From mboxrd@z Thu Jan 1 00:00:00 1970 From: DJ Delorie To: smorris@nexen.com Cc: cygwin@sourceware.cygnus.com Subject: Re: Cygwin license Date: Tue, 16 Mar 1999 12:21:00 -0000 Message-id: <199903162021.PAA20648@envy.delorie.com> In-reply-to: < 199903161757.MAA12041@brocade.nexen.com > (message from SteveMorris on Tue, 16 Mar 1999 12:57:28 -0500 (EST)) References: <19990316130132.20506.rocketmail@send105.yahoomail.com> <19990316104140.A1113@cygnus.com> <199903161757.MAA12041@brocade.nexen.com> <199903161757.MAA12041@brocade.nexen.com> X-SW-Source: 1999-03/msg00513.html > I have seldom found RMS's thoughts to be compelling. You always have > to take his adgenda into account and his adgenda is quite complex. I > do suspect that without LGPL gcc would be a minor player. Nothing about gcc is LGPL. Only glibc is LGPL, and the only reason *that* is LGPL is because it provides no value beyond what any other C runtime provides, so there is no incentive to choose it over another. > Every deveoper I have ever met that volunteered to work on gcc cut > their gcc teeth using gcc in a place of employment that used gcc to > create licensed binary distributed code. This has nothing to do with GPL vs LGPL. The GPL just doesn't cover the results of *using* a GPL'd program, only using the program's sources. This has always been the case. > These people wouldn't be part of the free software movement without > LGPL. But they were part of the movement before there was an LGPL, and the LGPL doesn't apply to gcc anyway. > I think RMS wants to have it both ways, the broad distribution that > comes with people using gcc commercially plus the forcing of > software into the free software domain. He doesn't like to admit the > part that binary distributors play in supporting free software. RMS doesn't care about how broad the distribution is. He only cares that the software he writes and uses is free to be used by everyone. I tend to agree - I have no philanthropic interest in convering the world to my views as long as *I* have the programs *I* need. > As has been stated earlier on this list it is not clear that Cygnus > can restrict the distribution of code that runs under Cygwin. Since any cygwin application includes a significant portion of the runtime within it, and it's a runtime that does not simply interface you to the Win32 API (which is the platform you're running on), there's a clear value added that's part of the source upon which your binary was built (not including the dll). It is primarily this code that makes your application GPL. > Courts have held in other cases that glue software required for > inter-operability can be used regardless of license conflicts. The cygwin startup code is not just glue, and there's nothing else for us to be inter-operable with. Do you know of any other companies making a cygwin dll? If the code were simply a way of hooking you do the Win32 API, I might agree, but we're providing a significant amount of extra functionality on top of the Win32 API. The fact that said functionality emulates a POSIX environment is irrelevent; Win32 is not a posix environment so our code isn't there to make you interoperable with the OS. > If you consider sourceware cygwin.dll to be a general platform for > running UNIX code We do not. It is a tool that helps you port unix code to windows. The result is not a unix program any more, but a windows program. > If it is effectively impossible (i.e. prohibitively expensive) to > distribute code without including Cygnus licensed glue I suspect > that the license is unenforcable. Just because cygwin makes a hard job easy doesn't mean we're legally obliged to guarantee we'll make it easy for everyone. We have chosen a particular set of terms which you may agree to or not. If you don't agree to them, you simply have to find another way. The video game analogy is inappropriate - it *is* impossible to run game X on system Y without interfacing to system Y's APIs. It *is* possible to write program A to run on Win32 without interfacing with Cygwin's APIs. > In other words when Cygnus decided to make cygwin.dll freely > available they probably let the cat out of the bag for third party > applications. Nope. It has always had the GPL's restrictions on it; you are free to not use cygwin if you don't agree with the terms and you will still be able to write Win32 applications (just not with Cygwin). > (NOTE: It would probably depend largely on the license fee > arrangement. The game manufacturers wanted steep per cartridge > royalties.) Irrelevent in our case, since cygwin isn't a mandatory component of writing a Win32 program. > Think about it. If cygwin.dll is free and people can distribute third > party software with out sending money to Cygnus, why the heck did they > waste their stakeholders money investing in it? Well, the original reason was because *we* needed it to port GNUPro to NT. > I personally hope that my legal arguement is invalid and that > Cygnus' two track release strategy is legally supportable, i.e. you > can use the sourceware version of cygwin to develop sourceware, and > the commercial version to develop commercial software. Commercial > vendors end up paying for the support of sourceware. I think your argument is invalid because cygwin isn't a *mandatory* part of developing Win32 programs. It's a tool that makes it easier to do certain things, but people have been writing Win32 programs long before cygwin came along. Now, if *Microsoft* tried to license the import libraries for Win32, *then* I think there would be a legal issue. > In other words my legal hat is having a major arguement with my > sourceware hat. Am I showing signs of split