On Fri, Jul 04, 2003 at 11:37:37AM +0200, Jean-Marc Lasgouttes wrote: > Some context: Ruurd would like us to distribute this port (the binaries for > now), but this is only possible if we change our license to explicitly allow > for linking against Qt non-commercial binary-only version.
Unless I'm grossly mistaken about the wide range of code contributions received over LyX's many, many years, it is not feasible to "change" the license in a manner clean and comprehensive enough to pass muster. All it does is raise questions about whether the LyX community is serious about enforcing rights under the GPL. That hurts the community, and it does harm to the GPL generally. And it provides another example of the GPL failing to meet the requirements of evolving software applications -- more grist for those critical of open source. The problem here is not with the GPL, but with the Trolltech business model and licensing practices, which puts open source applications under the GPL in this untenable position if developers wish to release Windows versions. Almost two years ago (?) I raised questions on this list (with strident language that I later came to regret; my apologies) about all the time, effort and (IMHO) diversion from potential enhancements, involved in making LyX "toolkit independent". I was rather shrill about the QT effort, in particular. In light of Trolltech's dual support for Unix and Windows, the QT frontend has always seemed destined to run aground on precisely this issue of linking against non-commercial libraries. Well, here we are. The "distribution without linking" to Windows users seems like a copout. What will happen, in reality, is that a handful of users will build linked binaries and distribute them to others. Some of this may fall within the dubious inside-the-same-organization GPL "exemption", but most will not. You can't deny this with a straight face.
