Sure, and don't blame Windows for it's crashes either then. Those are all caused by bad drivers as well.
We are talking about VMWare, which should have been able to run without additional drivers at all. Part of the reason it doesn't is probably the IA32 design, that simply makes virtualization more difficult than it ought to be. However if you throw such stuff into your kernel, you simply cannot expect it to be as stable as it used to be.
They want to pay to have no bugs.
In that case the first thing to do is to buy well documented hardware that actually works like the documentation says. When that has done it wouldn't be expensive to get bugs fixed as they are found. (Everybody ought to know, that a bug can never be fixed before it is found). With Windows OTOH you cannot pay somebody to fix the bugs as they are found.
Most crashes are either something wonky with the hardware or bad drivers or bad programming on the software manufacturer (not the OS, the app).
The last statement simply doesn't make sense. If an application program running without administrator priveleges can crash the OS, it is by definition a bug in the OS (or hardware).
But has it ever been taken to court? (Not that I would want to be the first one.)
If you have a license and want to use it on a different computer, you could just do it. It is not like Microsoft is ever going to find out about that. I wonder under which circumstances this would be taken to court.
Microsoft is going to be very, very interested in auditing these
guys.
Sure, but I certainly hope Microsoft doesn't get everything they want. I could
name a lot of stuff I would be interested in, that I'm probably never going to
get. What rights do Microsoft have to perform such an audit?
Probably the VMware
drivers.
So don't blame Linux for that.
Gentoo crashed on me. ATA raid drivers, the open source ones. The closed source ones work flawlessly.
Interesting case. There are multiple possible explanations. Could be that the drivers where not just tested well enough, doesn't sound like a piece of hardware everybody have. So having one you should at least provide a decent bug report, otherwise don't expect it to be fixed. It could of course also be the case, that no usable documentation is available for the device. That could explain why the closed source ones worked, the authors simply had some knowledge about the device that wasn't documented, and was required for stable operation.
SMP + SCSI is a no no under Linux in some kernels.
We have a dual Xeon (which means four logical CPUs in total) NFS server using SCSI, and it works just fine. High load has not yet been a problem, the load is rarely above 0.09.
I can't recall the last time my xp blue screened except for runnin a non certified video driver.
If that is true, it is really not much different from Linux. When running RH7.3 I had frequent crashes caused by the video driver, luckily it was solved when upgrading to RH9. In my experience most Linux crashes are caused by bad hardware and/or drivers. With Linux there is no such thing as a certified driver, but in reallity it probably compares to closed source drivers which will taint the kernel. So if we translate your statement into Linux terms, we will get a statement the I guess multiple Linux users could say: I can't recall the last time my Linux paniced except for runing a closed source video driver. In my particular case I was lucky enough, that the driver was open source, and a minor change eliminated the panics. But many video chips don't have open source drivers, and no specs are available for one to be developed. So if you want a fair comparision between the stability of Linux and Windows, you really should compare with a Windows using non certified video drivers, because that is the reality in the Linux world. Most vendors supply "non certified" Linux drivers (if any) for their graphics chips.
My windows xp box has been up and heavily used for 4 weeks.
Not impressive. At our institute we have 9 dual CPU Linux boxes for running applications, so the older and slower UNIX workstations could be degraded to X terminals. With on average 7 users on each box, they are probably more heavily used than your single user workstation. And right now all nine have an uptime of more than 7 weeks.
My Linux workstation at home have been up for 25 days now, and IIRC last time it was rebooted was because of a networking problem, that turned out to be caused by a defective cable modem. So that reboot was even superfluous.
I guess what they will be selling is just a paper where they promise not to persecute you for using Linux
Ha! If anybody really want to buy such a piece of paper I will sell one as well. And I'd even sell it at half the price of SCOs (whatever their price is). And you know what: I'd even promise not to persecute you even if you use Windows. When I can sell a piece of paper compatible with Linux as well as Windows, it has gotta be more valuable than what SCO can produce. And at half the price I cannot see why anybody in the world would not want to buy this piece of paper. Unfortunately I don't think everybody in the world read slashdot, and in particular not this comment. So I have to find another way to tell everybody about the piece of paper I'm selling, perhaps I can find everybody's email address somewhere.
that would not necessarily mean that one could not contractually (e.g., via agreeing to a license agreement) give up one's right to reverse engineer.
