On Thu, 15 Feb 2007 17:43:05 -0000, "Charles Moir" wrote: > And this focus on us releasing CDraw source implies that for some people > at least, perhaps the 'use elsewhere' is their goal and forget about the > Xara Xtreme product on Linux. There's no doubt that a release of CDraw to the Free Software community might lead to some interesting new applications of that code. In my view, that's one of the fundamental aspects that makes hacking on Free Software so rewarding---the knowledge that there aren't limits on coming up with new things to do with code. So here are three thoughts in response to the sentiment above: 1. Even if I couldn't fathom any interesting application of the code---that wouldn't change any of the analysis I provided before. Until it becomes a Free Software program, (meaning, something that doesn't depend on any non-free portions), Xara Xtreme will never really get picked up by the Free Software community. 2. Releasing the source code to CDraw is not a necessary part to Xara Xtreme becoming Free Software. If the current copyright holders are serious about wanting to engage the community, then making Xara Xtreme entirely free is essential. But CDraw doesn't have to be part of that. As discussed before, Xara Xtreme could be ported to use cairo and libCDraw.a could disappear along with all of its problems. And Xara still could maintain the use of CDraw in its commercial, proprietary offerings as a differentiating factor. If you're interested in extracting value from the secret aspect of CDraw, then that's a strategy for a proprietary software release, not a free software release. Now, Xara Xtreme+cairo might be painfully slow (perhaps even unusable) but I would still predict it would be much more well accepted that way than the current situation. I for one, (along with the rest of the cairo community), are already very motivated to improving cairo's performance. So embarrassing us with an application that would be extremely compelling if it weren't for cairo's performance bugs would be an excellent thing to do. We're always interested in receiving more interesting performance cases for cairo, (and I don't think there could be much more interesting than some of the graphics I've seen in .xar files). If there's a fear of "nobody would touch it if it's unusably slow" then you should recognize that you're currently trying to attract free software developers based on the characteristics of the part of the application that's not free. Do you see why that doesn't work? 3. Releasing CDraw under the GPL likely won't have the disastrous "use elsewhere" consequences you are imagining. I have to speculate here a bit, because I don't know exactly what the disastrous consequences are that you imagine. Is it that CDraw might get sucked into cairo, and then applications like inkscape could quickly start benefitting from it, and then inkscape might take developer mindshare instead of Xara? First, I'll ignore the fact that inkscape is already acknowledged as winning that mindshare already, even without any benefit from CDraw. Second, if CDraw were made available under the GPL, then it couldn't be immediately used in cairo, (not without cairo changing its license from LGPL/MPL to GPL, and that's quite unlikely to happen). For example, right now Zack Rusin at Trolltech is doing some really interesting things inside Qt with a vector-graphics library that's very much like cairo. It's available under the GPL, and the cairo community hasn't even looked at the code because of that, (in spite of the fact that Zack has recently added some stuff that we've been wanting in cairo for a while). Now, I actually don't love the fact that there's a bunch of duplicated effort in the community already between cairo and the Qt stuff. But I think the evidence is pretty good that GPL libraries do tend to maintain a sort of separate existence. (Interestingly, GPL applications wouldn't have any problem using GPL libraries, so maybe at some point GPL application authors will all start collaborating enough on GPL libraries that they start dominating, and things like LGPL libraries start disappearing.) Anyway, that's all mostly just to say that there's another way to provide what the community wants without necessarily changing the license of CDraw, (you could just rip it out instead). -Carl
Attachment:
pgpFYAB0PGiA1.pgp
Description: PGP signature