From: Austin Ziegler Date: 2003-07-10T06:18:48+09:00 Subject: Re: My brief and torrid affair with Ruby. -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thu, 10 Jul 2003 05:14:31 +0900, Sean O'Dell wrote: > I don't believe the GPL is specific about the library sharing the > same memory space as the linking application; for example, the > library may be linked through a TCP/IP pipe where the application > is running on a different machine than the library is. If the > application links intimately with the GPL library, the application > must also be GPL unless the library is distributed as a normal > part of a development environment or operating system. The FSF isn't really the arbiter of this. Their claim that dynamic linking creates a derivative work has not been resolved in any court of law or by legislation (I'm not sure that it's been shown for static linking, either, but that's much more likely to be shown as derivative work creating). Further, the GPL itself applies to *distribution*, not to *execution.* So if I distribute an application (A) that requires library (L), but do not distribute the library (L), then I am technically not in violation of the GPL, because the putative derivative work (A+L) exists only in memory and is not distributed. Note that GPL2 says nothing about remote access (e.g., GPLed programs that are accessed via the web or via SOAP or some other mechanism); GPL3 is likely to say something about this, although it will force me to immediately change any software that I own to "GPL2", not "GPL2 or later". If the GPL3 is changed the way that the Affero licence has been changed, then it truly is viral and anticommercial. - -austin - -- austin ziegler * austin@halostatue.ca * Toronto, ON, Canada software designer * pragmatic programmer * 2003.07.09 * 16:48:47 -----BEGIN PGP SIGNATURE----- Version: PGP 8.0 - not licensed for commercial use: www.pgp.com iQA/AwUBPwyGnmZOwWepqGDiEQLtOwCfTjtBvXLny0PfR4th4bwyv+nPHF4AoJSA ONrdEplEggzTR8ufW8kT5OYX =meu8 -----END PGP SIGNATURE-----