Static v. Dynamic Linking -- redux

David Johnson david at
Sat Mar 16 21:20:42 UTC 2002

On Saturday 16 March 2002 01:16 am, Emiliano wrote:

> > Your first point is interesting. I'm wondering how one could demonstrate
> > that a library was meant to be invoked by non-derivative works. Here the
> > criteria that I would use: the library is a separate distinct package,
> > and the interface for the library is documented. These two criteria tell
> > me that the library was intended to be used for multiple programs by
> > multiple authors.
> Which still leaves it open to wrapping. LGPL-licensed wrapper service (with
> a distinct interface from the DLL) calls functionality from GPL DLL. This
> proxy interface too can be documented, and as said, alternate
> implementations could be whipped up in a snap, with the actual program
> being 'designed to' link against the proxy service. Unless LGPL servers
> can't link against GPL DLLs; I'm not entirely sure on that.

I don't think an LGPL wrapper changes anything. It may now be "legal" to link 
to the wrapper, but it still isn't legal for the application to link to the 
library via the wrapper. The manner in which an author presents his library 
is a strong indication of the author's intention for that library. The 
intention does not change just because a third person has written a wrapper 
for it.

David Johnson
pgp public key on website
license-discuss archive is at

More information about the License-discuss mailing list