The Danish "lov om ophavsret" (copyright law) says in paragraph 37:
Stk. 3. Bestemmelserne i stk. 1 og 2 kan ikke fraviges ved aftale. which translated is the provisions in stk. 1 and 2 cannot be departed from by agreeement. In other words the license agreement you suggest would be invalid according to Danish law. I have heard, that other European countries have similar laws.
I belive the argument is that the header files and stub code inserted by the linker constitute a derived work.
Some people think it is like that. I don't think so. A header file is supposed to be an interface definition. I don't consider compiling a header file against to be making a copy/derivated work of the library, just like I don't consider it to be a derivated work of the compiler. Inserting large inline functions in header files is an abuse of header files. People using the header files for compilation are not required to read those header files frist, they should not expect such functions. If you later attempt to use those functions as reason for making requirements about the license of code using the library, I believe you are commiting fraud (might not be the exact right term, I don't know, IANAL).
In other words. By placing a function in a header file, you are implicitly saying that you consider copying that function to be fair use.
Now, if you reverse engineer (clean room of course) the function signatures from the library
It is so much easier to reverse engineer with access to the source.:-)
Nobody can ever prove whether the original or reverse engineered header files were used. Neither whether the writer of the reverse engineered header files have looked on the originals.
and load it at runtime (rather than at loadtime)
I cannot see why that would make any difference. Both are going to happen on the end user's system, not the developer's system. In fact I don't even think you can make a clear definition of the difference between the two cases.
your software is clearly deriving in some ways from the GPLed work no matter what the technicality of how you link to it.
But as long as you don't copy any of the original code, that doesn't matter at all.
This is one reason I stick with the LGPL for libraries - it's how I think the GPL should work, but just in case I'm wrong in a legal sense it's good to be covered.
Could be a good idea, if things turn out differently from what we expect.
That depends on who posts the tinurl link. In that particular case the link should be avoided. But there is for example nothing wrong with 6cr. I do however see no reason for using tinyurl for slashdot comments. Tinyurl is nice for usenet, but on slashdot you should rather use the final url.
But I haven't changed it! I'm only linking to it dynamically at
runtime!
In that case I don't think the GPL on the library can affect the program using the library. Since the linking is done on runtime, it is done by the end user, which is permited by the GPL no matter what license the other code has. FSF like to say the user of a dynamically linked library GPL library must be GPL as well. I don't believe them. The developer of the program using the library can distribute his own code under whatever license he want, since it doesn't contain any of the library code he cannot be violating the copyright on the library.
Who says 'self-aware networks' are even possible? I've seen no evidence to show that they are.
Where is the evidence, that any part of the human brain can do anything that cannot be simulated by a computer. Surely one computer to simulate each brain cell is unrealistic, because we don't have that many computers. But with sufficient parallelity there is no reason to think they couldn't get self-aware.
Compiling under Windows does not require any special privilege, you only need Administrator level to install software, and some software doesn't need even that.
How do they dare trying to steal an abbreviation that for years has meant Internet Protocol. They talk about rights, but who is it, that go out and steal abbreviations. And BTW please God make ignorants stop using the term TCP/IP when they really don't know what they are talking about.
Sure tar does retain some metadata, but only one well defined set of metadata. If a filesystem introduces new types of metadata, they don't magically get supported by tar, neither do they get supported by all other utilities.
For a long time I had to do that each day, finally one day even that did not work. All my computer would see was a USB unit that timed out, there was no way it could see which kind of USB device was attached. So I called my ISP. They told me, that when this has been working for half a year without problems, and then one day for no particular reason stops working, and rebooting all the equipment doesn't help. It surely could not be a hardware problem, it had got to be a software problem. Of course they first figured that out after I told them I was using Linux (and BTW had tested with two different computers already - running two different versions of RedHat).
Slashdot Mirror
Sure, and don't blame Windows for it's crashes either then. Those are all caused by bad drivers as well.
We are talking about VMWare, which should have been able to run without additional drivers at all. Part of the reason it doesn't is probably the IA32 design, that simply makes virtualization more difficult than it ought to be. However if you throw such stuff into your kernel, you simply cannot expect it to be as stable as it used to be.
