain you will have little recourse when a competitor uses it to reverse engineer your hardware.
The GPL is not public domain
If you release a binary and your competitor disassembles it you would have a greater chance at legal recourse.
How are you going to prove to a court that someone took your binary, disassembled it and used it as an example to write their own driver? Even if they were to simply decompile it then recompile it, after changing any obvious like a copyright statement, it may well well wind up quite different.
The issue is, obfuscation DOES contributes to protection - at least in this case - theoretically you can always dissect the interface, hiding it permanently is NOT the goal - You only need to delay your competitors doing so until you roll out your next generation product.
You also delay yourself, since you need to write and debug obfuscated code...
For that scheme, companies generally put serial numbers in each hardware unit sold. When the user pays for extra users etc. they get a code that is specific to that serial number.
There is a whole community of people who hack programs to remove all sorts to protection, be it serial numbers, dongles, what have you.
Open source might mean that variable pricing (i.e. crippled versions of the software for cheap customers) would be easier to break.
If you want to provide a crippled version of the program then the obvious solution is to provide one which only contains the code you want them to have in.
many Taiwanese and Chinese manufacturers would be happy to adapt somebody other's drivers, since copying the hardware is the easier part for them.
An interesting claim, considering you need access to quite a bit of expensive hardware to copy hardware. If you copy the hardware then a binary copy of the driver will be fine. Anyway you don't need the source to be able to alter or reverse engineer a driver.
well, say you have two products. one product costs $500 and is the budget version and another product is $1000 for the "high-end" version. Now both of these products are actually the same, the only thing that separates them is the driver for the device.Sounds like the manufacturer is enguaging in some shady practice. Like having the same hardware with different firmware at different prices, trying to hide how much a piece of hardware uses the CPU, optimising for benchmarks, etc. If someone works out that they can use the high end driver with the cheaper product, then tough luck.
The goal for the media companies is to keep us from having "perfect digital copies" What they fail to realize is 90+% of people don't care about perfect. If they did they wouldn't be trading 128 or 192 mp3's. The loss from a good analog audio cable is much less then the loss from a 128k mp3.
Unless someone is prepared to spend lots of money on amp and speakers they might well not notice much difference anyway.
Working with the same OS, and having a clone of NVidia's hardware, a competitor would still be seriously lacking in performance, conformance and stability if they didn't have Nvidia's driver. That driver has a lot of voodoo in it for efficiently doing 3d graphics that's not entirely directly related to accessing hardware I/O on the card.
Having a binary only driver strops it being cloned exactly how? You don't think it might be possible to work out this "voodoo" by looking an the binary.
In situation's like Nvidia's, where if ATI saw their driver source code they could readily apply lessons learned to their own driver and reap benefits, it makes sense for nvidia to keep it closed.
What stops ATI buying a Nvidia card retail and examining exactly what the driver does?
if you're a hardware company, then until everyday people copy your hardware, there is no need to place restrictions on device drivers. even opening up the source in most cases will not hurt your company, unless you're trying to protect how the device communicates.
You may ever sell more hardware, since even people with systems you havn't provided your own drivers for will still be able to consider using your hardware. If anyone really wants to find out how your hardware communicates with a bus they can do so.
Of course that's true, but you ignore the whole reason why a company might not want to release their code under the GPL, which is that they don't want to give other people rights to their code.
The situation would only happen if they wanted to incorporate GPL code, most likely to cut their own development times and costs, in the first place. (N.B. they can freely use any GPL licenced software which functions as a "tool" without any licencing issues.) If they don't like the "price" of using GPL code in their software product, then they can alternativly write the whole thing from scratch, licence some proprietary code or get in contact with the author(s) of the GPL code they want to use. Though to be honest considering driver code to be "valuable" looks more like some dogma about proprietary software being universally good than anything else.
authors of works released under the GPL grant it on condition that you do the same for your work. You might find the price too high,
Against this would be what is the likely value of the code you have produced, possibly not that much on its own (especially in the case of a driver) and what it would cost you to use (including writing from scratch) non GPL code for the rest of your program.
The point is: without a license of some sort, you have NO permission to use the existing work in any way, not to copy, not to quote, not to use to derive new work from.
The only way would be to claim "fair use", but if you were using the copyright material directly in a commercial product that probably wouldn't hold up.
You're leaving out a very important point: The author is the one who decides to use GPL, and so it is the author who 'takes' these rights from himself.
The author had a free choice about using GPL code or not.
