Tzahi Fadida wrote:
Is there a case study where an "API" (please note the use of " ") from a non-GPL to GPL was successfully defended?Personally, I'm not even familiar with a case in which an API was even contended.
not particularly in court but also in a public knowledge settlement out of court.
What is the closest one ("API") to the code?
using sockets? some kind of pipes? files? etc...
Regards,
tzahi.
Mind you, and API CAN be an API. I don't think calling method have anything to do with whether this is derived work or not.
Cases where you might claim that derived work relationship exist, and no (serious) contention happened:
- All kernel binary modules.
- Anyone still using MySQL from a commercial environment without paying (personally, I think claiming "derived work" here is utter rubbish, but that's just me).
- Anyone using MySQL from a non-GPL open source project
Cases where it's clear cut that no derived work status exists:
- User space Linux programs (by common sense, as well as by explicit statement in the license where the Linux kernel authors state that they view user space API usage to be non-derived from the kernel).
- Documented API usage in any library or OS - no one is claiming that my Win32 API is a derived work of Windows, or that by buying PkLib from PkWare, they become copyright holders on my program.
In short, cases where the API is well documented, and development is done against the documentation (vs. against the library's source code) are pretty clear cut "non-derived". You can see that most binary Linux kernel modules DON'T fall under that definition, as they do not use well documented APIs, and they DO usually have to look at the kernel's sources in order to develop them.
Shachar
-- Shachar Shemesh Lingnu Open Source Consulting ltd. http://www.lingnu.com/
================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