personality? ;-) Yes. No. ;-) -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com From mboxrd@z Thu Jan 1 00:00:00 1970 From: DJ Delorie To: smorris@nexen.com Cc: cygwin@sourceware.cygnus.com Subject: Re: Cygwin license Date: Wed, 31 Mar 1999 19:45:00 -0000 Message-ID: <199903162021.PAA20648@envy.delorie.com> References: <19990316130132.20506.rocketmail@send105.yahoomail.com> <19990316104140.A1113@cygnus.com> <199903161757.MAA12041@brocade.nexen.com> X-SW-Source: 1999-03n/msg00513.html Message-ID: <19990331194500.tLWe1HH9HDGAPHoe2cDe2F1haVvXzKvzUbbmR5p4H1U@z> > I have seldom found RMS's thoughts to be compelling. You always have > to take his adgenda into account and his adgenda is quite complex. I > do suspect that without LGPL gcc would be a minor player. Nothing about gcc is LGPL. Only glibc is LGPL, and the only reason *that* is LGPL is because it provides no value beyond what any other C runtime provides, so there is no incentive to choose it over another. > Every deveoper I have ever met that volunteered to work on gcc cut > their gcc teeth using gcc in a place of employment that used gcc to > create licensed binary distributed code. This has nothing to do with GPL vs LGPL. The GPL just doesn't cover the results of *using* a GPL'd program, only using the program's sources. This has always been the case. > These people wouldn't be part of the free software movement without > LGPL. But they were part of the movement before there was an LGPL, and the LGPL doesn't apply to gcc anyway. > I think RMS wants to have it both ways, the broad distribution that > comes with people using gcc commercially plus the forcing of > software into the free software domain. He doesn't like to admit the > part that binary distributors play in supporting free software. RMS doesn't care about how broad the distribution is. He only cares that the software he writes and uses is free to be used by everyone. I tend to agree - I have no philanthropic interest in convering the world to my views as long as *I* have the programs *I* need. > As has been stated earlier on this list it is not clear that Cygnus > can restrict the distribution of code that runs under Cygwin. Since any cygwin application includes a significant portion of the runtime within it, and it's a runtime that does not simply interface you to the Win32 API (which is the platform you're running on), there's a clear value added that's part of the source upon which your binary was built (not including the dll). It is primarily this code that makes your application GPL. > Courts have held in other cases that glue software required for > inter-operability can be used regardless of license conflicts. The cygwin startup code is not just glue, and there's nothing else for us to be inter-operable with. Do you know of any other companies making a cygwin dll? If the code were simply a way of hooking you do the Win32 API, I might agree, but we're providing a significant amount of extra functionality on top of the Win32 API. The fact that said functionality emulates a POSIX environment is irrelevent; Win32 is not a posix environment so our code isn't there to make you interoperable with the OS. > If you consider sourceware cygwin.dll to be a general platform for > running UNIX code We do not. It is a tool that helps you port unix code to windows. The result is not a unix program any more, but a windows program. > If it is effectively impossible (i.e. prohibitively expensive) to > distribute code without including Cygnus licensed glue I suspect > that the license is unenforcable. Just because cygwin makes a hard job easy doesn't mean we're legally obliged to guarantee we'll make it easy for everyone. We have chosen a particular set of terms which you may agree to or not. If you don't agree to them, you simply have to find another way. The video game analogy is inappropriate - it *is* impossible to run game X on system Y without interfacing to system Y's APIs. It *is* possible to write program A to run on Win32 without interfacing with Cygwin's APIs. > In other words when Cygnus decided to make cygwin.dll freely > available they probably let the cat out of the bag for third party > applications. Nope. It has always had the GPL's restrictions on it; you are free to not use cygwin if you don't agree with the terms and you will still be able to write Win32 applications (just not with Cygwin). > (NOTE: It would probably depend largely on the license fee > arrangement. The game manufacturers wanted steep per cartridge > royalties.) Irrelevent in our case, since cygwin isn't a mandatory component of writing a Win32 program. > Think about it. If cygwin.dll is free and people can distribute third > party software with out sending money to Cygnus, why the heck did they > waste their stakeholders money investing in it? Well, the original reason was because *we* needed it to port GNUPro to NT. > I personally hope that my legal arguement is invalid and that > Cygnus' two track release strategy is legally supportable, i.e. you > can use the sourceware version of cygwin to develop sourceware, and > the commercial version to develop commercial software. Commercial > vendors end up paying for the support of sourceware. I think your argument is invalid because cygwin isn't a *mandatory* part of developing Win32 programs. It's a tool that makes it easier to do certain things, but people have been writing Win32 programs long before cygwin came along. Now, if *Microsoft* tried to license the import libraries for Win32, *then* I think there would be a legal issue. > In other words my legal hat is having a major arguement with my > sourceware hat. Am I showing signs of split personality? ;-) Yes. No. ;-) -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com