They want to pay to have no bugs.
In that case the first thing to do is to buy well documented hardware that actually works like the documentation says. When that has done it wouldn't be expensive to get bugs fixed as they are found. (Everybody ought to know, that a bug can never be fixed before it is found). With Windows OTOH you cannot pay somebody to fix the bugs as they are found.
Most crashes are either something wonky with the hardware or bad drivers or bad programming on the software manufacturer (not the OS, the app).
The last statement simply doesn't make sense. If an application program running without administrator priveleges can crash the OS, it is by definition a bug in the OS (or hardware).
But has it ever been taken to court? (Not that I would want to be the first one.)
If you have a license and want to use it on a different computer, you could just do it. It is not like Microsoft is ever going to find out about that. I wonder under which circumstances this would be taken to court.
Microsoft is going to be very, very interested in auditing these guys.
Sure, but I certainly hope Microsoft doesn't get everything they want. I could name a lot of stuff I would be interested in, that I'm probably never going to get. What rights do Microsoft have to perform such an audit?
Probably the VMware drivers.
So don't blame Linux for that.
Gentoo crashed on me. ATA raid drivers, the open source ones. The closed source ones work flawlessly.
Interesting case. There are multiple possible explanations. Could be that the drivers where not just tested well enough, doesn't sound like a piece of hardware everybody have. So having one you should at least provide a decent bug report, otherwise don't expect it to be fixed. It could of course also be the case, that no usable documentation is available for the device. That could explain why the closed source ones worked, the authors simply had some knowledge about the device that wasn't documented, and was required for stable operation.
SMP + SCSI is a no no under Linux in some kernels.
We have a dual Xeon (which means four logical CPUs in total) NFS server using SCSI, and it works just fine. High load has not yet been a problem, the load is rarely above 0.09.
I can't recall the last time my xp blue screened except for runnin a non certified video driver.
If that is true, it is really not much different from Linux. When running RH7.3 I had frequent crashes caused by the video driver, luckily it was solved when upgrading to RH9. In my experience most Linux crashes are caused by bad hardware and/or drivers. With Linux there is no such thing as a certified driver, but in reallity it probably compares to closed source drivers which will taint the kernel. So if we translate your statement into Linux terms, we will get a statement the I guess multiple Linux users could say: I can't recall the last time my Linux paniced except for runing a closed source video driver. In my particular case I was lucky enough, that the driver was open source, and a minor change eliminated the panics. But many video chips don't have open source drivers, and no specs are available for one to be developed. So if you want a fair comparision between the stability of Linux and Windows, you really should compare with a Windows using non certified video drivers, because that is the reality in the Linux world. Most vendors supply "non certified" Linux drivers (if any) for their graphics chips.
My windows xp box has been up and heavily used for 4 weeks.
Not impressive. At our institute we have 9 dual CPU Linux boxes for running applications, so the older and slower UNIX workstations could be degraded to X terminals. With on average 7 users on each box, they are probably more heavily used than your single user workstation. And right now all nine have an uptime of more than 7 weeks.
My Linux workstation at home have been up for 25 days now, and IIRC last time it was rebooted was because of a networking problem, that turned out to be caused by a defective cable modem. So that reboot was even superfluous.
Capitalism is a rather nice thing when it works.
Microsoft is really the best proof that capatalism is not always good.
I guess what they will be selling is just a paper where they promise not to persecute you for using Linux
Ha! If anybody really want to buy such a piece of paper I will sell one as well. And I'd even sell it at half the price of SCOs (whatever their price is). And you know what: I'd even promise not to persecute you even if you use Windows. When I can sell a piece of paper compatible with Linux as well as Windows, it has gotta be more valuable than what SCO can produce. And at half the price I cannot see why anybody in the world would not want to buy this piece of paper. Unfortunately I don't think everybody in the world read slashdot, and in particular not this comment. So I have to find another way to tell everybody about the piece of paper I'm selling, perhaps I can find everybody's email address somewhere.
that would not necessarily mean that one could not contractually (e.g., via agreeing to a license agreement) give up one's right to reverse engineer.
