On Sun, 2011-09-25 at 05:42 +0900, zxq9 wrote:
I hadn't considered that. And this makes sense as concerns complete programs.
I'm confident, however, that the intent of the GPLv3 when it was written was not to hamstring the resurrection and improvement of GPLv2 code by its own users with an intent to release to other users (the GPLv2 having been written before any of this was thought up)
Again, the intent that matters is that of the projects that chose to release under GPLv2-only and GPLv3-only, respectively, because neither wanted their work being used under the other license.
by way of forbidding inclusion of new code which amounts to nothing more than a format interpretation library but not an application or even a complete program of its own. The GPL attitude toward system libraries seems to strongly indicate this as well. But a format interpretation library is hardly "system level" so it doesn't qualify for the exclusions provided for system libraries explicit in the GPL.
I feel your pain, and personally believe that FSF's claim about dynamic linking is bogus, under which assumption GPLvN would become essentially equivalent to LGPLvN. (Though like the Fedora project, I am playing it safe and not acting in reliance on this.) But with respect to the intent of QCad CE, based on their choice of GPLv2 and not LGPLv2, I can only presume that they did intend to prohibit the kind of linking you are now pursuing (to the extent possible under copyright law). Sorry.
-
Matt McCutchen