As for incorperating GPL code into your proprietary software, that would be theft on your part unless you either abide by the GPL, or negotiate an alternate license with the author.
Actually "copyright infringement" which isn't the same thing, however much the large media companies might like it to be.
Not true. It takes away the right of the author: - to keep his source code secret.
The GPL only does this if the software author creates a piece of software derived from GPL code. If they want to keep their source secret then they can either write the thing from scratch or licence code which allows them to keep the source secret. Both of which could easily cost them considerably more than using code from the GPL library.
- to limit the distribution of his work.
The GPL says nothing to either encourage or discourage distribution of software. Nor does copyright law for that matter, it simply prevents you copying works, once you have sold a copy the doctorine of first sale applies.
- to pursue a traditional software business model.
The traditional software business model is you contract someone to build a system to do something which makes makes your business work better. Either according to your on spec or as worked out by a systems analyst. GPL software can benefit both customer and contractor here.
If you copy with permission other than the GPL, no problem. If you copy under fair use, also no problem. The GPL only applies if permission and fair use do not apply.
Without the GPL your use would be copyright infringement ("software piracy"). Copyright law says that in the absence of permission you have no more right to copy parts of an open source program than you have to copy bits out of a novel or magazine. What the GPL does is grant you conditional permission to distribute copyright software and works derived from that copyright software. The conditions are that if you distribute binaries then you must also make sources available and that any parts you add to create a derived work (regardless of if you or the author of the code you are modifying hold the copyright on the modifications) must be subject to a GPL compatable copyright licence. There is no "infection" having to use the GPL for your own code is the price you pay for having access to a huge library of GPL code. Most likely using the GPL code saves you the money and time you'd need to write an entire whatever from scratch. This saving can often be considerably more than you could have ever hoped to have made than by selling licences to a piece of proprietary software When it comes to drivers what are you actually selling? A piece of hardware or software to interface it? If you supplied interface specs and an open source driver (maybe not even an especially good one) then you'd have a bigger potential market for your hardware. Without the expense of having to write and update drivers for platforms you possibly don't even known exist.
What is MS comparing that to? Building everything from scratch with MS tools?
In some cases using Microsoft's development tools means that you can't use any copyright licence you like. If anything these Microsoft licences are "viral".
NONE of the kernel developers will be able (or even willing) to assist you, because you're dealing with closed source module that will taint the kernel. They cannot help you because they simply cannot see what your driver does, i.e. see the source code.
Considering that the point of a driver is to enable hardware to work, what exactly do you lose by letting people know how to interface with your hardware? Anyone who really wants to know can pull your driver apart with a disassembler anyway.
In NVidia's case, this have gone tolerably, although many developers aren't happy with their closed source release, but I guess there are some neat tricks they don't want to share with ATI and others.
If there are "neat tricks" in the software driver then ATI, etc can simply look at the NVidia drivers. If they copied them you'd just end up with two parties arguing over who broke who's EULA. Does it actually make sense to consider how to interface a certain piece of hardware (possibly including uploading firmware) proprietary information? Especially since the entities involved are selling then hardware.
Let's say the FBI is investigating a poisoning. One of their possible suspects is known to have visited a library a few days before the murder. An agent gets a warrant and asks the librarian for a copy of the suspect's borrowing records. The records indicate that the suspect borrowed a medical text about the dosages and effects of various poisons, including the same poison used in the murder.
Here you have a specific crime and specific suspect(s). Which is rather different from seeing who might have borrowed "bad books".
Law enforcement has to have some particular reason to suspect YOU specifically before it probes through generally accepted expectations of privacy. The depth of the intrusion is propotional to the persuasiveness of the evidence. BUT NO FISHING EXPEDITIONS.
Or at least this is the way it should work. Similarly if they want to specifically spy on you or search your private property then in theory they have to convince a judge that there is some evidence to begin with.
The Patriot Act relies on a hysterical and ill-defined notion of a future terrorist threat to provide justification. This has been characteristic of many "emergency measures" in many countries over the years -- you know, we have to shut down the presses because it might cause trouble, etc. Now, it's been fairly quiet for over a year in the States -- when do you think they'll dilute the Act?
Governments like to grab power a lot more than they like to give it up. Sometimes the only way to get this to happen is a revolution.
A recent example abroad -- the Russian gov't interfered with internet and print press in the wake of the theater hostage-taking crisis. Although antiterrorism was the justification, a good portion of this appears to have been to save face for the gov't. They politely call this censorship "media restrictions."
Now, are we aiming to be more like the Russians, or more like us?
