Brian H Wilson <[email protected]> writes: > MY understanding is that you can use GPL libraries without making your > derived program GPL -- spatialite or GDAL in this case -- see the "GPL > licensing exception". This gets into the whole GPL V1/V2/V3 > thing. Which license covers liblwgeom? It looks like it's GPL > V2. Parts are under more permissive licenses. (eg GDAL). > > From the wikipedia page: http://en.wikipedia.org/wiki/GPL_linking_exception
But postgis does not have an exception; it's pure GPLV2 > Spatialite itself is also GPL V2 _or_ LGPL _or_ Mozilla Public > license. I consider this a little confusing but I am not worried! They > just want to make it easy for you to use and prevent it from being > locked up or patented. Really I think they are trying to make the code usable with various other licenses by allowing the code to be copied under any of multiple licenses. There is a real problem of license incompatibility of free licenses where there really isn't an intent to prohibit some uses, but it comes out that way. This "disjunctive dual (triple) licensing" is a way to avoid that. > I think the real answer to Jim is "no" -- if you create a work that > uses spatialite and liblwgeom as libraries then you are working > entirely within the framework spelled out by GPL V2. Go for it! Is > there someone (Paul Ramsey? PostGIS foundation?) who can make this > unequivocal? If you don't change the library you are free to use it > in your own program without that program becoming GPL. That's incorrect. First, note that the GPL does not really restrict *creating* a derived work. It really addresses distribution. So if you create a derived work (your program linked with a GPL library, assuming that's really a derived work under copyright law) and distribute it, then you are in violation of the library's license, unless you also grant a license under the GPL to the entire source of the derived work. Note that this does not say "has to *become* GPL". You are free to offer a license under the GPL and also a BSD license, letting recipients choose. What you may not do is distribute the derived work where the recipient does not at least have permission to copy under the GPL. There's a subtlety, which is that people read "have to make GPL". Copyright law doesn't work that way. It instead says "if you don't offer GPL on the derived work, you are violating copyright law by coyping/deriving-from the original work without a license." It's sort of the same thing operationally, but the "have to" expression is (very common) fuzzy thinking > When I use open source code in a library and I make a change to the > library, I always send the change back. That's very nice of you (I meant that sincerely), but not all open source licenses require that. It's an interesting observation in that BSD culture, people still want changes back, but have chosen social pressure rather than a license requirement. > Personally I want people to use the work I do, so my own leaning is > towards licenses like X/MIT, Apache, BSD and so on that are not > confusing. As in: "Do anything you want with this. If you break it you > can keep both pieces." Certainly that's a valid choice for you. I don't think it's really about confusing, but about whether you want to permit others to create non-free derived works.
pgpLU62A1EJEx.pgp
Description: PGP signature
_______________________________________________ postgis-users mailing list [email protected] http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