The Danish "lov om ophavsret" (copyright law) says in paragraph 37: Stk. 3. Bestemmelserne i stk. 1 og 2 kan ikke fraviges ved aftale. which translated is the provisions in stk. 1 and 2 cannot be departed from by agreeement. In other words the license agreement you suggest would be invalid according to Danish law. I have heard, that other European countries have similar laws.
Died last year, I think.
That is correct.
The redhat iso is about 0.00000000001% of all bittorrent traffic.
99.6% of all statistics are made up.
I belive the argument is that the header files and stub code inserted by the linker constitute a derived work.
:-)
Some people think it is like that. I don't think so. A header file is supposed to be an interface definition. I don't consider compiling a header file against to be making a copy/derivated work of the library, just like I don't consider it to be a derivated work of the compiler. Inserting large inline functions in header files is an abuse of header files. People using the header files for compilation are not required to read those header files frist, they should not expect such functions. If you later attempt to use those functions as reason for making requirements about the license of code using the library, I believe you are commiting fraud (might not be the exact right term, I don't know, IANAL).
In other words. By placing a function in a header file, you are implicitly saying that you consider copying that function to be fair use.
Now, if you reverse engineer (clean room of course) the function signatures from the library
It is so much easier to reverse engineer with access to the source.
Nobody can ever prove whether the original or reverse engineered header files were used. Neither whether the writer of the reverse engineered header files have looked on the originals.
and load it at runtime (rather than at loadtime)
I cannot see why that would make any difference. Both are going to happen on the end user's system, not the developer's system. In fact I don't even think you can make a clear definition of the difference between the two cases.
your software is clearly deriving in some ways from the GPLed work no matter what the technicality of how you link to it.
But as long as you don't copy any of the original code, that doesn't matter at all.
This is one reason I stick with the LGPL for libraries - it's how I think the GPL should work, but just in case I'm wrong in a legal sense it's good to be covered.
Could be a good idea, if things turn out differently from what we expect.
Parent is link to tubgirl.
That is not true.
Never trust TinyURL.
That depends on who posts the tinurl link. In that particular case the link should be avoided. But there is for example nothing wrong with 6cr. I do however see no reason for using tinyurl for slashdot comments. Tinyurl is nice for usenet, but on slashdot you should rather use the final url.
But I haven't changed it! I'm only linking to it dynamically at runtime!
In that case I don't think the GPL on the library can affect the program using the library. Since the linking is done on runtime, it is done by the end user, which is permited by the GPL no matter what license the other code has. FSF like to say the user of a dynamically linked library GPL library must be GPL as well. I don't believe them. The developer of the program using the library can distribute his own code under whatever license he want, since it doesn't contain any of the library code he cannot be violating the copyright on the library.
Who says 'self-aware networks' are even possible? I've seen no evidence to show that they are.
Where is the evidence, that any part of the human brain can do anything that cannot be simulated by a computer. Surely one computer to simulate each brain cell is unrealistic, because we don't have that many computers. But with sufficient parallelity there is no reason to think they couldn't get self-aware.
Compiling under Windows does not require any special privilege, you only need Administrator level to install software, and some software doesn't need even that.
That is no different from other systems.
How do they dare trying to steal an abbreviation that for years has meant Internet Protocol. They talk about rights, but who is it, that go out and steal abbreviations. And BTW please God make ignorants stop using the term TCP/IP when they really don't know what they are talking about.
(c)God Himself
And of course you don't care infriging his rights?
(like tar) which DOES retain metadata.
Sure tar does retain some metadata, but only one well defined set of metadata. If a filesystem introduces new types of metadata, they don't magically get supported by tar, neither do they get supported by all other utilities.
powercycle your cable modem
For a long time I had to do that each day, finally one day even that did not work. All my computer would see was a USB unit that timed out, there was no way it could see which kind of USB device was attached. So I called my ISP. They told me, that when this has been working for half a year without problems, and then one day for no particular reason stops working, and rebooting all the equipment doesn't help. It surely could not be a hardware problem, it had got to be a software problem. Of course they first figured that out after I told them I was using Linux (and BTW had tested with two different computers already - running two different versions of RedHat).
The "Linux 8.0" was a tad odd as well.
We know that 2.6 testing has just begun.
I mean, ethics, there is prior art for that right?
Obviously the patent office never heard about ethics.