Do you really think the US (, British, French, German or wherever) government is any less interested in "face saving" than the Russians. People in power like to be portrayed as somehow superhuman and never making mistakes.
Most other democratic countries form coalition governments, where the parties are forced to reach some common ground and consensus.
In most democratic countries there are more than two political parties. Also there are frequently sub-national political parties.Including those such as the SNP, who's platform is for Scotland to be the independent nation state it hasn't been for nearly 400 years. Nothing like this in the US, even in Hawaii.
Yet, all this hubbub about the "differences" between the Democrats and Republicans pulls the wool over our eyes. How were the parties different on NAFTA, on bombing Afghanistan, or passing the DMCA? (They weren't.)
The "War on (some) drugs" is also a common slashdot example.
If you think that there's more than minor degrees of separation between the Republicans and Democrats, you need to take a healthy step back, my friend.
Especially where you have so called "open primaries" where candidates for either party are picked by the voters for both.
Similarly, public libraries exist in part to support an educated citizenry. If citizens are afraid to check out "dangerous" books to educate themselves, we're stifling the democratic process which requires free access to information.
If you have an entrenched ruling class, including career politicans, then an educated citizenship can be rather frightening. They might do things like investigating things themselves, rather than simply taking what those in government say as being correct. It's likely to be a bit harder to witch hunt Communism or Islam in a population which knows all about the principles of Communism or has read the Koran. Similarly claiming Iraq is a huge enemy of the world dosn't really wash with anyone who knows the recent history of that country.
Slashdot Mirror
Of course, the firmware should actually be in a different file that's not even part of the module, with as restrictive a license as you want, but whatever.
Thus also makes easier such things as upgrading the firmware or allowing the save driver to operate with different hardware (which requires different firmware.)
That being said, I can see why companies would not want to release driver source code. I once worked on a project at another company that had a Xilinx chip onboard. On load, the driver would program the chip through it's jtag port. Dunno if I would want people poking around in that. Yeah, it could be reverse engineered...but you'd have to know what you were looking for. With the source it would be a no-brainer.
This might make some sense with custom hardware, if you are using off the shelf chips there are typically data sheets/books available.
Betcha NVidia and their ilk have firmware in their drivers. That's probably why they have a bug in their shorts about releasing code.
Only if they are using off the shelf hardware. Otherwise all you can see is that you have a string of bytes uploaded at the start and after it is uploaded the interface behaves in such and such a way.
The other protection offered by binary only is legal protection. If release the code into the public domain you will have little recourse when a competitor uses it to reverse engineer your hardware.
The GPL is not public domain
If you release a binary and your competitor disassembles it you would have a greater chance at legal recourse.
How are you going to prove to a court that someone took your binary, disassembled it and used it as an example to write their own driver? Even if they were to simply decompile it then recompile it, after changing any obvious like a copyright statement, it may well well wind up quite different.
The issue is, obfuscation DOES contributes to protection - at least in this case - theoretically you can always dissect the interface, hiding it permanently is NOT the goal - You only need to delay your competitors doing so until you roll out your next generation product.
You also delay yourself, since you need to write and debug obfuscated code...
For that scheme, companies generally put serial numbers in each hardware unit sold. When the user pays for extra users etc. they get a code that is specific to that serial number.
There is a whole community of people who hack programs to remove all sorts to protection, be it serial numbers, dongles, what have you.
Open source might mean that variable pricing (i.e. crippled versions of the software for cheap customers) would be easier to break.
If you want to provide a crippled version of the program then the obvious solution is to provide one which only contains the code you want them to have in.
many Taiwanese and Chinese manufacturers would be happy to adapt somebody other's drivers, since copying the hardware is the easier part for them.
An interesting claim, considering you need access to quite a bit of expensive hardware to copy hardware. If you copy the hardware then a binary copy of the driver will be fine. Anyway you don't need the source to be able to alter or reverse engineer a driver.
well, say you have two products. one product costs $500 and is the budget version and another product is $1000 for the "high-end" version. Now both of these products are actually the same, the only thing that separates them is the driver for the device.Sounds like the manufacturer is enguaging in some shady practice. Like having the same hardware with different firmware at different prices, trying to hide how much a piece of hardware uses the CPU, optimising for benchmarks, etc.
If someone works out that they can use the high end driver with the cheaper product, then tough luck.
The goal for the media companies is to keep us from having "perfect digital copies" What they fail to realize is 90+% of people don't care about perfect. If they did they wouldn't be trading 128 or 192 mp3's. The loss from a good analog audio cable is much less then the loss from a 128k mp3.
