On Mon, 2002-08-05 at 16:55, Dalibor Topic wrote: > /* 2.1 */ > The second case is quite clear, I think. Take a > program explicitely using a GPLd library, for example > a java program using Qt through JNI. In that case the > license of the interpreter should not matter, as the > code is explicitely linking to a GPLd library.
You must be careful with this line of reasoning. For instance, someone just released a Qt implementation of AWT. I can't see that the existence of this Qt implementation has any effect on proprietary programs using the AWT. They might use the Qt AWT library and would never know that they weren't using the standard edition. Part of showing that malicious linking is showing that the distributor of the program "knows" invalid linking is occuring. A version of Debian with Kaffe, Sun JVM, non-GPL java, GPL java, the Qt AWT implementation, a standard AWT implementation and a proprietary AWT implementation is a reasonable and legal aggregation of code. There is a chance that someone might configure their box to use the Qt AWT and then run proprietary Java code but neither the original author of the non-GPL code had a clear intent to create this situation. This is one of the reasons I find RMS's resistance to digital rights management fascinating. The one facility that could insure GPL compliance would be DRM features built into the library loaders (ld and ClassLoaders) but that's another story. -- _____________________________________________________________________ Ean Schuessler [EMAIL PROTECTED] Brainfood, Inc. http://www.brainfood.com