Unless someone is prepared to spend lots of money on amp and speakers they might well not notice much difference anyway.
You're not proving anything. The infection *is* the price.
:)
By this kind of reasoning banks and stock exchanges should be considered bio-warfare factories
Working with the same OS, and having a clone of NVidia's hardware, a competitor would still be seriously lacking in performance, conformance and stability if they didn't have Nvidia's driver. That driver has a lot of voodoo in it for efficiently doing 3d graphics that's not entirely directly related to accessing hardware I/O on the card.
Having a binary only driver strops it being cloned exactly how? You don't think it might be possible to work out this "voodoo" by looking an the binary.
In situation's like Nvidia's, where if ATI saw their driver source code they could readily apply lessons learned to their own driver and reap benefits, it makes sense for nvidia to keep it closed.
What stops ATI buying a Nvidia card retail and examining exactly what the driver does?
if you're a hardware company, then until everyday people copy your hardware, there is no need to place restrictions on device drivers. even opening up the source in most cases will not hurt your company, unless you're trying to protect how the device communicates.
You may ever sell more hardware, since even people with systems you havn't provided your own drivers for will still be able to consider using your hardware.
If anyone really wants to find out how your hardware communicates with a bus they can do so.
Aren't we talking about a driver here though? Few people would buy hardware if there wasn't software to make the device work.
Without the hardware the driver isn't of much practical use either.
Of course that's true, but you ignore the whole reason why a company might not want to release their code under the GPL, which is that they don't want to give other people rights to their code.
The situation would only happen if they wanted to incorporate GPL code, most likely to cut their own development times and costs, in the first place. (N.B. they can freely use any GPL licenced software which functions as a "tool" without any licencing issues.)
If they don't like the "price" of using GPL code in their software product, then they can alternativly write the whole thing from scratch, licence some proprietary code or get in contact with the author(s) of the GPL code they want to use.
Though to be honest considering driver code to be "valuable" looks more like some dogma about proprietary software being universally good than anything else.
authors of works released under the GPL grant it on condition that you do the same for your work. You might find the price too high,
Against this would be what is the likely value of the code you have produced, possibly not that much on its own (especially in the case of a driver) and what it would cost you to use (including writing from scratch) non GPL code for the rest of your program.
The point is: without a license of some sort, you have NO permission to use the existing work in any way, not to copy, not to quote, not to use to derive new work from.
The only way would be to claim "fair use", but if you were using the copyright material directly in a commercial product that probably wouldn't hold up.
You're leaving out a very important point: The author is the one who decides to use GPL, and so it is the author who 'takes' these rights from himself.
The author had a free choice about using GPL code or not.
As for incorperating GPL code into your proprietary software, that would be theft on your part unless you either abide by the GPL, or negotiate an alternate license with the author.
Actually "copyright infringement" which isn't the same thing, however much the large media companies might like it to be.
Not true. It takes away the right of the author:
- to keep his source code secret.
The GPL only does this if the software author creates a piece of software derived from GPL code. If they want to keep their source secret then they can either write the thing from scratch or licence code which allows them to keep the source secret. Both of which could easily cost them considerably more than using code from the GPL library.
- to limit the distribution of his work.
The GPL says nothing to either encourage or discourage distribution of software. Nor does copyright law for that matter, it simply prevents you copying works, once you have sold a copy the doctorine of first sale applies.
- to pursue a traditional software business model.
The traditional software business model is you contract someone to build a system to do something which makes makes your business work better. Either according to your on spec or as worked out by a systems analyst. GPL software can benefit both customer and contractor here.
If you copy with permission other than the GPL, no problem. If you copy under fair use, also no problem. The GPL only applies if permission and fair use do not apply.
Without the GPL your use would be copyright infringement ("software piracy"). Copyright law says that in the absence of permission you have no more right to copy parts of an open source program than you have to copy bits out of a novel or magazine.
What the GPL does is grant you conditional permission to distribute copyright software and works derived from that copyright software.
The conditions are that if you distribute binaries then you must also make sources available and that any parts you add to create a derived work (regardless of if you or the author of the code you are modifying hold the copyright on the modifications) must be subject to a GPL compatable copyright licence.
There is no "infection" having to use the GPL for your own code is the price you pay for having access to a huge library of GPL code. Most likely using the GPL code saves you the money and time you'd need to write an entire whatever from scratch. This saving can often be considerably more than you could have ever hoped to have made than by selling licences to a piece of proprietary software
When it comes to drivers what are you actually selling? A piece of hardware or software to interface it? If you supplied interface specs and an open source driver (maybe not even an especially good one) then you'd have a bigger potential market for your hardware. Without the expense of having to write and update drivers for platforms you possibly don't even known exist.
What is MS comparing that to? Building everything from scratch with MS tools?
In some cases using Microsoft's development tools means that you can't use any copyright licence you like. If anything these Microsoft licences are "viral".
NONE of the kernel developers will be able (or even willing) to assist you, because you're dealing with closed source module that will taint the kernel. They cannot help you because they simply cannot see what your driver does, i.e. see the source code.
Considering that the point of a driver is to enable hardware to work, what exactly do you lose by letting people know how to interface with your hardware? Anyone who really wants to know can pull your driver apart with a disassembler anyway.
In NVidia's case, this have gone tolerably, although many developers aren't happy with their closed source release, but I guess there are some neat tricks they don't want to share with ATI and others.
If there are "neat tricks" in the software driver then ATI, etc can simply look at the NVidia drivers. If they copied them you'd just end up with two parties arguing over who broke who's EULA.
Does it actually make sense to consider how to interface a certain piece of hardware (possibly including uploading firmware) proprietary information? Especially since the entities involved are selling then hardware.
Let's say the FBI is investigating a poisoning. One of their possible suspects is known to have visited a library a few days before the murder. An agent gets a warrant and asks the librarian for a copy of the suspect's borrowing records. The records indicate that the suspect borrowed a medical text about the dosages and effects of various poisons, including the same poison used in the murder.
Here you have a specific crime and specific suspect(s). Which is rather different from seeing who might have borrowed "bad books".
Yes, the library is a public place, but what I look at and what I check out is my private business,
Private between you and the library.
and unless I'm already under investigation, they have no right to this information.
Even if you are under investigation, someone should still have to make the case that your book borrowing is relevent.
Law enforcement has to have some particular reason to suspect YOU specifically before it probes through generally accepted expectations of privacy. The depth of the intrusion is propotional to the persuasiveness of the evidence. BUT NO FISHING EXPEDITIONS.
Or at least this is the way it should work. Similarly if they want to specifically spy on you or search your private property then in theory they have to convince a judge that there is some evidence to begin with.
The Patriot Act relies on a hysterical and ill-defined notion of a future terrorist threat to provide justification. This has been characteristic of many "emergency measures" in many countries over the years -- you know, we have to shut down the presses because it might cause trouble, etc. Now, it's been fairly quiet for over a year in the States -- when do you think they'll dilute the Act?
Governments like to grab power a lot more than they like to give it up. Sometimes the only way to get this to happen is a revolution.
A recent example abroad -- the Russian gov't interfered with internet and print press in the wake of the theater hostage-taking crisis. Although antiterrorism was the justification, a good portion of this appears to have been to save face for the gov't. They politely call this censorship "media restrictions."
Now, are we aiming to be more like the Russians, or more like us?
Do you really think the US (, British, French, German or wherever) government is any less interested in "face saving" than the Russians.
People in power like to be portrayed as somehow superhuman and never making mistakes.
Most other democratic countries form coalition governments, where the parties are forced to reach some common ground and consensus.
In most democratic countries there are more than two political parties. Also there are frequently sub-national political parties.Including those such as the SNP, who's platform is for Scotland to be the independent nation state it hasn't been for nearly 400 years. Nothing like this in the US, even in Hawaii.
Yet, all this hubbub about the "differences" between the Democrats and Republicans pulls the wool over our eyes. How were the parties different on NAFTA, on bombing Afghanistan, or passing the DMCA? (They weren't.)
The "War on (some) drugs" is also a common slashdot example.
If you think that there's more than minor degrees of separation between the Republicans and Democrats, you need to take a healthy step back, my friend.
Especially where you have so called "open primaries" where candidates for either party are picked by the voters for both.
Similarly, public libraries exist in part to support an educated citizenry. If citizens are afraid to check out "dangerous" books to educate themselves, we're stifling the democratic process which requires free access to information.
If you have an entrenched ruling class, including career politicans, then an educated citizenship can be rather frightening.
They might do things like investigating things themselves, rather than simply taking what those in government say as being correct.
It's likely to be a bit harder to witch hunt Communism or Islam in a population which knows all about the principles of Communism or has read the Koran. Similarly claiming Iraq is a huge enemy of the world dosn't really wash with anyone who knows the recent history of that country.